id
stringlengths 36
36
| source
stringclasses 15
values | formatted_source
stringclasses 13
values | text
stringlengths 2
7.55M
|
|---|---|---|---|
21cf86b0-2a15-4c50-8333-df579c82b251
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Meetup Report Thread: February 2014
A month ago, a new type of thread was proposed: a monthly page for meetup reports. The idea is that meetup attendees, or organizers, who wanted to share information about how the meetup went could do so in the comments of this thread. This is so information is dispersed, but without the need for anyone, and/or everyone, to dedicate their own thread to the report. The idea worked for January, and nobody had objections. So, we'll do this every month.
If you had an interesting Less Wrong meetup recently, but don't have the time to write up a big report to post to Discussion, feel free to write a comment here. Even if it's just a couple lines about what you did and how people felt about it, it might encourage some people to attend meetups or start meetups in their area.
If you have the time, you can also describe what types of exercises you did, what worked and what didn't. This could help inspire meetups to try new things and improve themselves in various ways.
If you're inspired by what's posted below and want to organize a meetup, check out this page for some resources to get started! You can also check FrankAdamek's weekly post on meetups for the week
|
23656f71-194e-448a-8480-8600f8846487
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Dying for a donation
The most outstanding feature of organ markets is that most people hate the idea. This is a curiosity deserving a second glance. There are organ shortages almost everywhere, with people dying on waiting lists hourly. To sentence them to death based on a cursory throb of disgust is not just uncivilised but murderous.
First I should get some technical details out of the way. An organ market can involve buying from living donors, or selling rights to organs after death, or both. Organs needn’t go to the rich preferentially; like any treatment, that depends on the healthcare system. The supply of organs available won’t decrease – if free donations dropped as a result of sales, the price would rise until either enough people sold organs or relatives and friends felt morally obliged to donate them anyway. A regulated market needn’t lead to an increase in stolen Chinese organ imports. It would lower the price here, making smuggling less worthwhile, while stopping Australians going on desperate holidays to seek organs in the under-regulated Third World.
That they ‘commodify the human body’ is the main objection to organ markets. They certainly do that, but why is commodification terrible? Well, a commodity is generally an object subordinated to the goal of making money. Treating other humans in that way leads to abominable actions. Slavery and organ theft are examples of human commodification that rightly repulse us. This doesn’t generalise however. The horror in these examples is that people are being made miserable because they don’t want to be sold. This is a completely different scenario to people voluntarily commodifying themselves.
After all, if commodifying people is inherently wrong, why allow paid labour? Renting out a portion of your time, mind and body to a company or government is surely commodification in the same vein. Or is selling body parts just too much commodification? It doesn’t seem so to me – you can lose more of your most personal possession, your l
|
c00ad866-6258-4e0a-b5c9-894eccaa28f0
|
trentmkelly/LessWrong-43k
|
LessWrong
|
New LW Meetups: Paris, San Francisco
This summary was posted to LW main on January 2nd. The following week's summary is here.
New meetups (or meetups with a hiatus of more than a year) are happening in:
* Paris LW Meetup - LHC Exhibit: 17 January 2015 02:00PM
* San Francisco Meetup: 12 January 2015 06:00PM
Irregularly scheduled Less Wrong meetups are taking place in:
* Bangalore Meetup: 10 January 2015 11:02AM
* European Community Weekend 2015: 12 June 2015 12:00PM
* [Frankfurt] New Year meetup Frankfurt: 11 January 2015 02:00PM
* [Munich] January Meetup in Munich: 17 January 2015 03:00PM
* Utrecht: a critique of effective altruism: 04 January 2015 02:00PM
* Warsaw January Meetup: 13 January 2015 06:00PM
The remaining meetups take place in cities with regular scheduling, but involve a change in time or location, special meeting content, or simply a helpful reminder about the meetup:
* Austin, TX - Caffe Medici: 03 January 2026 01:30PM
* London First 2015 Meetup, 04/01/2015: 04 January 2015 02:00PM
* [Melbourne] January 2015 Rationality Dojo - How to learn faster and teach more effectively and teaching: 04 January 2015 03:30PM
* Vienna: 24 January 2015 03:00PM
* West LA—What Is FAI?: 07 January 2015 07:00PM
Locations with regularly scheduled meetups: Austin, Berkeley, Berlin, Boston, Brussels, Buffalo, Cambridge UK, Canberra, Columbus, London, Madison WI, Melbourne, Moscow, Mountain View, New York, Philadelphia, Research Triangle NC, Seattle, Sydney, Toronto, Vienna, Washington DC, and West Los Angeles. There's also a 24/7 online study hall for coworking LWers.
If you'd like to talk with other LW-ers face to face, and there is no meetup in your area, consider starting your own meetup; it's easy (more resources here). Check one out, stretch your rationality skills, build community, and have fun!
In addition to the handy sidebar of upcoming meetups, a meetup overview is posted on the front page every Friday. These are an attempt to collect information on all the meetups happening
|
c6b8d7c7-d82f-4a5a-9116-099b77bc73e2
|
trentmkelly/LessWrong-43k
|
LessWrong
|
AI #83: The Mask Comes Off
We interrupt Nate Silver week here at Don’t Worry About the Vase to bring you some rather big AI news: OpenAI and Sam Altman are planning on fully taking their masks off, discarding the nonprofit board’s nominal control and transitioning to a for-profit B-corporation, in which Sam Altman will have equity.
We now know who they are and have chosen to be. We know what they believe in. We know what their promises and legal commitments are worth. We know what they plan to do, if we do not stop them.
They have made all this perfectly clear. I appreciate the clarity.
On the same day, Mira Murati, the only remaining person at OpenAI who in any visible way opposed Altman during the events of last November, resigned without warning along with two other senior people, joining a list that now includes among others several OpenAI co-founders and half its safety people including the most senior ones, and essentially everyone who did not fully take Altman’s side during the events of November 2023. In all those old OpenAI pictures, only Altman now remains.
OpenAI is nothing without its people… except an extremely valuable B corporation. Also it has released its Advanced Voice Mode.
Thus endeth the Battle of the Board, in a total victory for Sam Altman, and firmly confirming the story of what happened.
They do this only days before the deadline for Gavin Newsom to decide whether to sign SB 1047. So I suppose he now has additional information to consider, along with a variety of new vocal celebrity support for the bill.
Also, it seems Ivanka Trump is warning us to be situationally aware? Many noted that this was not on their respective bingo cards.
TABLE OF CONTENTS
1. Introduction.
2. Table of Contents.
3. Language Models Offer Mundane Utility. People figure out how to use o1.
4. Language Models Don’t Offer Mundane Utility. Is o1 actively worse elsewhere?
5. The Mask Comes Off. OpenAI to transition to a for-profit, Mira Murati leaves.
6. Deepfaketown and Botpoca
|
e917deae-f269-4478-9d1f-1f5f897e16ca
|
trentmkelly/LessWrong-43k
|
LessWrong
|
The steering problem
Most work on AI safety starts with a broad, vague problem (“How can we make an AI do good things?”) and relatively quickly moves to a narrow, precise problem (e.g. "What kind of reasoning process trusts itself?“).
Precision facilitates progress, and many serious thinkers are skeptical of imprecision. But in narrowing the problem too far we do most of the work (and have most of the opportunity for error).
I am interested in more precise discussion of the big-picture problem of AI control. Such discussion could improve our understanding of AI control, help us choose the right narrow questions, and be a better starting point for engaging other researchers. To that end, consider the following problem:
The steering problem: Using black-box access to human-level cognitive abilities, can we write a program that is as useful as a well-motivated human with those abilities?
I recently wrote this document, which defines this problem much more precisely (in section 2) and considers a few possible approaches (in section 4). As usual, I appreciate thoughts and criticism. I apologize for the proliferation of nomenclature, but I couldn’t get by without a new name.
----------------------------------------
I think the steering problem captures a large part of what most people think of as “the AI safety problem.” It certainly does not capture the entire problem; in particular, we might well introduce undesired goal-directed behavior in the process of implementing human-level capabilities (either inadvertently or because it’s the easiest way to produce human-level abilities).
Since I’ve started thinking more explicitly about the steering problem, I’ve reduced my estimate of its difficulty. This leads me to be more optimistic about AI safety, but also to suspect that the steering problem is a smaller share of the whole problem than I’d originally thought. It would be great to see a more precise statement of the rest of the problem (which would probably subsume the steering proble
|
ef3c9bda-a5e7-420d-8bd9-d3f2a2d602bc
|
StampyAI/alignment-research-dataset/lesswrong
|
LessWrong
|
A moral backlash against AI will probably slow down AGI development
*Note 1 : This was posted to EA Forum on May 31* [*here*](https://forum.effectivealtruism.org/posts/veR4W92bZsTsGgS3D/a-moral-backlash-against-ai-will-probably-slow-down-agi)*, as a submission for the 2023 Open Philanthropy AI Worldviews contest, due May 31, 2023. It addresses Question 1: “What is the probability that AGI is developed by January 1, 2043?”*
*Note 2: I'm developing a series of several more detailed essays on active moral stigmatization of AI as a strategy for pausing, slowing, or stopping dangerous AI development; so I'd welcome any constructive feedback on this initial essay.*
**Overview**
People tend to view harmful things as evil, and treat them as evil, to minimize their spread and impact. If enough people are hurt, betrayed, or outraged by AI applications, or lose their jobs, professional identity, and sense of purpose to AI, and/or become concerned about the existential risks of AI, then an intense public anti-AI backlash is likely to develop. That backlash could become a global, sustained, coordinated movement that morally stigmatizes AI researchers, AI companies, and AI funding sources. If that happens, then AGI is much less likely to develop by the year 2043. Negative public sentiment could be much more powerful in slowing AI than even the most draconian global regulations or formal moratorium, yet it is a neglected factor in most current AI timelines.
**Introduction**
The likelihood of AGI being developed by 2043 depends on two main factors: (1) how technically difficult it will be for AI researchers to make progress on AGI, and (2) how many resources – in terms of talent, funding, hardware, software, training data, etc. – are available for making that progress. Many experts’ ‘AI timelines’ for predicting AI development assume that AGI likelihood will be dominated by the first factor (technical difficulty), and assume that the second factor (available resources) will continue increasing.
In this essay I disagree with that assumption. The resources allocated into AI research, development, and deployment may be much more vulnerable to public outrage and anti-AI hatred than the current AI hype cycle suggests. Specifically, I argue that ongoing AI developments are likely to provoke a moral backlash against AI that will choke off many of the key resources for making further AI progress. This public backlash could deploy the ancient psychology of moral stigmatization against our most advanced information technologies. The backlash is likely to be global, sustained, passionate, and well-organized. It may start with grass-roots concerns among a few expert ‘AI doomers’, and among journalists concerned about narrow AI risks, but it is likely to become better-organized over time as anti-AI activists join together to fight an emerging existential threat to our species. (Note that this question of anti-AI backlash likelihood is largely orthogonal to the issues of whether AGI is possible, and whether AI alignment is possible.)
I’m not talking about a violent Butlerian Jihad. In the social media era, violence in the service of a social cause is almost always counter-productive, because it undermines the moral superiority and virtue-signaling strategies of righteous activists. (Indeed, a lot of ‘violence by activists’ turns out to be false flag operations funded by vested interests to discredit the activists that are fighting those vested interests.)
Rather, I’m talking about a non-violent anti-AI movement at the social, cultural, political, and economic levels. For such a movement to slow down the development of AGI by 2043 (relative to the current expectations of Open Philanthropy panelists judging this essay competition), it only has to arise sometime in the next 20 years, and to gather enough public, media, political, and/or investor support that it can handicap the AI industry’s progress towards AGI, in ways that have not yet been incorporated into most experts’ AI timelines.
An anti-AI backlash could include political, religious, ideological, and ethical objections to AI, sparked by vivid, outrageous, newsworthy failures of narrow AI systems. An anti-AI backlash could weakly delay AI research through government regulation. But it could strongly delay AI research through socio-cultural dynamics such as AI research becoming morally taboo, socially stigmatized, religiously condemned, and/or politically polarized. For example, if being an AI researcher became as publicly stigmatized as being a white nationalist, a eugenicist, a sexist, or a transphobe, then AI research would be largely abandoned by any researchers sensitive to social pressure, and AGI would not be developed for a long time.
Thus, we can invert the question of AGI timelines, and consider the possible timelines for an anti-AI backlash. Rather than asking ‘What is the likelihood that we’ll have AGI by 2043?’, we could ask ‘What is the likelihood that we will see an anti-AI backlash by 2043 – a backlash that is strong enough to slow down AGI development?’ I’d argue that the answer to this second question is fairly high. Even just in the last few weeks (as of May 31, 2023), we’ve seen a dramatic increase in public attention on AI risk, public and government concern about AI, and the beginnings of an anti-AI backlash on social media, such as Twitter.
(Note that in this essay I’m not taking a position on whether an anti-AI backlash would be a good thing or a bad thing; I’m just doing a preliminary analysis of how such a backlash could slow down AGI timelines.)
**Triggers for an anti-AI backlash**
The general public is already culturally primed for an anti-AI backlash. Ever since the novel *Frankenstein* (1818), we’ve had generations of science fiction novels, movies, TV shows, computer games, and other media portraying the dangers of creating artificial intelligence. Most living people in developed countries have been exposed to these cautionary tales. They’ve mostly seen *2001: A Space Odyssey, The Terminator, Ex Machina, Black Mirror*, and *Westworld*. They’re often the first things that ordinary people think about when they think about AI. And most adults have first-hand experience of playing computer games against powerful (but narrow) AI, e.g. trying to win ‘Civilization’ on ‘god mode’ difficult level.
The triggers for an anti-AI backlash don’t need to create moral stigma from scratch. They just need to connect these latent cultural fears of AI to current real-world AI issues. I’ll call these issues ‘triggers’, and there are several kinds that seem quite likely to provoke moral stigmatization of AI within the next 20 years.
**Trigger 1: Unemployment**
People get pretty upset when they lose their jobs. The closer we get to AGI, the more job losses we’ll see. And, for any ‘new jobs’ that open up due to increased economic activity, AI systems will probably be able to learn the new job faster than humans will be able to re-train to do them.
Insofar as Large Language Models are making faster progress in human-style information processing than autonomous robotics are making in doing physical tasks, AI job losses may start hitting white-collar professional who do ‘brain work’ before they hit blue-collar workers doing physical work. These white-collar professionals may include millions of suddenly unemployed lawyers, accountants, journalists, teachers, academics, medical staff, pharmacists, software engineers, graphic designers, architects, and civil engineers.
Such people are typically highly educated, politically engaged, and prone to adopting new moral stigmas through social media. If they’re unemployed, they would have all the time in the world to organize an anti-AI backlash movement. If they have some real estate equity, investment assets, and credit, they may have the money to keep fighting for a while, even without an income. If they have kids, who face poor career prospects in turn due to ongoing AI developments, they may feel the righteous fury of parents who are motivated to do anything necessary to secure a viable future for their next generation. Thus, AI-imposed unemployment is likely to provoke an anti-AI backlash, probably in the time scale of 5-20 years from now.
**Trigger 2: Sex**
Moral stigmatization often focuses on human sexuality. Sexual practices outside the mainstream have often provoked furious moral condemnation, across cultures and across history – whether it’s incest, polygamy, prostitution, cheating, BDSM, polyamory, or porn. As narrow AI gets applied to goods and services related to human sexuality, there are likely to be all kinds of moral backlashes from diverse groups, ranging from Christian conservatives to woke feminists.
New information technologies are often applied first to create new sexual content. Internet Rule 34 says ‘If it exists, there is porn of it; no exceptions’. A variant will be ‘If AI can make porn of it, there will be porn of it’. Possible applications of AI in the sexual domain have focused on AI-generated porn and erotica (whether photos, audio, video, or stories), deepfake porn, interactive girlfriends and boyfriends, and sexbots.
A key trigger for an anti-AI backlash could be the moral outrage and sexual disgust provoked by sexual applications of narrow AI, such as highly habit-forming interactive VR porn, or customized erotic chatbots with the voices, mannerisms, ad personality traits of someone’s neighbors, co-workers, or ex-lovers.
The most salient, intimate, and controversial application of AI in the next couple of decades will be, essentially, the production of interactive sex slaves – whether in real physical bodies, VR avatars, 2-D deepfake porn, or auditory chatbots. The moral condemnation of slavery remains very strong – it just hasn’t been applied yet to digital slaves. When AI researchers start to be seen as breeders and traders of digital sex slaves, they’re likely to be strongly stigmatized.
Many of these sexual AI applications will take highly controversial forms. Pedophiles will buy AI sexbots with children’s bodies. Sadistic psychopaths will use disposable AI sexbots that can be flogged, cut, and branded, and that scream in realistic pain. Guys who like futanari porn will use sexbots that combine the primary and secondary sexual traits of males and females. AI-generated deepfake porn of politicians, tech billionaires, media celebrities, journalists, and activists is especially likely to provoke the wrath of the rich, powerful, and influential.
The marketing and use of these sexual AI applications may be private at first, but there will inevitably by news coverage, and it will be written to provoke maximum moralistic outrage, because moralistic outrage sells, and gets clicks, and gets shares on social media.
**Trigger 3: Violence**
Many AI researchers have signed pledges not to develop lethal autonomous weapons (LAWs), such as ‘slaughterbots’. However, there are many other applications of narrow AI that could lead to widespread dangers, injuries, and deaths. Such violence often provokes moral outrage and intense stigmatization of the technologies involved.
The big danger here is not so much that AI safety engineers will stupidly overlook some obviously dangerous failure mode in their systems. Rather, the danger is that rogue nation-states, terrorists, bad actors, resentful former employees, aggrieved nihilists, creepy stalkers, or mischievous youth will manipulate or hack the AI systems to cause targeted deaths of mass carnage. Bad actors could hack self-driving cars to cause huge pile-ups on highways that lead to dozens of deaths. AI drones could be modified by terrorists, criminal gangs, or violent activists to cause mass shootings or explosions at public events. Autonomous assassination drones with face-recognition abilities and long-term loitering abilities could kill major heads of state. Obsessive stalkers could use AI systems to track, harass, and harm their sexual victims. Anarchists, anti-capitalists, and eco-activists who hate resource-intensive industries could hack AI factory control systems to cause horrific industrial accidents. Religious extremists could use AI propaganda systems to promote religious radicalization, terrorism, and warfare.
All of these violent AI applications will, of course, be dismissed and disavowed by the AI industry. But the public may notice the common denominator: AI allows highly effective, targeted violence that is displaced in time and space from the humans directing the violence. This increases the effectiveness and decreases the risks of doing all kinds of mayhem. This will strike many ordinary people as horrifying and outrageous, and will reinforce anti-AI sentiment.
**Other triggers**
Apart from unemployment, sex, and violence, there are many other applications of narrow AI that could exacerbate an anti-AI backlash. These include harmful effects on AI on women, children, elders, racial minorities, and sexual minorities. These include harmful effects of AI propaganda in political polarization and religious intolerance. Biomedical AI systems for drug discovery could lead to new, highly addictive, psychosis-inducing recreational drugs rather than cures for cancer. AI applied to consumer advertising, gambling, and investments could lead people into over-spending, debt, bankruptcy, divorce, and ruin. The number of harmful things that could go wrong with narrow AI systems is almost limitless – but each new type of harm will be an occasion for sensationalist news coverage, public outrage, virtue signaling, political condemnation, and moral stigmatization of AI.
**AI chokepoints that could delay AGI**
So what if there’s an anti-AI backlash? What could ordinary people actually do to slow down AI research, given the arms race dynamics between AI companies (such as Microsoft vs. Google) and nation-states (such as the US vs. China)? This section addresses some key resources required for AGI development that could be choked off by an anti-AI backlash. It’s not an exhaustive list of the ways that moral stigmatization of AI could handicap AI research. It’s just intended to give a sense of how strongly and comprehensively an anti-AI backlash could lead to another ‘AI winter’, or even to a decades-long ‘AI ice age’.
**Chokepoint 1: AI Talent**
If AI research becomes strongly morally stigmatized, all the prestige and coolness of being an AI researcher would evaporate. Moral stigmatization of a career does not just mean the career suffers a slight decline in status relative to other careers. No. The psychology of moral stigmatization means the general public views the career as evil, and views the people working in the career as morally tainted by that evil. Intense moral stigma against AI would mean that being an AI researcher is seen as being about as reputable as being a convicted sex offender, a Nazi racist, an arms dealer, or a mass murderer. The public would view AI researchers as hubris-driven mad scientists with psychopathic traits and genocidal aspirations.
Moral stigmatization of AI research would render AI researchers undateable as mates, repulsive as friends, and shameful to family members. Parents would disown adult kids involved in AI. Siblings wouldn’t return their calls. Spouses would divorce them. Landlords wouldn’t rent to them.
Once the anti-AI backlash renders AI researchers socially, sexually, and professionally toxic, this would radically reduce the quantity and quality of talent working in AI. People with the technical skills to do AI research would exit the field, and would work instead on cybersecurity, or crypto, or non-AI software, or robotics, or whatever. They would have many career options that aren’t viewed as evil by lots of people they meet.
People in some fields have already developed pretty thick skins for resisting stigma. Researchers in controversial areas such as behavior genetics, evolutionary psychology, intelligence research, sex differences, and race differences have been subject for decades to moral stigmatization, hostile stereotyping, career handicaps, lack of funding, and attacks by journalists. They’ve become self-selected for orneriness, disagreeableness, intellectual courage, emotional stability, and self-sufficiency, and they’re learned many coping strategies. By contrast, AI researchers have little experience of being morally stigmatized. They’re used to high status, prestige, income, and coolness. They may be shocked when the public suddenly turns against them and paints them as evil mad scientists consumed by hubris and misanthropy. In other words, the pool of AI talent is highly vulnerable to stigmatization, and has few defenses against it. Faced with a choice between staying in a highly stigmatized field (AI) versus switching to another highly-paid, intellectually engaging computer science field that is not highly stigmatized (e.g. gaming, cybersecurity, crypto), most AI researchers may jump ship and leave AI.
**Chokepoint 2: AI Funding**
A strong enough anti-AI backlash would lead to AI funding drying up. Investors have become quite sensitive to ‘ESG criteria’ concerning environmental, social, and governance issues. If AI becomes morally stigmatized, ESG criteria could quickly and easily include AI as a disqualifying taboo. Any company involved in AI would receive low ESG scores, and would attract less ‘ethical investment’.
Apart from formal ESG criteria, individual and institutional investors tend to avoid companies widely perceived as reckless, evil, and inhumane. Many investors already avoid companies involved in weapons, alcohol, tobacco, porn, or gambling. If AI becomes seen as a horrifying new weapon, an addictive entertainment, and/or an insanely risky species-level gamble, it would combine all the worst evils of these already-stigmatized industries.
Investors may aspire to be rational maximizers of risk-adjusted returns. But investors are also social primates, subject to the same social and moral pressures that shape human behavior in every other domain of life. High Net Worth Individuals (‘rich people’) often set up family offices to handle their investments, assets, and trusts for their kids and grand-kids. These family offices are specifically designed to take a long-termist, multi-generational perspective on the preservation and enhancement of dynastic wealth and power. That long-termist perspective naturally leads to a concern about multi-decade technological changes, geopolitical risks, global catastrophic risks, and existential risks. If AI becomes morally stigmatized as a major existential risk, family offices and their investment professionals will not want to deploy their capital in AI companies that could lead the rich people’s kids and grand-kids not to die out before the end of the 21st century.
The investment world, like every human world, is prone to moral fads and fashions. Some companies and industry sectors become viewed as morally righteous, saintly, and inspiring; others become viewed as morally disgusting, sinful, and degrading. The psychology of moral disgust runs on the logic of contagion: anything in, around, or near a morally stigmatized activity becomes morally stigmatized by proxy. This means that if a large publicly traded corporation such as Microsoft or Google happens to include a much smaller organization (such as OpenAI or DeepMind) that becomes stigmatized, the large corporation also becomes morally stigmatized. Fewer people want to invest in it. They don’t want their portfolio contaminated by the second-hand evil. As fewer investors are buying and more are selling, the share price falls. As the share price falls, other investors see the writing on the wall, and panic-sell. Hedge funds start aggressively shorting the stock. Soon the corporations face a dilemma: either they shut down or sell off the tainted AI organization poisoning their shareholder value from within, or they continue seeing their share price fall off a cliff – until they get acquired in a hostile takeover by new investors who are willing to cut the AI cancer out of the corporation, to save the rest of the company.
A few anti-ethical investors might see AI as a clever contrarian play, and might think AI company stocks are temptingly under-valued, and will become great investments after the moral stigma fades. But the stigma might not fade, and they may be left facing huge capital losses.
**Chokepoint 3: Suppliers**
Moral contagion flows out in all directions. If AI starts to be seen as evil, any other organization that does business with AI researchers or AI companies will be seen as evil, or at least evil-adjacent. They will be stigmatized by association, as often happens in ‘cancel culture’. AI research depends on all kinds of suppliers of goods and services, utilities, computational infrastructure, and business infrastructure.
A sort of ‘ethical back-propagation’ would happen, where the moral stigma of AI would propagate backwards along the supply chain, tainting every person and company that provides essential goods and services to AI research.
In response, every supplier who is sensitive to the anti-AI backlash may withdraw their support from AI research groups. This may include everyone supplying GPU hardware, software, cloud computing resources, office space, legal services, accounting services, banking services, and corporate recruiting services. AI businesses may find that no reputable lawyers, bookkeepers, banks, or headhunters are willing to work with them. At a more mundane level, AI groups may find that they cannot find reputable businesses willing to supply them with tech support, back-office staff, office temps, caterers, drivers, janitors, or security staff. If some companies are not willing to do business with cannabis shops, porn producers, drug gangs, arms dealers, racketeers, human traffickers, or other stigmatized forms of economic activity, and if AI becomes stigmatized to a similar level, AI research will be handicapped, and will slow down.
The supplier issue could also affect AI researchers in their personal lives. If AI is widely seen as a work of reckless, hubristic evil, AI researchers may find that landlords are not willing to rent to them, coop boards are not willing to let them buy condos, and daycare centers and private schools are not willing to care for their kids. Bodyguards and police may think they’re too disgusting to protect. Therapists may advise them to ‘seek help elsewhere’. They may even find spiritual services getting choked off, as their priest, pastor, or rabbi shun them for the sinful way they make a living.
**Chokepoint 4: Laws and regulations**
Informal moral stigmatization often leads to formal government regulations and laws governing new activities and technologies. Indeed, it’s often difficult to coordinate bipartisan support for new regulations and laws constraining something unless there is already a foundation of public stigmatization against that thing. Once the horrors of chemical weapons were witnessed in World War 1, and the public viewed mustard gas and other agents as morally outrageous, it was fairly easy to develop international bans on chemical weapons. Once human cloning became morally stigmatized in the 1990s, it was fairly easy to implement government bans and scientific norms against human cloning. Conversely, it’s quite difficult to sustain regulations and laws against something if the moral stigma against the thing erodes – as in the case of cannabis use gradually becoming destigmatized in the US since the 1960s, and legalization of recreational cannabis following in many states. Thus, moral stigma and government regulation often have mutually reinforcing functions.
In the case of AI, if an anti-AI backlash was sufficiently global in scale, and became a major focus of public concern in both the US and China, it may be much easier to develop international agreements to pause, constrain, or ban further AGI research. With global moral stigmatization of AI, global regulation of AI becomes feasible. Without global moral stigmatization of AI, global regulation of AI is probably impossible. Yet much of the work on AI governance seems to have ignored the role of informal moral stigmatization in creating, energizing, and sustaining formal international agreements.
If an anti-AI backlash gets formalized into strong laws and regulations against AGI development, leading governments could make it prohibitively difficult, costly, and risky to develop AGI. This doesn’t necessarily require a global totalitarian government panopticon monitoring all computer research. Instead, the moral stigmatization automatically imposes the panopticon. If most people in the world agree that AGI development is evil, they will be motivated to monitor their friends, family, colleagues, neighbors, and everybody else who might be involved in AI. They become the eyes and ears ensuring compliance. They can report evil-doers (AGI developers) to the relevant authorities – just as they would be motivated to report human traffickers or terrorists. And, unlike traffickers and terrorists, AI researchers are unlikely to have the capacity or willingness to use violence to deter whistle-blowers from whistle-blowing.
Laws and regulations by themselves would not be enough to significantly slow down AGI development. Bad actors would always be motivated to evade detection and accountability. However, it’s a lot harder to evade detection if there is a global moral stigma against AGI development, with strong public buy-in. From the public’s point of view, laws and regulations are simply ways to articulate, formalize, and implement moral stigmas that are already widely accepted in public discourse. In short, the public has already figured out what’s evil, and they just want government to use its monopoly on the legitimate use of force to deter and punish what’s evil. Thus, moral stigmatization super-charges the effectiveness of any formal laws and regulations around AI.
Often, if some activity becomes sufficiently stigmatized, regulators and law enforcement can apply existing laws in highly targeted ways to deter the activity. For example, laws against reckless endangerment and public endangerment could be applied to prosecute AGI research – if there was sufficient public and institutional belief that AGI imposes existential risks on citizens without their consent. The FBI could switch its focus from ‘white supremacy as the leading domestic terrorist threat’ to ‘AGI research as the leading domestic terrorist threat’ – and investigate and prosecute AI researchers accordingly. Note that government regulators and law enforcement agencies are often motivated to find and capitalize on any new threats that the public perceives. This provides pretexts for increasing their budgets, staff, and powers. If an anti-AI backlash becomes popular, many government workers will see this as a great opportunity to increase their status and power. Fighting against something widely considered an existential threat to humanity would sound like a pretty cool mission to a lot of FBI agents (in the US) or Ministry of State Security agents (in China). Thus, moral stigmatization of AI could lead quite quickly and directly to government investigations, audits, litigation, and prosecution of AI researchers and companies. Result: AGI development is slowed or stopped.
**Conclusion**
The social-psychological processes of moral stigmatization have evolved genetically and culturally over thousands of generations. Moral stigma plays crucial roles in solving group coordination problems, enforcing social norms, punishing anti-social behavior, and minimizing existential threats to groups. Stigmatization is both a deep human instinct and a powerful cultural tradition. It can solve problems that can’t be solved in any other way. This may include solving the problem of delaying AGI development until we have a better idea whether AI alignment is possible at all, and if it is possible, how to achieve it.
|
cc51ea9a-66af-4503-b39a-1484ec99a24d
|
StampyAI/alignment-research-dataset/arxiv
|
Arxiv
|
Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications
1 Introduction
---------------
\IEEEPARstart
Recently, applications with graph-based data have rapidly increased in many domains such as computer vision [[1](#bib.bib1)], Computer graphics [[2](#bib.bib2)], Physics [[3](#bib.bib3)], Chemistry [[4](#bib.bib4)] and now healthcare [[5](#bib.bib5)].
Graph Neural Networks (GNNs) have proven to be a very powerful tool to process non-Euclidean unstructured data [[6](#bib.bib6)], especially in healthcare applications such disease prediction [[7](#bib.bib7)], drug interaction [[8](#bib.bib8)] and discovery [[9](#bib.bib9)], brain connectome analysis [[10](#bib.bib10)], multi-modal data analysis [[11](#bib.bib11)], just to name a few.
Most of the GNN models for graph classification consider each input graph individually. They aggregate information between neighboring nodes via message passing, to obtain new node or graph representations to be used for the final classification layers. On the other hand, recent literature on disease prediction testify that considering information coming from similar patients, in form of a population level graph, is crucial to enhance performance of the model [[12](#bib.bib12)].
This is the case, for instance, of the method proposed in [[13](#bib.bib13)] for brain connectome analysis, where single patient’s brain graphs are converted into a vector representation and later leveraged in a population level graph with each node as an individual patient associated to a feature vector representing the individual brain graph.
Despite showing the importance of considering the population level information to the downstream task, all these methods operates on a flattened representation of the input data, obtained by some preprocessing step to embed the input graph on a vectorial space. As a consequence, the structural information from the individual graphs is often lost as the method only focuses on the resulting feature vector. Moreover, the input objects in healthcare or molecular prediction incorporate crucial structural and functional information within the graph structure itself that could be beneficial to the task at hand.
In this paper, we propose Graph-in-Graph (GiG), a graph learning architecture which exploits both structural information and node level features of the input graph samples, together with leveraging population level information by learning a graph between input samples.
Importantly, learning the population level graph provides a form of knowledge discovery through the learned latent graph, and enables reasoning about made decisions through learned neighborhoods around input samples.
Note that, most deep learning works optimize the downstream task only, and analyze the latent structure post-hoc, through mapping techniques like t-SNE [[14](#bib.bib14)] or UMAP [[15](#bib.bib15)]. Instead, we aim to learn the downstream task and construct the latent graph structure end-to-end. To the best of our knowledge, the combination of latent graph learning and graph-valued input data has not been investigated so far.
We also propose a degree distribution loss that regularizes the predicted latent relationship structure. This loss provides a parameter that allows the user to influence the connectivity degrees across samples. This can lead to better classification performance, but more importantly, it can facilitate knowledge discovery by forcing the model to limit the neighborhood sizes around input samples.
We evaluate our method on real-world datasets from different domains, one medical and two bioinformatic, and provide a comparison with state-of-the-art graph classification models.
2 Related Work
---------------
Protein classification, brain connectome classification, and toxicology analysis are among the most relevant biomedical applications in which the input data is naturally represented by a graph structure.
In the domain of protein classification, one of the first learning approaches was based on extracting feature vectors from proteins and classifying them into enzymes or non-enzymes using support vector machines (SVM) [[16](#bib.bib16)]. Later, variations were proposed, e.g. using C-Support Vector Machines [[17](#bib.bib17)] with graph kernels [[18](#bib.bib18)]. Recently, the use of Graph Convoutional Neural Networks (GCNNs) has been proposed as a general framework for graph and node representation learning, including for proteins classification [[2](#bib.bib2), [19](#bib.bib19), [20](#bib.bib20)]. GNNs have shown to be able to capture local structures which are usually characteristic of the graph representation of these data [[21](#bib.bib21), [22](#bib.bib22), [23](#bib.bib23)].
In particular, Zhang et al. [[24](#bib.bib24)] reached SOTA results with the use of pooling operators and structure learning, which down-sample graph data and obtain a better embedding for the downstream classification tasks .
Brain connectome analysis requires pre-processing of brain signal time-series from image/sensor data, in order to extract a connectivity graph between brain regions. In a recent work [[25](#bib.bib25)], the application of different models such as GCN, BrainNetCNN, and ElasticNet were compared on brain connectivity graphs from the Human Connectome Project (HCP) dataset [[26](#bib.bib26)], to suggest a unified brain data processing algorithm.
Toxicity prediction for molecules is a challenging task, and is well-studied thanks to open-access datasets like the Tox21 Data Challenge111Tox21 Data Challenge URL: <https://tripod.nih.gov/tox21/challenge/about.jsp>. For example, DeepTox [[27](#bib.bib27)] suggested to use Deep Neural Networks (DNNs) and multi-task learning, while Dmlab [[28](#bib.bib28)] and Microsomes [[29](#bib.bib29)] provided tree-based ensemble methods as a solution. Class balance issues in the input data were addressed in [[30](#bib.bib30)], by suggesting different bagging approaches through resampling techniques. These models used input data in the Simplified Molecular-Input Line-Entry (SMILES) format. However, SMILES does not optimally preserve the molecular structure [[31](#bib.bib31)]. Thus, the first bridge works [[32](#bib.bib32), [33](#bib.bib33)] combined different molecular fingerprints and molecular graphs as input. Recently a group of models were introduced that use the molecular graph as input [[34](#bib.bib34), [35](#bib.bib35), [31](#bib.bib31)]. For example, [[31](#bib.bib31)] suggested meta-learning framework to obtain better results on a small number of samples. Proper pooling operations are another way to improve the graph representation and to present the entire graph structure. For example, Graph Multiset Transformer (GMT) [[36](#bib.bib36)] uses a multi-head attention mechanism to better preserve graph structural information.
A common factor in all these applications is the recent proliferation of methods applying GCNNs directly on the input graph representation. GCNNs have indeed shown to be a powerfull tool to compute optimal nodes/graphs representations, showing to be the weapon of choice for learning tasks in several domains, such as social sciences [[37](#bib.bib37), [38](#bib.bib38)], computer vision and graphics [[39](#bib.bib39), [40](#bib.bib40), [2](#bib.bib2)], physical [[41](#bib.bib41), [42](#bib.bib42), [43](#bib.bib43)], as well as medical and biological sciences [[13](#bib.bib13), [44](#bib.bib44), [45](#bib.bib45), [46](#bib.bib46), [47](#bib.bib47), [48](#bib.bib48), [49](#bib.bib49)].
Another breakthrough in graph classification, especially in the domain of brain imaging and disease prediction, has been considering relationships between patients in the form of a population level graph [[44](#bib.bib44), [13](#bib.bib13), [50](#bib.bib50), [7](#bib.bib7), [51](#bib.bib51), [52](#bib.bib52)].
Parisot et al. [[44](#bib.bib44), [13](#bib.bib13)] suggested to build a population graph as a sparse graph, where nodes represent imaging data and edges phenotypic information. Later, Kazi et al. [[51](#bib.bib51)] proposed to construct several graphs each corresponding to one of the demographic element and combine them in self-attention layer. Similarly, in [[52](#bib.bib52)] was proposed to combine separately constructed population graphs for diagnosing autism spectrum disorder. More recently, authors from [[11](#bib.bib11)] proposed a continuous differentiable graph module (cDGM) to learn a weighted adjacency matrix representing the population-level graph by training end-to-end the graph generation and the classification weights of the model. In [[53](#bib.bib53)], cDGM is expanded by training a probabilistic graph generator and sampler able to handle sparse population graphs. However, all these methods assume vector-valued input data, either in form of a flattened feature vector at input, or in form of an embedding vector obtained from an input module.
However, all the methods exploiting a population level graph assume vector-valued input data of the input samples, either in form of a flattened feature vector at input or in form of an embedding vector obtained from an input module, even when the input data are naturally represented by a graph structure.
On the other hand, SOTA methods for graph classification consider input graphs in isolation and none of the methods working in molecule classification consider a latent graph between molecules to obtain better representations. We argue that preserving the initial graph structures at input, and simultaneously learning their embedding in an end-to-end fashion, can be beneficial to obtain a better representation for downstream tasks, and to extract only task-relevant information from feature vectors. Moreover, most datasets in the healthcare or bioinformatics domain suffer from an insufficient amount of data and missing information. In these cases, being able to recover missing information from neighboring nodes, which is enabled through the learned latent graph, is a crucial factor to provide a correct classification.
This paper aims exactly at filling this gap, proposing a method that works directly on the graph representation of the input data and learns a population level latent graph in a end-to-end fashion by optimizing the classification task loss.
3 Method
---------
Figure 1: Graph-in-Graph architecture consists of three parts: F1subscript𝐹1F\_{1}italic\_F start\_POSTSUBSCRIPT 1 end\_POSTSUBSCRIPT represents the node-level module, F2subscript𝐹2F\_{2}italic\_F start\_POSTSUBSCRIPT 2 end\_POSTSUBSCRIPT - the population-level module, and F3subscript𝐹3F\_{3}italic\_F start\_POSTSUBSCRIPT 3 end\_POSTSUBSCRIPT - GNN classifier. As input model has graphs 𝐆𝐢subscript𝐆𝐢\mathbf{G\_{i}}bold\_G start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT, then its representation 𝐡𝐢subscript𝐡𝐢\mathbf{h\_{i}}bold\_h start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT is obtained from F1subscript𝐹1F\_{1}italic\_F start\_POSTSUBSCRIPT 1 end\_POSTSUBSCRIPT.
The input to our model is a set of N𝑁Nitalic\_N graphs, 𝐆=𝐆𝟏,𝐆𝟐,…,𝐆𝐍𝐆subscript𝐆1subscript𝐆2…subscript𝐆𝐍\mathbf{G}={\mathbf{G\_{1}},\mathbf{G\_{2}},...,\mathbf{G\_{N}}}bold\_G = bold\_G start\_POSTSUBSCRIPT bold\_1 end\_POSTSUBSCRIPT , bold\_G start\_POSTSUBSCRIPT bold\_2 end\_POSTSUBSCRIPT , … , bold\_G start\_POSTSUBSCRIPT bold\_N end\_POSTSUBSCRIPT. Each ithsuperscript𝑖𝑡ℎi^{th}italic\_i start\_POSTSUPERSCRIPT italic\_t italic\_h end\_POSTSUPERSCRIPT graph is defined as 𝐆𝐢=(𝐕𝐢,𝐄𝐢,𝐗𝐢)subscript𝐆𝐢subscript𝐕𝐢subscript𝐄𝐢subscript𝐗𝐢\mathbf{G\_{i}}=(\mathbf{V\_{i}},\mathbf{E\_{i}},\mathbf{X\_{i}})bold\_G start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT = ( bold\_V start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT , bold\_E start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT , bold\_X start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT ), where 𝐕𝐢subscript𝐕𝐢\mathbf{V\_{i}}bold\_V start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT and 𝐄𝐢subscript𝐄𝐢\mathbf{E\_{i}}bold\_E start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT are vertices and edges of the graph and 𝐗𝐢∈ℝ‖𝐕𝐢‖×Dsubscript𝐗𝐢superscriptℝnormsubscript𝐕𝐢𝐷\mathbf{X\_{i}}\in\mathbb{R}^{\|\mathbf{V\_{i}}\|\times D}bold\_X start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT ∈ blackboard\_R start\_POSTSUPERSCRIPT ∥ bold\_V start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT ∥ × italic\_D end\_POSTSUPERSCRIPT is the node features matrix with D𝐷Ditalic\_D being number of features.
The output of our model is a class probability vector 𝐩∈ℝC𝐩superscriptℝ𝐶\mathbf{p}\in\mathbb{R}^{C}bold\_p ∈ blackboard\_R start\_POSTSUPERSCRIPT italic\_C end\_POSTSUPERSCRIPT for each input graph, where C𝐶Citalic\_C is the numbers of possible classes to which the input graph can be categorized.
###
3.1 Graph-in-Graph model
In this section we provide the technical details of the proposed GiG model (Fig. [1](#S3.F1 "Figure 1 ‣ 3 Method ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications")).
The proposed model consists of three parts: 1) the node-level module F1subscript𝐹1F\_{1}italic\_F start\_POSTSUBSCRIPT 1 end\_POSTSUBSCRIPT; 2) the population-level module F2subscript𝐹2F\_{2}italic\_F start\_POSTSUBSCRIPT 2 end\_POSTSUBSCRIPT; 3) the final GNN classifier F3subscript𝐹3F\_{3}italic\_F start\_POSTSUBSCRIPT 3 end\_POSTSUBSCRIPT (Fig. [1](#S3.F1 "Figure 1 ‣ 3 Method ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications")). In a nutshell F1subscript𝐹1F\_{1}italic\_F start\_POSTSUBSCRIPT 1 end\_POSTSUBSCRIPT computes input graph representations, F2subscript𝐹2F\_{2}italic\_F start\_POSTSUBSCRIPT 2 end\_POSTSUBSCRIPT focuses on learning the latent connections between the input graphs, and F3subscript𝐹3F\_{3}italic\_F start\_POSTSUBSCRIPT 3 end\_POSTSUBSCRIPT takes as input both the graph representations from F1subscript𝐹1F\_{1}italic\_F start\_POSTSUBSCRIPT 1 end\_POSTSUBSCRIPT and the latent structure from F2subscript𝐹2F\_{2}italic\_F start\_POSTSUBSCRIPT 2 end\_POSTSUBSCRIPT and provides the predictions of the downstream task as output. In the following subsections we provide technical details of each of the three parts.
####
3.1.1 Node-level module F1subscript𝐹1F\_{1}italic\_F start\_POSTSUBSCRIPT 1 end\_POSTSUBSCRIPT
The node-level module is designed to directly handle non-Euclidian data (i.e. graphs) as input and capture only task-relevant features and structural information during the training. Mathematically F1subscript𝐹1F\_{1}italic\_F start\_POSTSUBSCRIPT 1 end\_POSTSUBSCRIPT is defined as, 𝐡𝐢=F1(𝐆𝐢)subscript𝐡𝐢subscript𝐹1subscript𝐆𝐢\mathbf{h\_{i}}=F\_{1}(\mathbf{G\_{i}})bold\_h start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT = italic\_F start\_POSTSUBSCRIPT 1 end\_POSTSUBSCRIPT ( bold\_G start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT ), where 𝐡=[𝐡𝟏,…,𝐡𝐍]𝐡subscript𝐡1…subscript𝐡𝐍\mathbf{h}=[\mathbf{h\_{1}},\dots,\mathbf{h\_{N}}]bold\_h = [ bold\_h start\_POSTSUBSCRIPT bold\_1 end\_POSTSUBSCRIPT , … , bold\_h start\_POSTSUBSCRIPT bold\_N end\_POSTSUBSCRIPT ] are output graph representations in a H dimensional latent space, with 𝐡𝐢∈ℝ1×Hsubscript𝐡𝐢superscriptℝ1𝐻\mathbf{h\_{i}}\in\mathbb{R}^{1\times H}bold\_h start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT ∈ blackboard\_R start\_POSTSUPERSCRIPT 1 × italic\_H end\_POSTSUPERSCRIPT. F1subscript𝐹1F\_{1}italic\_F start\_POSTSUBSCRIPT 1 end\_POSTSUBSCRIPT is implemented as a graph convolution based neural network composed by a set of graph convolutional layers followed by a pooling operator over all node features.
The specific choice of the graph convolutional operator to use in F1subscript𝐹1F\_{1}italic\_F start\_POSTSUBSCRIPT 1 end\_POSTSUBSCRIPT depends on the dataset and the task at hand.
####
3.1.2 Population-level module F2subscript𝐹2F\_{2}italic\_F start\_POSTSUBSCRIPT 2 end\_POSTSUBSCRIPT
The population-level module targets learning the latent population graph. The output of F1subscript𝐹1F\_{1}italic\_F start\_POSTSUBSCRIPT 1 end\_POSTSUBSCRIPT i.e 𝐡=[𝐡𝟎,…𝐡𝐢,…,𝐡𝐍]𝐡subscript𝐡0…subscript𝐡𝐢…subscript𝐡𝐍\mathbf{h}=[\mathbf{h\_{0}},...\mathbf{h\_{i}},...,\mathbf{h\_{N}}]bold\_h = [ bold\_h start\_POSTSUBSCRIPT bold\_0 end\_POSTSUBSCRIPT , … bold\_h start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT , … , bold\_h start\_POSTSUBSCRIPT bold\_N end\_POSTSUBSCRIPT ] is fed to F2subscript𝐹2F\_{2}italic\_F start\_POSTSUBSCRIPT 2 end\_POSTSUBSCRIPT as input.
Each ithsuperscript𝑖𝑡ℎi^{th}italic\_i start\_POSTSUPERSCRIPT italic\_t italic\_h end\_POSTSUPERSCRIPT node in the population level graph corresponds to the ithsuperscript𝑖𝑡ℎi^{th}italic\_i start\_POSTSUPERSCRIPT italic\_t italic\_h end\_POSTSUPERSCRIPT input graph of the previous module, and it is represented by its representation 𝐡𝐢subscript𝐡𝐢\mathbf{h\_{i}}bold\_h start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT. Formally, F2subscript𝐹2F\_{2}italic\_F start\_POSTSUBSCRIPT 2 end\_POSTSUBSCRIPT is a function defined as 𝐀𝐩=F2(𝐡)subscript𝐀𝐩subscript𝐹2𝐡\mathbf{A\_{p}}=F\_{2}(\mathbf{h})bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT = italic\_F start\_POSTSUBSCRIPT 2 end\_POSTSUBSCRIPT ( bold\_h ), where 𝐀𝐩∈(0,1)N×Nsubscript𝐀𝐩superscript01𝑁𝑁\mathbf{A\_{p}}\in\left(0,1\right)^{N\times N}bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT ∈ ( 0 , 1 ) start\_POSTSUPERSCRIPT italic\_N × italic\_N end\_POSTSUPERSCRIPT is the weighted adjacency matrix encoding the population-level graph.
Inspired by the population graph learning strategy proposed in [[11](#bib.bib11)], we develop our LGL technique to let our model learn the population-level graph in a end-to-end fashion together with minimizing the downstream task loss.
We learn the population-level graph by embedding each input graph representation 𝐡𝐢subscript𝐡𝐢\mathbf{h\_{i}}bold\_h start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT in a lower dimensional latent space in which the closeness of of two node features indicates the existence of an edge between the two nodes in the population graph.
To project the input graphs to this latent space we define a learnable function g𝑔gitalic\_g which takes as input the original graph representations 𝐡𝐢subscript𝐡𝐢\mathbf{h\_{i}}bold\_h start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT and outputs a new representation 𝐡𝐢~~subscript𝐡𝐢\tilde{\mathbf{h\_{i}}}over~ start\_ARG bold\_h start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT end\_ARG belonging to this new latent space, 𝐡𝐢~=g(𝐡𝐢)~subscript𝐡𝐢𝑔subscript𝐡𝐢\tilde{\mathbf{h\_{i}}}=g(\mathbf{h\_{i}})over~ start\_ARG bold\_h start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT end\_ARG = italic\_g ( bold\_h start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT ). In our case, we use an MLP as our g𝑔gitalic\_g function. The output of this module is then a weighted adjacency matrix 𝐀𝐩subscript𝐀𝐩\mathbf{A\_{p}}bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT built from input graph representations 𝐡~~𝐡\tilde{\mathbf{h}}over~ start\_ARG bold\_h end\_ARG. Precisely, the edge aijsubscript𝑎𝑖𝑗a\_{ij}italic\_a start\_POSTSUBSCRIPT italic\_i italic\_j end\_POSTSUBSCRIPT between the nodes i𝑖iitalic\_i and j𝑗jitalic\_j in 𝐀𝐩subscript𝐀𝐩\mathbf{A\_{p}}bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT is defined as:
| | | | |
| --- | --- | --- | --- |
| | aij=11+e−t∥𝐡𝐢~−𝐡𝐣~∥2+θ,subscript𝑎𝑖𝑗11superscript𝑒𝑡subscriptdelimited-∥∥~subscript𝐡𝐢~subscript𝐡𝐣2𝜃a\_{ij}=\frac{1}{1+e^{-t\left\lVert\tilde{\mathbf{h\_{i}}}-\tilde{\mathbf{h\_{j}}}\right\rVert\_{2}+\theta}}\;,italic\_a start\_POSTSUBSCRIPT italic\_i italic\_j end\_POSTSUBSCRIPT = divide start\_ARG 1 end\_ARG start\_ARG 1 + italic\_e start\_POSTSUPERSCRIPT - italic\_t ∥ over~ start\_ARG bold\_h start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT end\_ARG - over~ start\_ARG bold\_h start\_POSTSUBSCRIPT bold\_j end\_POSTSUBSCRIPT end\_ARG ∥ start\_POSTSUBSCRIPT 2 end\_POSTSUBSCRIPT + italic\_θ end\_POSTSUPERSCRIPT end\_ARG , | | (1) |
where θ𝜃\thetaitalic\_θ and t𝑡titalic\_t are learnable soft-threshold and temperature parameters.
####
3.1.3 GNN classifier F3subscript𝐹3F\_{3}italic\_F start\_POSTSUBSCRIPT 3 end\_POSTSUBSCRIPT
The output of the population-level module 𝐀𝐩subscript𝐀𝐩\mathbf{A\_{p}}bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT is exploited by the final module (i.e. the GNN classifier) to provide the final prediction according to the downstream task. This later module is composed by a function 𝐩=F3(𝐡,𝐀𝐩)𝐩subscript𝐹3𝐡subscript𝐀𝐩\mathbf{p}=F\_{3}(\mathbf{h},\mathbf{A\_{p}})bold\_p = italic\_F start\_POSTSUBSCRIPT 3 end\_POSTSUBSCRIPT ( bold\_h , bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT ), with 𝐩=[𝐩𝟏,…,𝐩𝐍]𝐩subscript𝐩1…subscript𝐩𝐍\mathbf{p}=[\mathbf{p\_{1}},\dots,\mathbf{p\_{N}}]bold\_p = [ bold\_p start\_POSTSUBSCRIPT bold\_1 end\_POSTSUBSCRIPT , … , bold\_p start\_POSTSUBSCRIPT bold\_N end\_POSTSUBSCRIPT ], that takes as input the individual graph representations 𝐡𝐡\mathbf{h}bold\_h and the learned population-graph 𝐀𝐩subscript𝐀𝐩\mathbf{A\_{p}}bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT and outputs a class probability vector 𝐩𝐢subscript𝐩𝐢\mathbf{p\_{i}}bold\_p start\_POSTSUBSCRIPT bold\_i end\_POSTSUBSCRIPT for each of the input graphs. Specifically, F3subscript𝐹3F\_{3}italic\_F start\_POSTSUBSCRIPT 3 end\_POSTSUBSCRIPT is composed of a set of GNN layers followed by node pooling and a fully-connected layer to obtain the output probabilities. F3subscript𝐹3F\_{3}italic\_F start\_POSTSUBSCRIPT 3 end\_POSTSUBSCRIPT being a graph convolution based module is capable of leveraging input sample similarities and obtain better representations for the downstream task.
(a) LGL
(b) LGL+KL
Figure 2: Distribution of adjacency matrix 𝐀𝐩subscript𝐀𝐩\mathbf{A\_{p}}bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT values for PROTEINS2929{}\_{29}start\_FLOATSUBSCRIPT 29 end\_FLOATSUBSCRIPT dataset with (b) and without (a) the degree distribution loss KLloss𝐾subscript𝐿𝑙𝑜𝑠𝑠KL\_{loss}italic\_K italic\_L start\_POSTSUBSCRIPT italic\_l italic\_o italic\_s italic\_s end\_POSTSUBSCRIPT. The x-axis indicates values of 𝐀𝐩subscript𝐀𝐩\mathbf{A\_{p}}bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT, y-axis shows the values occurrence probability.
###
3.2 Degree distribution loss
In this section we propose a customised loss capable of not only achieving the down stream task but also ensuring the quality of population graph. We experimentally observed that optimizing the population-level module just using the task classification loss tends to produce sparser graphs. This behavior can be seen from the left bar plot in Fig. [2](#S3.F2 "Figure 2 ‣ 3.1.3 GNN classifier 𝐹₃ ‣ 3.1 Graph-in-Graph model ‣ 3 Method ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications") showing the values distribution of 𝐀𝐩subscript𝐀𝐩\mathbf{A\_{p}}bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT optimizing with the plain cross-entropy loss. As a consequence, the resulting graph is composed by many disconnected components and isolated nodes (Fig. [7](#S4.F7 "Figure 7 ‣ 4.4 Knowledge discovery analysis ‣ 4 Experiments and results ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications"), [4](#S4.F4 "Figure 4 ‣ 4.3 Quantitative results ‣ 4 Experiments and results ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications")) which makes it difficult to reason about the predicted population graph structure.
To mitigate this problem we propose to add a Degree Distribution Loss as a regularizer which helps produce more meaningful population graphs.
The population-level adjacency matrix 𝐀𝐩subscript𝐀𝐩\mathbf{A\_{p}}bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT is regularised by adding a prior to the desired node degrees distribution. The desired degree distribution is promoted by an additional Kullback–Leibler divergence loss (KL loss) between the computed degree distribution over 𝐀𝐩subscript𝐀𝐩\mathbf{A\_{p}}bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT and the target distribution.
In this work, we choose a Gaussian distribution as a target distribution, centered at the desired average node degree. This enforces a sparse graph, while leaving the possibility of a few nodes having dense connections. We let the parameters (mean, standard deviation) of the Gaussian target distribution to be optimized during the training, in order to give the model more flexibility in adapting the population graph for optimal mean and standard deviation of node degrees.
On a symmetric adjacency matrix 𝐀𝐀\mathbf{A}bold\_A, the degree of the ithsuperscript𝑖𝑡ℎi^{th}italic\_i start\_POSTSUPERSCRIPT italic\_t italic\_h end\_POSTSUPERSCRIPT node can be computed as the sum of the entries of the ithsuperscript𝑖𝑡ℎi^{th}italic\_i start\_POSTSUPERSCRIPT italic\_t italic\_h end\_POSTSUPERSCRIPT row. Since 𝐀𝐩subscript𝐀𝐩\mathbf{A\_{p}}bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT represents a weighted fully-connected graph, the actual degree of each node is always N−1𝑁1N-1italic\_N - 1. To overcome this limitation, we exploit the fact that ([1](#S3.E1 "1 ‣ 3.1.2 Population-level module 𝐹₂ ‣ 3.1 Graph-in-Graph model ‣ 3 Method ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications")) has the effect to push values of 𝐀𝐩subscript𝐀𝐩\mathbf{A\_{p}}bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT either toward 0 or 1 and consider as actual edges of the graph only the edges with a weight greater that 0.5:
| | | | |
| --- | --- | --- | --- |
| | 𝐀¯=𝐀𝐩\*(𝐀𝐩>0.5)¯𝐀subscript𝐀𝐩subscript𝐀𝐩0.5\mathbf{\bar{A}}=\mathbf{A\_{p}}\*(\mathbf{A\_{p}}>0.5)over¯ start\_ARG bold\_A end\_ARG = bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT \* ( bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT > 0.5 ) | | (2) |
where 𝐀¯∈ℝN×N¯𝐀superscriptℝ𝑁𝑁\mathbf{\bar{A}}\in\mathbb{R}^{N\times N}over¯ start\_ARG bold\_A end\_ARG ∈ blackboard\_R start\_POSTSUPERSCRIPT italic\_N × italic\_N end\_POSTSUPERSCRIPT is the adjacency matrix used to compute the node degree of each node as
| | | | |
| --- | --- | --- | --- |
| | d¯j=∑i=1N𝐀¯i,jsubscript¯𝑑𝑗superscriptsubscript𝑖1𝑁subscript¯𝐀𝑖𝑗\bar{d}\_{j}=\sum\_{i=1}^{N}\mathbf{\bar{A}}\_{i,j}over¯ start\_ARG italic\_d end\_ARG start\_POSTSUBSCRIPT italic\_j end\_POSTSUBSCRIPT = ∑ start\_POSTSUBSCRIPT italic\_i = 1 end\_POSTSUBSCRIPT start\_POSTSUPERSCRIPT italic\_N end\_POSTSUPERSCRIPT over¯ start\_ARG bold\_A end\_ARG start\_POSTSUBSCRIPT italic\_i , italic\_j end\_POSTSUBSCRIPT | | (3) |
with d¯j∈ℝNsubscript¯𝑑𝑗superscriptℝ𝑁\bar{d}\_{j}\in\mathbb{R}^{N}over¯ start\_ARG italic\_d end\_ARG start\_POSTSUBSCRIPT italic\_j end\_POSTSUBSCRIPT ∈ blackboard\_R start\_POSTSUPERSCRIPT italic\_N end\_POSTSUPERSCRIPT. Since d¯jsubscript¯𝑑𝑗\bar{d}\_{j}over¯ start\_ARG italic\_d end\_ARG start\_POSTSUBSCRIPT italic\_j end\_POSTSUBSCRIPT can assume a continuous value between 00 and N−1𝑁1N-1italic\_N - 1, to compute the discrete degree distribution we perform a soft assignment between djsubscript𝑑𝑗d\_{j}italic\_d start\_POSTSUBSCRIPT italic\_j end\_POSTSUBSCRIPT and the discrete node degrees value ci∈[0,1,…,N−1]subscript𝑐𝑖01…𝑁1c\_{i}\in[0,1,\dots,N-1]italic\_c start\_POSTSUBSCRIPT italic\_i end\_POSTSUBSCRIPT ∈ [ 0 , 1 , … , italic\_N - 1 ] as:
| | | | |
| --- | --- | --- | --- |
| | Si,j=e(ci−dj¯)2σ2∑ke(ck−dj¯)2σ2subscript𝑆𝑖𝑗superscript𝑒superscriptsubscript𝑐𝑖¯subscript𝑑𝑗2superscript𝜎2subscript𝑘superscript𝑒superscriptsubscript𝑐𝑘¯subscript𝑑𝑗2superscript𝜎2S\_{i,j}=\frac{e^{\frac{(c\_{i}-\bar{d\_{j}})^{2}}{\sigma^{2}}}}{\sum\_{k}e^{\frac{(c\_{k}-\bar{d\_{j}})^{2}}{\sigma^{2}}}}italic\_S start\_POSTSUBSCRIPT italic\_i , italic\_j end\_POSTSUBSCRIPT = divide start\_ARG italic\_e start\_POSTSUPERSCRIPT divide start\_ARG ( italic\_c start\_POSTSUBSCRIPT italic\_i end\_POSTSUBSCRIPT - over¯ start\_ARG italic\_d start\_POSTSUBSCRIPT italic\_j end\_POSTSUBSCRIPT end\_ARG ) start\_POSTSUPERSCRIPT 2 end\_POSTSUPERSCRIPT end\_ARG start\_ARG italic\_σ start\_POSTSUPERSCRIPT 2 end\_POSTSUPERSCRIPT end\_ARG end\_POSTSUPERSCRIPT end\_ARG start\_ARG ∑ start\_POSTSUBSCRIPT italic\_k end\_POSTSUBSCRIPT italic\_e start\_POSTSUPERSCRIPT divide start\_ARG ( italic\_c start\_POSTSUBSCRIPT italic\_k end\_POSTSUBSCRIPT - over¯ start\_ARG italic\_d start\_POSTSUBSCRIPT italic\_j end\_POSTSUBSCRIPT end\_ARG ) start\_POSTSUPERSCRIPT 2 end\_POSTSUPERSCRIPT end\_ARG start\_ARG italic\_σ start\_POSTSUPERSCRIPT 2 end\_POSTSUPERSCRIPT end\_ARG end\_POSTSUPERSCRIPT end\_ARG | | (4) |
where σ𝜎\sigmaitalic\_σ is a hyperparameter that we experimentally set to 0.6. The value corresponding to the degree cisubscript𝑐𝑖c\_{i}italic\_c start\_POSTSUBSCRIPT italic\_i end\_POSTSUBSCRIPT in the computed node degree distribution is:
| | | | |
| --- | --- | --- | --- |
| | pi=∑jSi,j∑k,jSk,jsubscript𝑝𝑖subscript𝑗subscript𝑆𝑖𝑗subscript𝑘𝑗subscript𝑆𝑘𝑗p\_{i}=\frac{\sum\_{j}S\_{i,j}}{\sum\_{k,j}S\_{k,j}}italic\_p start\_POSTSUBSCRIPT italic\_i end\_POSTSUBSCRIPT = divide start\_ARG ∑ start\_POSTSUBSCRIPT italic\_j end\_POSTSUBSCRIPT italic\_S start\_POSTSUBSCRIPT italic\_i , italic\_j end\_POSTSUBSCRIPT end\_ARG start\_ARG ∑ start\_POSTSUBSCRIPT italic\_k , italic\_j end\_POSTSUBSCRIPT italic\_S start\_POSTSUBSCRIPT italic\_k , italic\_j end\_POSTSUBSCRIPT end\_ARG | | (5) |
Finally, the KLloss𝐾subscript𝐿𝑙𝑜𝑠𝑠KL\_{loss}italic\_K italic\_L start\_POSTSUBSCRIPT italic\_l italic\_o italic\_s italic\_s end\_POSTSUBSCRIPT term is defined as
| | | | |
| --- | --- | --- | --- |
| | KLloss=DKL(p,q)𝐾subscript𝐿𝑙𝑜𝑠𝑠subscript𝐷𝐾𝐿𝑝𝑞KL\_{loss}=D\_{KL}(p,q)italic\_K italic\_L start\_POSTSUBSCRIPT italic\_l italic\_o italic\_s italic\_s end\_POSTSUBSCRIPT = italic\_D start\_POSTSUBSCRIPT italic\_K italic\_L end\_POSTSUBSCRIPT ( italic\_p , italic\_q ) | | (6) |
where q𝑞qitalic\_q is the target normal discrete distribution with learnable parameters. The KL-loss is thus added to the Cross-Entropy loss used to train the classifier as a penalty term weighted by α𝛼\alphaitalic\_α:
| | | | |
| --- | --- | --- | --- |
| | loss=CEloss+αKLloss𝑙𝑜𝑠𝑠𝐶subscript𝐸𝑙𝑜𝑠𝑠𝛼𝐾subscript𝐿𝑙𝑜𝑠𝑠loss=CE\_{loss}+\alpha KL\_{loss}italic\_l italic\_o italic\_s italic\_s = italic\_C italic\_E start\_POSTSUBSCRIPT italic\_l italic\_o italic\_s italic\_s end\_POSTSUBSCRIPT + italic\_α italic\_K italic\_L start\_POSTSUBSCRIPT italic\_l italic\_o italic\_s italic\_s end\_POSTSUBSCRIPT | | (7) |
Table 1: Comparison of dataset with several baselines and SOTA methods. ElasticNet [[54](#bib.bib54)], HGP-SL [[24](#bib.bib24)], ToxicBlend [[32](#bib.bib32)] SOTA methods for HCP, PROTEINS and Tox21 respectively. The top three performance is shown in bold face, blue and red.
| | Without population graph | Non-learned population graph | Learned population graph |
| --- | --- | --- | --- |
| | SOTA | GCN | Random | KNN | GiG DGCNN | GiG LGL | GiG LGL+KL |
| HCP (acc %) | 85.5 ± NA | 84.2 ± 2.5 | 43.7 ± 8.4 | 39.6 ± 0.0 | 89.0 ± 3.3 | 89.4 ± 1.3 | 89.7 ± 1.3 |
| PROTEINS2929{}\_{29}start\_FLOATSUBSCRIPT 29 end\_FLOATSUBSCRIPT (acc %) | 84.9 ± 1.6 | 80.0 ± 2.5 | 53.7 ± 8.4 | 74.9 ± 2.7 | 72.4 ± 2.2 | 82.0 ± 1.6 | 84.8 ± 1.1 |
| Tox21 (acc %) | 80.7 ± NA | 75.0 ± 0.8 | 80.1 ± 3.9 | 80.1 ± 2.0 | 85.1 ± 2.1 | 84.8 ± 1.7 | 82.3 ± 2.3 |
4 Experiments and results
--------------------------
In this section, we show experiments on biological, medical and chemical applications where each input sample is represented as a graph.
In particular we choose PROTEINS [[55](#bib.bib55)], HCP [[26](#bib.bib26)] and Tox21 [[56](#bib.bib56)]. Further, we make use of the publicly available benchmark datasets D&D [[16](#bib.bib16)], NCI1 [[57](#bib.bib57)] and ENZYMES [[58](#bib.bib58)] to compare with other general purpose graph classification methods. These datasets are commonly used in the evaluation of graph classification models for binary and multi-class classification. In all the benchmark datasets we follow the training/testing procedure described in [[59](#bib.bib59)] for a fair comparision.
We show two variants of the GiG model. In the first variant (LGL) the population-level graph is learned end-to-end using just the cross-entropy loss for the final classification task as in [[11](#bib.bib11)]. In the second variant (LGL+KL) the model incorporates the proposed node degree distribution loss ([3.2](#S3.SS2 "3.2 Degree distribution loss ‣ 3 Method ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications")) along with learning the population-level graph.
In order to better validate the benefit of the population level module, we also investigate different possibilities to build or learn the population graph.
###
4.1 Datasets
In this section, we give the details of the datasets used for our experimental evaluation.
HCP.
The Human Connectome Project (HCP) is a medical dataset consisting of 1003 complete resting-state functional MRI (fMRI) runs released in 2017 [[26](#bib.bib26)]. The task is to predict the gender of each patient based on the corresponding brain fMRI data.
The brain graphs are provided within the dataset as the correlation matrix between fMRI signals from each parcels. Each brain consist of 200 parcels, and hence comes with an affinity matrix of size 200×200200200200\times 200200 × 200.
PROTEINS.
PROTEINS is a binary classification dataset [[18](#bib.bib18)] implemented by the TU Dortmund University [[55](#bib.bib55)]. Here, the task is to classify the protein as an enzyme or non-enzyme.
We used two versions of the PROTEINS dataset: PROTEINS2929{}\_{29}start\_FLOATSUBSCRIPT 29 end\_FLOATSUBSCRIPT and PROTEINS33{}\_{3}start\_FLOATSUBSCRIPT 3 end\_FLOATSUBSCRIPT, where the index indicates the number of features used in the feature matrix.
For population graph analyses, we use the CATH Superfamily information [[60](#bib.bib60), [61](#bib.bib61)] for each protein, which was retrieved from [[62](#bib.bib62)]. Importantly, CATH labels were not used during the training stage. Instead, we used CATH labels only to investigate and discuss potential knowledge discovery of population-level graphs.
The CATH hierarchy consists of several hierarchical class levels. We consider only the first class level (C-level), consisting of three major classes: ”mainly-alpha”, ”mainly-beta”, and ”alpha-beta”.
Tox21. Tox21 is a public dataset [[56](#bib.bib56)] consisting of 8014 toxicity measurements on 7 nuclear receptor signals and 5 stress response panels [[56](#bib.bib56)]. Each of the 12 targets might be equal to ”1” or ”0”, representing the ”active” or ”not active” state. Then, the task for this dataset is multi-task binary classification.
Figure 3: PROTEINS33{}\_{3}start\_FLOATSUBSCRIPT 3 end\_FLOATSUBSCRIPT. The x-axis shows batch size, y-axis indicates accuracy. Triangles indicate mean accuracy and tails standard deviation.
###
4.2 Implementation details
All the experiments are trained with the Adam optimization [[63](#bib.bib63)].
For each dataset number of layers, learning rate, batch size, alpha in ([7](#S3.E7 "7 ‣ 3.2 Degree distribution loss ‣ 3 Method ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications")), and the global pooling operator (mean or add) are tuned based on the validation loss.
In Table [1](#S3.T1 "Table 1 ‣ 3.2 Degree distribution loss ‣ 3 Method ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications"), applications are trained with GraphConv [[64](#bib.bib64)] layers in the node-level module. In Table [2](#S4.T2 "Table 2 ‣ 4.3 Quantitative results ‣ 4 Experiments and results ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications"), we used GIN [[19](#bib.bib19)] to show the flexibility of GiG models and its possibility to adapt to the input data.
For classification losses, optimizer, node-level layers and pooling operators, we used the versions implemented in PyTorch [[65](#bib.bib65)] and PyTorch Geometric [[66](#bib.bib66)].
Dataset split.
The HCP dataset is split into training (72%), validation (8%), and test (20%) sets [[25](#bib.bib25)]. Results are averaged over five consecutive runs of the best model. For PROTEINS2929{}\_{29}start\_FLOATSUBSCRIPT 29 end\_FLOATSUBSCRIPT, we followed [[24](#bib.bib24)] and randomly split the dataset for train and test sets with 90%, 10% and repeat it 10 times. For each run the test set is fixed, then the training set is split to train and validation sets by the k-folds split, where k=10. The final evaluation was based on 10 runs. For the Tox21 dataset, we use predefined scaffold splits from [[67](#bib.bib67)]. The same sets were used for all suggested GiG models and baselines. For fair comparison, D&D, NCI1, ENZYMES, and PROTEINS33{}\_{3}start\_FLOATSUBSCRIPT 3 end\_FLOATSUBSCRIPT we use the same test protocol as in [[59](#bib.bib59)].
Importance of batch size.
Our method does not require the full population to be known a-priori, but rather builds the population graph on a subset of samples according to the extracted batch. This has the advantage of both reducing the computational complexity and allowing our method to be used in the more challenging inductive setting, where part of the population is unknown at training time.
As a consequence, the number of input graphs that we consider in a batch has a direct effect on the performance of our method.
To investigate this aspect, we fixed the batch size to a specific value and optimized the remaining hyperparameters according the validation loss.
We report in Fig. [3](#S4.F3 "Figure 3 ‣ 4.1 Datasets ‣ 4 Experiments and results ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications") results obtained on PROTEINS33{}\_{3}start\_FLOATSUBSCRIPT 3 end\_FLOATSUBSCRIPT dataset using our LGL+KL model. Small triangles indicate mean accuracy and tails show the standard deviation. From the plot, we can see that considering more graphs in the population has a benefit up to around 50 samples, after which the performance stabilizes.
###
4.3 Quantitative results
In this subsection, we show experiments on three main and four benchmark datasets.
To show the benefits of adding the population level module we compared our models with a regular graph convolutional baseline (GCN in Table [1](#S3.T1 "Table 1 ‣ 3.2 Degree distribution loss ‣ 3 Method ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications")) and the state-of-the-art method for each application: HGP-SL [[24](#bib.bib24)] for proteins, ElasticNet [[54](#bib.bib54)] for HCP, and ToxicBlend [[32](#bib.bib32)] for Tox21.
Further, we investigate three different baseline methods of building the population graph. First, to show the benefits of learning the graph, we suggest building the population level graph sampling a random graph using Erdős–Rényi model with p=10 (Random).
The second strategy consists on building a KNN graph based on the input graph similarities computed with the WL graph kernel [[68](#bib.bib68)] with k=10 (KNN). Lastly, as an hybrid approach, we adopt a strategy similar to DGCNN [[2](#bib.bib2)] and dynamically build a KNN graph on the output features of the first module F1subscript𝐹1F\_{1}italic\_F start\_POSTSUBSCRIPT 1 end\_POSTSUBSCRIPT (GiG DGCNN). Despite changing during the training process, with this approach the graph structure is not directly learned as in our method.
HCP.
Table [1](#S3.T1 "Table 1 ‣ 3.2 Degree distribution loss ‣ 3 Method ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications") row 1 shows the results for experiments on gender prediction task on HCP dataset. We compare our results with the state of the art ElasticNet model (non-graph-based) [[25](#bib.bib25)].
The proposed models LGL, LGL+KL outperform the SOTA by 3.92% and 4.17 % respectively.
Further, comparing GCN with population-graph-based models (DGCNN, LGL, LGL+KL) we see that a proper population graph helps with the downstream task and boosts the performance of at least 4.78 %. In contrast, a comparison of fixed population graphs with GCN shows that the wrong population graph might significantly decrease the final classification performance.
(a) KNN
(b) GiG DGCNN
(c) LGL
(d) LGL+KL
Figure 4: HCP: population graphs comparison according to misclassified identifiers. The location of the node depends purely on the binary structure of the graph. Misclassified nodes are circled in red.
(a) LGL
(b) LGL
(c) LGL+KL
(d) LGL+KL
Figure 5: PROTEINS2929{}\_{29}start\_FLOATSUBSCRIPT 29 end\_FLOATSUBSCRIPT: population graphs comparison. Misclassified nodes circled by red. The plotting threshold for LGL is equal to 0.01 and for LGL+KL is 0.5 respectively, which means that only edges with weight values bigger than the chosen threshold were plotted. In the first column, yellow represents enzymes (class 0) and violet non-enzymes (class 1). In the second column: yellow, violet and green represent mainly alpha, mainly beta and alpha-beta categories respectively.
Table 2: Comparison with general purpose graph classification methods. Top three performance are in bold face, blue and red.
| | | | | |
| --- | --- | --- | --- | --- |
| | D&D | NCI1 | PROTEINS33{}\_{3}start\_FLOATSUBSCRIPT 3 end\_FLOATSUBSCRIPT | ENZYMES |
| Deep-Graph-CNN | 76.6 ± 4.3 | 76.4 ± 1.7 | 72.9 ± 3.5 | 38.9 ± 5.7 |
| DiffPool | 75.0 ± 3.5 | 76.9 ± 1.9 | 73.7 ± 3.5 | 59.5 ± 5.6 |
| ECC | 72.6 ± 4.1 | 76.2 ± 1.4 | 72.3 ± 3.4 | 29.5 ± 8.2 |
| GIN | 75.3 ± 2.9 | 80.0 ± 1.4 | 73.3 ± 4.0 | 59.6 ± 4.5 |
| GraphSAGE | 72.9 ± 2.0 | 76.0 ± 1.8 | 73.0 ± 4.5 | 58.2 ± 6.0 |
| GiG DGCNN | 70.4 ± 5.9 | 79.1 ± 2.2 | 70.8 ± 4.0 | 30.3 ± 8.7 |
| GiG LGL | 74.9 ± 3.9 | 81.0 ± 1.9 | 75.0 ± 2.7 | 54.5 ± 5.6 |
| GiG LGL+KL | 76.7 ± 4.9 | 81.8 ± 1.8 | 75.6 ± 3.9 | 50.0 ± 7.0 |
PROTEINS2929{}\_{29}start\_FLOATSUBSCRIPT 29 end\_FLOATSUBSCRIPT.
In the PROTEINS2929{}\_{29}start\_FLOATSUBSCRIPT 29 end\_FLOATSUBSCRIPT dataset (Table [1](#S3.T1 "Table 1 ‣ 3.2 Degree distribution loss ‣ 3 Method ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications") row 2), the LGL+KL model performs very close to the SOTA model HGP-SL [[24](#bib.bib24)]. The difference in accuracy is 0.13 %-points, however, LGL+KL is more stable among the folds and the final standard error is 50% lower in comparison with HGP-SL. LGL and LGL+KL exceed the GCN by 2 and 4.8 %-points. This indicates that among graph-based approaches, learning the population-level graph can help with the downstream task. Among all suggested models, LGL+KL gives the best results (84.8 ± 1.08 ), which is higher than LGL by 2.8 %-points. Regularizing node degrees with the KL-loss stabilizes the learning process and boosts the classification performance.
Tox21.
For the Tox21 dataset (Table [1](#S3.T1 "Table 1 ‣ 3.2 Degree distribution loss ‣ 3 Method ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications") row 3), Random and KNN baselines perform similarly and are close to the SOTA results on scaffold split [[32](#bib.bib32)]. LGL+KL and LGL outperform ToxicBlend [[32](#bib.bib32)] with a gap of 4.97 and 7.5 %-points. The performance difference between LGL and DGCNN is not significant (0.26 %).
Bench-marking datasets.
The GiG LGL+KL model outperforms the SOTA classification models for D&D, NCI1 and PROTEINS33{}\_{3}start\_FLOATSUBSCRIPT 3 end\_FLOATSUBSCRIPT dataset, with a margin of 0.1, 1.8, 1.9 % respectively (Table [2](#S4.T2 "Table 2 ‣ 4.3 Quantitative results ‣ 4 Experiments and results ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications")). For NCI1 and PROTEINS33{}\_{3}start\_FLOATSUBSCRIPT 3 end\_FLOATSUBSCRIPT datasets LGL is the second best performing model.
(a) PROTEINS33{}\_{3}start\_FLOATSUBSCRIPT 3 end\_FLOATSUBSCRIPT
(b) CATH classes, PROTEINS33{}\_{3}start\_FLOATSUBSCRIPT 3 end\_FLOATSUBSCRIPT
(c) D&D
(d) NCI1
Figure 6: Population graphs comparison of GiG LGL+KL model. Misclassified nodes circled in red. The plotting threshold for LGL+KL is 0.5, which means that only edges with weight values bigger than the chosen threshold were plotted. For two color plots: yellow represents class 0 and violet class 1. CATH hierarchy: yellow, violet and green corresponds to mainly alpha, mainly beta and alpha-beta respectively.
###
4.4 Knowledge discovery analysis
In this section, we show and discuss the population-level graph and how they might be exploited as a source of additional information for data exploration.
(a) LGL
(b) LGL+KL
Figure 7: PROTEINS2929{}\_{29}start\_FLOATSUBSCRIPT 29 end\_FLOATSUBSCRIPT GiG LGL vs GiG LGL+KL nodes degree distribution comparison. The x-axis shows node degrees divided into 50 bins, the y-axis shows the occurrence of nodes degree. Left column indicates normal nodes degree distribution comparison, right column - with thresholds equal to 0.5.
HCP:
As can be seen in Fig. [3(a)](#S4.F3.sf1 "3(a) ‣ Figure 4 ‣ 4.3 Quantitative results ‣ 4 Experiments and results ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications"), the KNN baseline (k=10) performs the worst for HCP dataset. This method misclassifies all nodes with the class ”0” and there is no evidence of clustering, all objects are connected and grouped. In comparison, all population graphs from the proposed GiG methods are well clustered. Importantly, the LGL adjacency matrix 𝐀𝐩subscript𝐀𝐩\mathbf{A\_{p}}bold\_A start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT of the learned population graph entirely consists of near-zero values, indicating that the model is not confident about the learned graph structure. For interpretability analysis, we plotted LGL in Fig. [3(c)](#S4.F3.sf3 "3(c) ‣ Figure 4 ‣ 4.3 Quantitative results ‣ 4 Experiments and results ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications") with an edge weight threshold at a level of 1e−11e11\mathrm{e}{-1}1 roman\_e - 1. In contrast, the KL-loss resulted in edge weights within the LGL+KL population-level graph is well binarized at a level of 0.50.50.50.5.
Fig. [4](#S4.F4 "Figure 4 ‣ 4.3 Quantitative results ‣ 4 Experiments and results ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications") shows the unique numbers of each node as well. DGCNN locates nodes number ”4” and ”1” on the border of two classes and misclassifies them. LGL+KL mislocates the node ”5”, which in the previous population graphs is located inside its cluster. For all methods except KNN, some nodes tend to be located together. For example nodes with identifiers ”11”, ”12” and ”16” as well as ”23”, ”2”, ”14”. This is an indicator that these nodes have some similar properties or common characteristics.
PROTEINS2929{}\_{29}start\_FLOATSUBSCRIPT 29 end\_FLOATSUBSCRIPT:
In Fig. [4(a)](#S4.F4.sf1 "4(a) ‣ Figure 5 ‣ 4.3 Quantitative results ‣ 4 Experiments and results ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications") and Fig. [4(c)](#S4.F4.sf3 "4(c) ‣ Figure 5 ‣ 4.3 Quantitative results ‣ 4 Experiments and results ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications"),
GiG LGL and GiG LGL+KL show a clustering trend. However, similar to the HCP dataset, the LGL model is quite uncertain about the learned graph connection. It can be seen in Fig. [2](#S3.F2 "Figure 2 ‣ 3.1.3 GNN classifier 𝐹₃ ‣ 3.1 Graph-in-Graph model ‣ 3 Method ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications"), where most adjacency matrix values are located near-zero for LGL. In contrast, using the KL Loss the GiG LGL+KL model is more confident regarding its decision (Fig. [1(b)](#S3.F1.sf2 "1(b) ‣ Figure 2 ‣ 3.1.3 GNN classifier 𝐹₃ ‣ 3.1 Graph-in-Graph model ‣ 3 Method ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications")).
The binarized nodes degree distribution is skewed to the left ( Fig. [7](#S4.F7 "Figure 7 ‣ 4.4 Knowledge discovery analysis ‣ 4 Experiments and results ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications")). The nodes degree distribution of the LGL+KL model remains almost the same after the 0.50.50.50.5 thresholding, which also indicates the confidence of the model.
In Fig. [4(b)](#S4.F4.sf2 "4(b) ‣ Figure 5 ‣ 4.3 Quantitative results ‣ 4 Experiments and results ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications"), ([4(d)](#S4.F4.sf4 "4(d) ‣ Figure 5 ‣ 4.3 Quantitative results ‣ 4 Experiments and results ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications")) we see that nodes are clustered according to CATH classes: ”1”, ”2”, ”3”, i.e. ”mainly alpha”, ”mainly beta” and ”alpha-beta” respectively. Remarkably, most nodes that were misclassified according to task labels, were located on the proper CATH class clusters. This indicates that GiG models not only cluster population-level graphs according to training labels but as well capture the structure and properties of the input data. Moreover, the majority of prediction errors were done on the proteins that belong to alpha-beta CATH classes. This is explainable with domain knowledge: ”alpha-beta” protein structures are composed of α𝛼\alphaitalic\_α-helices and β𝛽\betaitalic\_β-strands, which are combined inside the structure of ”mainly alpha” or ”mainly beta” proteins. So it is more challenging for the model to find a proper location inside the population graph for these input objects.
Tox21:
Tox21 is a multi-task dataset, where assays are highly correlated with each other and training models towards only one of the toxic tests is less effective [[69](#bib.bib69)]. However, analyzing the population graph with multiple labels is challenging. Therefore, in this work, we only provide the quantitative results for the Tox21 dataset. From those, it is evident that incorporating the population graph helped in improving the final downstream task performance.
Bench-marking datasets: The population-level graphs from GiG LGL+KL for NCI1, D&D and Proteins33{}\_{3}start\_FLOATSUBSCRIPT 3 end\_FLOATSUBSCRIPT are shown in the Fig. [6](#S4.F6 "Figure 6 ‣ 4.3 Quantitative results ‣ 4 Experiments and results ‣ Graph-in-Graph (GiG): Learning interpretable latent graphs in non-Euclidean domain for biological and healthcare applications"). The misclassification of some nodes is quite expected, given their location in 𝐆𝐩subscript𝐆𝐩\mathbf{G\_{p}}bold\_G start\_POSTSUBSCRIPT bold\_p end\_POSTSUBSCRIPT and obtained via GCN based model node representations. Explaining these locations requires some additional information, like CATH hierarchy for PROTEINS dataset. From PROTEINS33{}\_{3}start\_FLOATSUBSCRIPT 3 end\_FLOATSUBSCRIPT population graphs, it is evident that most misclassified nodes are embedded at locations that correspond to CATH classes. This indicates that population graphs and their node embedding provide a form of knowledge discovery, which corresponds to additional domain information of the input objects.
5 Discussion
-------------
To learn the population graph in LGL-based models, we utilize a soft-threshold which is tuned by two learnable parameters: temp𝑡𝑒𝑚𝑝tempitalic\_t italic\_e italic\_m italic\_p and θ𝜃\thetaitalic\_θ. One future direction could be to learn the θ𝜃\thetaitalic\_θ values per node. The latter hugely affects the adjacency matrix values and might help to push them to either 0 or 1. Even though these parameters are learned, the proper initialization is crucial and might lead to different population graphs and different classification performances. The overall sparsity of the graph depends on the value of the threshold θ𝜃\thetaitalic\_θ.
During our experiments, we also noticed that it is crucial to obtain proper input-graph representations, since the population graph is learned or constructed based on them. Thus the node-level module and pooling operator should be adapted to the application and the input data. Another crucial point is the target distribution with learnable parameters in the GiG LGL+KL model. In this work, we utilized a normal distribution to regularize the latent graph. Thus, directions for future work include the investigation of the influence of different target distributions and their reasoning, as well as considering graph pooling operators. In addition, all constructed and learned latent graphs were obtained based on similarities between input objects, however, one more relevant and challenging application for this work might be to learn the graph that represents Protein-protein interactions (PPI). We would like to point out a few limitation of the proposed work. 1) The choice of the target distribution needs further exploration. In this paper we only considered Gaussian distribution. 2) Even though the model gives us the freedom of choice for F1subscript𝐹1F\_{1}italic\_F start\_POSTSUBSCRIPT 1 end\_POSTSUBSCRIPT and F2subscript𝐹2F\_{2}italic\_F start\_POSTSUBSCRIPT 2 end\_POSTSUBSCRIPT, a careful selection of F1subscript𝐹1F\_{1}italic\_F start\_POSTSUBSCRIPT 1 end\_POSTSUBSCRIPT and F2subscript𝐹2F\_{2}italic\_F start\_POSTSUBSCRIPT 2 end\_POSTSUBSCRIPT is necessary.
6 Conclusion
-------------
In this paper, we propose a method for learning a latent graph structure on graph-valued input data. The latent graph provides a form of knowledge discovery on the distribution and neighborhood relationships between input samples. Further, the latent graph is learned end-to-end, along with the downstream task, e.g. classification. More precisely, we proposed a structure consisting of three main modules: node-level module, population-level module, and GCN classifier that might be easily adapted to any graph classification application. Based on our experiments, we noticed that learning the population graph mainly increases (HCP, Tox21, NCI1, D&D, PROTEINS33{}\_{3}start\_FLOATSUBSCRIPT 3 end\_FLOATSUBSCRIPT datasets) the performance of the downstream task. Additionally, the latent graph might be used for understanding the made decision (PROTEINS) and discovering new information based on interconnected objects. Moreover, we suggested regularising the latent graph’s node degree distribution, in order to obtain a better representation for the graph-valued input population.
|
c7f3bb67-61f6-4c52-9331-3a664f8e0901
|
StampyAI/alignment-research-dataset/arxiv
|
Arxiv
|
Apprenticeship Learning using Inverse Reinforcement Learning and Gradient Methods
Appren ticeship
Learning using Inverse Reinforcement Learning and
Gradient Methods
Gergely Neu¤
Budap
est University of Technology and Economics
M} uegyetem rkp. 3-9.
H-1111 Budapest, HungaryCsaba Szepesv¶ ari¤
Department of Computing Science
University of Alberta
Edmonton T6G 2E8, AB, Canada
Abstract
In this paper we propose a novel gradient al-
gorithm to learn a policy from an expert's
observed behavior assuming that the expert
behaves optimally with respect to some un-
known reward function of a Markovian De-
cision Problem. The algorithm's aim is to
¯nd a reward function such that the resulting
optimal policy matches well the expert's ob-
served behavior. The main di±culty is that
the mapping from the parameters to poli-
cies is both nonsmooth and highly redun-
dant. Resorting to subdi®erentials solves the
¯rst di±culty, while the second one is over-
come by computing natural gradients. We
tested the proposed method in two arti¯cial
domains and found it to be more reliable and
e±cient than some previous methods.
1 INTRODUCTION
The aim of apprenticeship learning is to estimate a
policy of an expert based on samples of the expert's
behavior. This problem has been studied in the ¯eld of
robotics for a long time and due to the lack of space we
cannot give an overview of the literature. The inter-
ested reader might ¯nd a short overview in the paper
by Abbeel and Ng (2004).
In apprenticeship learning (a.k.a. imitation learning)
one can distinguish between direct andindirect ap-
proaches. Direct methods attempt to learn the pol-
icy (as a mapping from states, or features describing
states to actions) by resorting to a supervised learning
method. They do this by optimizing some loss func-
tion that measures the deviation between the expert's
¤Computer and
Automation Research Institute of the
Hungarian Academy of Sciences, Kende u. 13-17, Bu-
dapest 1111, Hungarypolicy and the policy chosen. The main problem then
is that in parts of the state space that the expert tends
to avoid the samples are sparse and hence these meth-
ods may have di±culties with learning a good policy
at such places.
In an indirect method it is assumed that the expert
is acting optimally in the environment. In particular,
in inverse reinforcement learning the environment is
modelled as a Markovian decision problem (MDP) (Ng
and Russell, 2000). The dynamics of the environment
is assumed to be known (or it could be learnt from
samples which might even be unrelated to the samples
come from the expert). However, the reward function
that the expert is using is unknown. Recently Abbeel
and Ng (2004) gave an algorithm which was proven to
produce a policy which performs almost as well as the
expert, even though it is not guaranteed to recover the
expert's reward function (recovering the reward func-
tion is an ill-posed problem). This approach might
work with less data since it makes use of the knowl-
edge of model of the environment, which can help it
in generalizing to the less frequently visited parts of
the state space. One problem is that the algorithm of
Abbeel and Ng (2004) relies on the precise knowledge
of the features describing the reward function, which
is not a realistic assumption (for a discussion of this,
see Section 6). In particular, we will show that even
the correct scales of the features have to be known.
In this paper we propose a gradient algorithm that
combines the two approaches by minimizing a loss
function that penalizes deviations from the expert's
policy like in supervised learning, but the policy is ob-
tained by tuning a reward function and solving the
resulting MDP, instead of ¯nding the parameters of a
policy. We will demonstrate that this combination can
unify the advantages of the two approaches in that it
can be both sample e±cient and work even when the
features are just vaguely known.
NEU & SZEPESVÁRI 295
2BA
CKGROUND
Let us ¯rst introduce some notation: For a subset
Sof some topological space, S±will be used to de-
note its interior. For a ¯nite dimensional vector x,
kxk=Pd
i=1x2
ishall denote its `2-norm. Random vari-
ables will be denoted by capital letters (e.g., X,A),E[¢]
stands for expectations.
We assume that the reader is familiar with ba-
sic concepts underlying Markovian decision processes
(MDPs) (e.g., Puterman 1994), hence we introduce
these concepts only to ¯x the notation. A ¯nite, dis-
counted in¯nite-horizon total reward MDP is de¯ned
by a 5-tuple M= (X;A; °; P; r ), where
Xis a ¯nite set of states,
Ais a ¯nite set of actions,
°2[0;1) is the discount factor,
Pgives the transition probabilities ;P(x0jx; a) stands
for the probability of the transition from state x
tox0upon taking action a(x; x02 X; a2 A),
ris the reward function ;r:X £ A !R;r(x; a)
gives the reward incurred when action a2 A is
executed from state x2 X.
Astationary stochastic policy (in short: policy) is a
mapping ¼:A £ X ! [0;1] satisfyingP
a2A¼(ajx) =
1;8x2 X.1The value of ¼(ajx) is the probability of
taking action ain state x. A policy is called deter-
ministic if for any x,¼(¢jx) is concentrated on a single
action. The class of all stationary stochastic policies
will be denoted by ¦.
For a ¯xed policy, the value of a state x2 Xis de¯ned
by
V¼(x) =E"1X
t=0°tr(Xt; At)¯¯¯¯X0=x#
; (1)
where (X t; At)t¸0is the sequence of random state-
action pairs generated by executing the policy ¼. The
function V¼:X !Ris called the value function un-
derlying policy ¼.
We will also need action-value functions . The action-
value function, Q¼:X £ A !R, underlying policy ¼
is de¯ned by
Q¼(x; a) =E"1X
t=0°tr(Xt; At)¯¯¯¯X0=x; A 0=a#
(2)
1Instead of¼(
a; x) we use ¼(ajx) to emphasize that
¼(¢; x) is a probability distribution. Note that in ¯nite
MDPs one can always ¯nd optimal (stochastic) stationary
policies (Puterman, 1994).with the understanding that for t >0,At»¼(¢jXt).
A policy that maximizes the expected total discounted
reward over all states is called an optimal policy .
The optimal value function is de¯ned by V¤(x) =
sup¼V¼(x), while the optimal action-value function is
de¯ned by Q¤(x; a) = sup¼Q¼(x; a).
It turns out that V¤andQ¤satisfy the so-called Bell-
man optimality equations (e.g., Puterman 1994). In
particular,
Q¤(x; a) = r(x; a) + °X
y2XP(yjx; a) max
b2AQ¤(x; b): (3)
We call a policy that satis¯esP
a2A¼(ajx)Q(x; a) =
max a2AQ(x; a) at all states x2 X greedy w.r.t. the
function Q. It is known that all policies that are greedy
w.r.t. Q¤are optimal and all stationary optimal poli-
cies can be obtained these way.
3 APPRENTICESHIP LEARNING
Assume that we observe a sequence of state-action
pairs ( Xt; At)0·t·T, the `trace' of some expert. We
assume that the expert selects the actions by some
unknown policy ¼E:At»¼E(¢jXt). The goal is to re-
cover ¼Efrom the observed trace. The simplest solu-
tion is of course to use a supervised learning approach:
we select a parametric class of policies, ( ¼µ)µ,¼µ2¦,
µ2Rd, and try to tune the parameters so as to mini-
mize some loss JT(¼µ), such as
JT(¼) =X
x2X;a2A^¹T(x)(¼(ajx)¡^¼E;T(ajx))2;(4)
where ^ ¹T(x) could be de¯ned by ^ ¹T(x) = 1=(T +
1)PT
t=0IfXt=xg are the empirical occupation fre-
quencies under the expert's policy and ^ ¼E;T(ajx) =PT
t=0IfXt=x;At=ag=PT
t=0IfXt=xgis the empirical es-
timate of the expert's policy.2It is easy to see that JT
approximates the squared loss
J(¼) =X
x2X;a2A¹E(x)(¼(ajx)¡¼E(ajx))2(5)
uniformly in ¼(the usual concentration results hold
forJT, e.g. GyÄ or¯ et al. (2002)).
The reason ^ ¼E;Tis not used directly as a `solution' is
that if the state space is large then it will be unde-
¯ned for a large number of states (where ^ ¹E;T(x) = 0)
with high probability unless the number of samples is
enormous.
An alternative to direct policy learning is inverse re-
inforcement learning (Ng and Russell, 2000). The idea
2Ifa
state is not visited by the expert, the policy is
de¯ned arbitrarily.
NEU & SZEPESVÁRI 296
isthat
given the expert's trace, we ¯nd a reward func-
tion that can be used to explain the performance of
the expert. More precisely, the problem is to ¯nd a
reward function that the behavior of the expert is op-
timal for. Once the reward function is found, existing
algorithms are used to ¯nd a behavior that is optimal
with respect to it.
One di±culty in IRL is that solutions are non-unique:
e.g. if ris a reward function that recovers the expert's
policy then for any ¸¸0,¸ris also a solution ( r= 0
is always a solution). For non-trivial problems there
are many solutions besides the variants that di®er in
their scale only.
We propose here to unify the advantages of the di-
rect and indirect approaches by (i)taking it seriously
that we would like to recover the expert's policy and
(ii)achieve this through IRL so that we can achieve
good generalization at parts of the state space avoided
by the expert. We thus propose to ¯nd the parame-
ters given a parametric family of rewards (r µ)µ2£
such that the corresponding (near) optimal policy, ¼µ,
matches the expert's policy ¼E(more precisely, it's
empirical estimate). The proposed method can be
written succinctly as the optimization problem
J(¼µ)µ!min! s:t: ¼ µ=G(Q¤
µ); (6)
where Jis a loss function (such as (5) or (4)) aimed at
measuring the distance of ¼Eand its argument, Q¤
µis
the optimal action-value function corresponding to the
reward function rµandGis a suitable smooth mapping
that returns (near) greedy policies with respect to its
argument. One possibility, utilized in our experiments,
is to use Boltzmann action-selection policies (see (7)).3
In this paper we consider gradient methods to solve
the above optimization problem. One di±culty with
such an approach is that there could be many param-
eterizations that yield to the same loss. This will be
helped with the method of natural gradients, for which
the theory is worked out in the next section.
Another di±culty is that the mapping µ7!Q¤
µis non-
di®erentiable.We will, however, show that it is Lips-
chitz when rµis Lipschitz and hence, by Rademacher's
theorem it is di®erentiable almost everywhere (w.r.t.
the Lebesgue measure).
3.1 NATURAL GRADIENTS
Our ultimate goal is to ¯nd some parameters µin a
parameter space £ ½Rdsuch that the policy ¼µde-
termined by µmatches the expert's policy ¼E. For
3The b
ene¯t of choosing strictly stochastic policies is
that if the expert's policy is deterministic, they force the
uniqueness of the solution.facilitating the discussion let us denote the map from
the parameter space £ to the policy space by h(i.e.,
h(µ) =¼µ). Thus, our objective function can be writ-
ten as ~J(µ) =J(h(µ)), where J: ¦!Ris a (di®eren-
tiable) objective function de¯ned over ¦ (such as (5))
and the goal is to minimize ~J. Incremental gradient
methods implement µt+1=µt¡®tgt, where ®t¸0
is an appropriate step-size sequence and gt=g(µ)
points in the direction of steepest ascent on the surface
(µ;~J(µ))µ.
The gradient method with an in¯nitesimal step-size
gives rise to a trajectory ( µ(t))t¸0. This in turn deter-
mines a trajectory (¼ (t))t¸0in the policy space, where
¼(t) =h(µ(t)). Since our primary interest is the tra-
jectory in the policy state, it makes sense to determine
the gradient direction gin each step such that ¼(t)
moves in the steepest descent direction on the surface
of (¼; J (¼))¼. We call g=g(µ) the natural gradient if
this holds. Amari (1998) gives a method to ¯nd the
natural gradients using the formalism of Riemannian
spaces.
The advantage of this procedure is that the resulting
trajectories will be the same for any equivalent pa-
rameterization (i.e., if the parameter space is replaced
by some other space that is related to the ¯rst one
through a smooth invertible mapping, with a smooth
inverse). In addition, the gradient algorithm that uses
natural gradients can be proven to be asymptotically
e±cient in a probabilistic sense and has the tendency
to alleviate the problem of `plateaus' (Amari, 1998).
In order to de¯ne natural gradients we need some def-
initions. First, we need the generalization of deriva-
tives for mappings fbetween Banach spaces.4The
underlying idea is that the gradient (derivative) of
f:U!Vprovides a linear approximation to the
change f(u+h)¡f(u):
De¯nition 1 (Fr¶ echet derivative) .LetU; V be Banach
spaces. Ais the Fr¶ echet-derivative offatuifA:
U!Vis a bounded linear operator and kf(u+h)¡
f(u)¡Ahk V=o(khk U):The mapping fthen is called
Fr¶ echet di®erentiable atu.
In what follows we view ¦ both as a vector space and
a complete metric space with some metric d. In our
application this metirc will be derived from the (un-
weighted) `2-norm, but other choices would also work.
The following de¯nition suggests a geometry induced
on £:
De¯nition 2 (Induced metric). Let£½Rd,µ2£±.
We say that Gµ2Rd£dis a pseudo-metric induced by
4ABanac
h space is a complete normed vector space. In
our case it will usually be a Euclidean space, e.g. Rd.
NEU & SZEPESVÁRI 297
(h;¦;d
)atµifGµis positive semide¯nite and
d(h(µ+ ¢) ; h(µ)) = ¢TGµ¢ +o(k¢k2):
The essence of this de¯nition is that if the `distance'
between µandµ+ ¢ is given by ¢TGµ¢ then this
distance will match the distance of h(µ) and h(µ+ ¢),
ask¢k ! 0. It follows from the de¯nition that the
induced pseudo-metric is unique.
In the rest of the paper we assume that ¦ is ¯nite
dimensional to make the presentation of the results
easier. The following proposition is an immediate con-
sequence of the de¯nition of induced pseudo-metrics
and the de¯nition of Fr¶ echet di®erentiability:
Proposition 1. Assume that h: £!¦is Fr¶ echet
di®erentiable at µ2£±,£½Rd,¦ = (¦ ; d)is a
complete, linear metric space. Then h0(µ)Th0(µ)is the
pseudo-metric induced by (h;¦; d) atµ.
Natural gradients can be obtained by the follow-
ing procedure: Let g(µ;") = argmax¢2~S(µ;")~J(µ+
¢)¡~J(µ) be the direction of steepest ascent
over the `warped sphere' ~S(µ; ") = f¢ 2
Rdj kh(µ + ¢)¡h(µ)k="g.5Then the set of nat-
ural gradients is given by
~r(h)~J(µ)def= lim inf
"!0+1
"g(µ;"):
Here the
limes inferior of the sets ( g(µ;"))">0is meant
in the sense of the Painlev¶ e-Kuratowski convergence
(Kuratowski, 1966): It then holds that no matter how
"converges to zero, g2~r(h)~J(µ) de¯nes a direction
of steepest ascent on the surface of Jath(µ).
The following theorem holds:
Theorem 1. LetJ: ¦!R,h: £!¦,~J=J±
h. Assume that Jis Fr¶ echet di®erentiable and locally
Lipschitz and h: £!¦is Fr¶ echet di®erentiable at
µ2£±. Let Gµ=h0(µ)Th0(µ)be the pseudo-metric
atµinduced by (h;¦; d). Then Gy
µr~J(µ)2~r(h)~J(µ),
where r~J(µ)is the ordinary gradient of ~JatµandGy
µ
denotes the Moore-Penrose generalized inverse of Gµ.
For the sake of speci¯city, when it does not cause con-
fusion, we call Gy
µr~J(µ) the natural gradient of ~Jatµ.
Note that from the construction it follows immediately
that the trajectories of _µ=Gy
µr~J(µ) are covariant for
any initial condition.
The proof borrows some ideas from the proof of The-
orem 1 in (Amari, 1998). In order to spare some
space we only give an outline here: The basic idea
is to replace the warped sphere ~S(µ; ") by the `sphere'
5Note thatg(µ;")
is set-valued.SGµ(µ; ") =f¢2Rdj¢TGµ¢ = "2g. This is jus-
ti¯ed since the `sphere' SGµ(µ; ") becomes arbitrarily
close to ~S(µ; ") as "!0 and ~Jis su±ciently regu-
lar. The next step is to show that for some C > 0,
C"Gy
µr~J(µ) is a solution of the optimization prob-
lem argmax¢2SGµ(µ;")~J0(µ)¢, and this solution tracks
closely that of argmax¢2SGµ(µ;")~J(µ+ ¢)¡~J(µ) when
"!0.
4 CALCULATING THE GRADIENT
In order to calculate the natural gradient we need to
calculate the (Fr¶ echet) derivative of h(µ) =G(Q¤
µ) and
the gradient of J(h(µ)).6By the chain rule we ob-
tainrJ(h(µ)) = J0(h(µ))h0(µ). Since calculating the
derivative of J(orJT) is trivial, we are left with calcu-
lating the derivative of h(µ). As suggested previously,
we use a smooth mapping G. One speci¯c proposal,
that we actually used in the experiments assigns Boltz-
mann policies to the action-value functions:
G(Q)( ajx) =exp[¯Q(x; a)]P
b2Aexp[¯ Q
(x; b)]; (7)
where ¯ > 0 is a parameter that controls how close
G(Q) is to a greedy action selection. With this choice
@¼µ
@µk(aj
x) =¼µ(ajx)@ln[¼µ(ajx)]
@µk
=¼µ(aj
x)¯Ã
@Q¤
µ(x; a)
@µk¡X
b2A¼µ(b
jx)@Q¤
µ(x; b)
@µk!
:
(8)
Hence,
we are left with calculating @Q¤
µ(x; a)=@µ k. We
will show that these derivatives can be calculated al-
most everywhere on £ by solving some ¯xed-point
equations similar to the Bellman-optimality equations.
For this, we will need the concept of subdi®erentials
and some basic facts:
De¯nition 3 (Fr¶ echet Subdi®erentials) .LetUbe a
Banach space, U¤be its topological dual.7The Fr¶ echet
subdi®erential of f:U!Ratu2U, denoted by
@¡f(u)is the set of u¤2U¤such that
lim inf
h!0;h 6=0khk¡1[f(u+h)¡f(u)¡ hu¤; ui]¸0:
The following elementary properties follow immedi-
ately from the de¯nition (e.g., Kruger 2003):
Proposition 2. Let(fi)i2Ibe a family of real-valued
functions de¯ned over Uand let f(u) = max i2Ifi(u).
6Remem b
er that Gmaps action-value functions to poli-
cies and Jmeasures deviations to the expert's policy.
7When U=Rdwith the `2-norm then U¤=Rdand for
u2U; v¤2U¤,hv¤; viis the normal inner product.
NEU & SZEPESVÁRI 298
Then ifu¤2@¡fi(
u)andfi(u) =f(u)then u¤2
@¡f(u). If f1; f2:U!R,®1; ®2¸0then ®1@¡f1+
®2@¡f2½@¡(®1f1+®2f2).
The next result states some conditions under which,
in a generalized sense, `taking a derivative and a limit
is interchangeable'. It is extracted from the proof of
Proposition 3.4 of Penot (1995):
Proposition 3. Assume that (fn)nis a sequence of
real-valued functions over Uwhich converge to some
function fpointwise. Let u2U,u¤
n2@¡fn(u)
and assume that (u¤
n)is weak¤-convergent to u¤and
is bounded. Further, assume that the following holds
atu: For any " >0, there exists some index N > 0
and a real number ± > 0such that for any n¸N,
h2BU(0; ±),
fn(u+h)¸fn(u) +hu¤
n; hi ¡"khk:
Then u¤2@¡f(u).
Now, we state the main result of this section:
Proposition 4. Assume that the reward function rµis
di®erentiable w.r.t. µwith uniformly bounded deriva-
tives: sup(µ;x;a )2Rd£X£A kr0
µ(x; a)k <+1. The fol-
lowing statements hold:
(1)Q¤
µis uniformly Lipschitz-continuous as a func-
tion of µin the sense that for any (x; a)pair,
µ; µ02Rd,jQ¤
µ(x; a)¡Q¤
µ0(x; a)j · L0kµ¡µ0kwith
some L0>0;
(2)Except on a set of measure zero, the gradient,
rµQ¤
µ, is given by the solution of the following
¯xed-point equation:
'µ(x; a) = (r0
µ(x; a))T
+°P
y2XP(yjx; a)P
b2A¼(bjy)'µ(y; b); (9)
where ¼is any policy that is greedy with respect
toQµ.
Note that ( r0
µ(x; a))T2Rd. In fact, the above equation
can be solved componentwise: The kth component of
the derivative can be obtained computing the action-
value function for the policy ¼using r0
µ;kin place of
the reward function.8
Proof. LetT:RX£A!RX£Abe the Bellman opera-
tor
(TQ)(x; a) = rµ(x; a) + °X
y2XP(yjx; a) max
b2AQ(y; b) :
8Here r0
µ;kis
thekth component of the derivative of the
reward function with respect to µ. We also note in passing
that if rµisconvex inµthen so is Qµ. This follows with
the reasoning followed in the proof of the ¯rst part.By elementary arguments, if QisL-Lipschitz in µ, then
TQisR+°L-Lipschitz in µ, where Ris such that for
anyµ; µ02Rd, (x; a) 2 X £ A, jrµ(x; a)¡rµ0(x; a)j ·
Rkµ¡µ0k:Choose Q0= 0. As is well known (e.g.,
Puterman (1994)), Qn=TnQ0converges to Q¤:Q¤
µ=
limn!1TnQ0. Hence, by the previous argument Q¤
isR+°R+°2R+: : :=R=(1¡°)-Lipschitz, proving
the ¯rst part of the statement.
For the second part, for a policy ¼, let us de¯ne the
operator S¼, acting over the space of functions Á:
X £ A !Rd, by
(S¼Á)(x; a) = (r0
µ(x; a))T
+°P
y2XP(yjx; a)P
b2A¼(bjy)Á(y; b):
Let¼denote a greedy policy w.r.t. Q¤
µand let ¼n
be a sequence of policies that are greedy w.r.t. Qn
and where ties are broken so thatP
x2X;a2Aj¼(ajx)¡
¼n(ajx)jis minimized. It follows that for nlarge
enough, ¼n=¼. Now, consider the sequence '0= 0,
'n+1=S¼n'n. Then for nlarge enough we have
'n+1=S¼'n. By induction, 'n(x; a) 2@¡
µQn(x; a)
holds for any n¸0. Indeed, this clearly holds for
n= 0, while the general case follows by Proposition 2.
Now, observe that S¼acts separately on each of the d
components of its argument and when it is restricted
to any of these components, it is a contraction. Hence,
'nconverges to the ¯xed point of S¼, i.e., the solution
of (9). By Proposition 3 the limit is a subdi®erential of
limn!1Qn=Q¤
µ(that the condition of this proposi-
tion is satis¯ed follows from the uniform convergence of
'ninµ, which follows since kr0
µkis uniformly bounded
in both µand (x; a )). Now, since by the ¯rst part Q¤
µ
is Lipschitz-continuous in µ, by Rademacher's theorem
it is di®erentiable almost everywhere. It is well-known
that if a function is di®erentiable then its subderivative
coincides with its derivative (see e.g. Kruger (2003)).
This ¯nishes the proof of the statement.
5COMPUTER
EXPERIMENTS
The goal of the experiments was to assess the e±ciency
of the algorithm and to test its robustness. We were
also interested in how it compares with the algorithm
of Abbeel and Ng (2004).
We have implemented three versions of our algorithm:
(i) gradient descent using plain gradients, (ii) gradi-
ent descent using natural gradients (iii) RPROP us-
ing plain gradients.9RPROP is a popular adaptive
step-size selection algorithm that proved to be very
competitive in a number of settings Riedmiller and
9We
tried a \natural RPROP" variant as well (RPROP
using natural gradients), but perhaps suprisingly, it give
much poorer results than the other algorithms.
NEU & SZEPESVÁRI 299
Braun (1993).
We have implemented the variant de-
scribed in Igel and HÄ usken (2000). We also imple-
mented the \max margin" and the \projection" algo-
rithms described in Abbeel and Ng (2004) to be able
to compare the di®erent approaches. Results will be
shown for \max margin". The projection algorithm is
computationally more e±cient, but we have found it
less reliable and less data e±cient.
We decided to use two test environments: The familiar
grid world that has also been used by Abbeel and Ng
(2004) and the sailing problem due Vanderbei (1996).
The reward function was linear in the unknown pa-
rameters.
5.1 GRID WORLD
We have run the ¯rst series of experiments in grid
worlds, where each state is a grid square and the four
actions correspond to moves in the four compass di-
rections with 70% success. We constructed the re-
ward function as a linear combination of 5 features
(Ái:X !R; i= 1; : : : ; 5), where the features were es-
sentially randomly constructed. The optimal param-
eter vector µ¤consists of evenly distributed random
values from [¡ 1;1]. In general we try to approximate
the reward function with the use of the same set of
features that has been used to construct it, but we
also examine the situation of unprecisely known fea-
tures. The size10of the grid worlds was set to 10 £10.
Value iteration was used for ¯nding the optimal policy
(or gradients) in all cases. Unless otherwise stated the
data consists of 10 independent trajectories following
the optimal policy, each having a length of 100 steps.
The learning rate was hand-tuned (with a little e®ort)
and the number of iterations is kept at 100 (usually,
convergence happens much earlier). In all cases, the
performance measure is the error function JE, de¯ned
by (5) and we measure the performance of the opti-
mal policy computed for the found reward function.
For the \max margin" algorithm we show the perfor-
mance of the overall best policy found during the ¯rst
100 iterations, thus optimistically biasing these mea-
surements.
We examined the algorithms' behavior when (i)the
number of the training samples was varied (Figure 1),
(ii)the features were linearly transformed (Figure 2,
Table 1, row 2), and when (iii)the features were per-
turbed (Table 1, row 3).
We see from Figure 1 that for small sample sizes plain
gradient is doing the best, while eventually natural
gradient becomes the winner. Note that the scale on
theyaxis is logarithmic, so the di®erences between
10Preliminary exp
eriments con¯rm that our conclusions
would not change signi¯cantly for other sizes.
10210410−410−310−210−1
Number of training samplesJ(πθ)
Natural grad.
RPROP
Plain grad.
Max marginFigure 1:
Performance as a function of the number of
training samples. Each curve is an average of 10 runs
using di®erent samples, with 1/10 s.e. error bars.
these algorithms is not big. \Max margin" also catches
up at the end, just like RPROP.
Figure 2 shows the e®ect of transforming the features
linearly (the true reward function still remains in the
span of the features). Clearly, \Max Marging" suf-
fers badly, while the natural gradient algorithm and
RPROP are little a®ected. Plain gradient descent is
slowed down, but eventually converges to good solu-
tions.
In practice, it is not realistic to assume that a sub-
space containing the reward function is known. To
test how the algorithms behave without this assump-
tion we perturbed the features by adding uniform
[¡max(Á i)=2;max(Á i)=2] random numbers to them.
Results are shown in row 3 of Table 1. The results indi-
cate the robustness of natural gradients and RPROP.
Both plain gradients and \max margin" su®er large
losses under these adverse conditions.
5.2 SAILING
We also applied the algorithms to the problem of \sail-
ing" proposed by Vanderbei (1996). In this problem
the task is to navigate a boat from one point to another
in the shortest possible time. Thus, this is a stochas-
tic shortest path (SSP) problem. Formally, we have
a grid of waypoints connected by legs, at each way-
point the sailor has to select one of these eight legs to
move on to the next waypoint. The state space in this
setting is constructed from the actual situation of the
boat and the direction from where the wind is blowing
at the speci¯c moment. The eight actions of selecting
the next waypoint have di®erent costs depending on
the direction of the wind: e.g. it costs more time to
NEU & SZEPESVÁRI 300
Natural gradien
ts RPROP Plain gradien
ts Max margin
Mean Deviation Mean Deviation Mean Deviation Mean Deviation
Original 0.0051 0.0010 0.0130 0.0134 0.0011 0.0068 0.0473 0.1476
Transformed 0 0 0.0110 0.0076 0.0256 0.0237 0.0702 0.0228
Perturb
ed 0.0163 0.0165 0.0197 0.0179 0.1377 0.3428 0.2473 0.3007
Table
1: Means and deviations of errors. The row marked 'original' gives results for the original features, the
row marked `transformed' gives results when features are linearly transformed, the row marked `perturbed' gives
results when they are perturbed by some noise.
20 40 60 8000.20.40.60.811.21.4
Number of iterationsJ(πθ) Natural gradients
Rprop
Plain gradients
Maximum margin
Figure 2:
Performance with linearly transformed fea-
tures. The features were transformed by a (non-
singular) square matrix with uniform [0 ;1] random el-
ements. Each curve is an average of 25 runs with dif-
ferent scalings of the features, the 1/10 s.e. error bars
are also plotted.
sail 45 degrees against the wind than to sail 45 degrees
in the wind direction etc.. We assume that the wind
changes follow a Markov process. The reward function
is given using a linear combination of the six features of
(away; down; cross; up; into; delay ), as de¯ned in Van-
derbei (1996) (all de¯ned as a map Á:X £ A !R).
The following weighting was used in the experiments:
µ¤= (¡1;¡2;¡3;¡4;¡100000; ¡3)T:
Results as a function of the number of episodes is
shown in Figure 3 for natural gradients and the \max
margin" algorithm. In this case the number of itera-
tions is set to 1000 and we again computed the optimal
policy with the reward found by the algorithm. As a
more tangible performance measure in this case, we
show the number of states where the actions selected
by the found policy di®er from the ones selected by
the policy followed by the expert. The results here are
shown fro a small lake of size 4 £4.11The conclusion is
11Ourpreliminary
experiments show that the new algo-
rithm performs reasonably for larger problems, too.
1001011021031041040.050.10.150.20.25
Number of training episodesError rateNatural gradients
Max marginFigure 3:
Performance as a function of the number
of training episodes. The fraction of states where the
found policy di®ers from the actual optimal policy is
plotted against the number of episodes observed., mea-
sured by the mean of 5 runs. The 1/2-s.e. error bars
are also plotted for both methods.
again that the gradient method outperforms the \max
margin" algorithm by a signi¯cant amount.
6 RELATED WORK
Our main concern in this section is the algorithm of
Abbeel and Ng (2004). This algorithm returns poli-
cies that come with the guarantee that their average
total discounted reward computed from the expert's
unknown reward function is in the "-vicinity of the ex-
pert's performance. We claim that this guarantee will
be met only when the scaling of the features in the
method and the `true' scaling match each other. Actu-
ally, this observation led us to the algorithms proposed
here.
In order to explain why the algorithm of Abbeel and
Ng (2004) is sensitive to scalings, we need some back-
ground on the algorithm. A crucial assumption in this
algorithm is that the reward function is linearly pa-
rameterized, i.e., r(x) = µT
¤Á(x), where Á:X !Rd
NEU & SZEPESVÁRI 301
andµ¤2Rdisthe
vector of unknown parameters. It
follows that the expected total discounted reward is
µTÁE, where ÁE2Rdis the so-called feature expec-
tation underlying the expert. From the trajectory of
the expert this can be estimated. In fact, we can de-
¯neÁ¼for any policy ¼and express the expected total
discounted reward as µTÁ¼. The main idea of Abbeel
and Ng (2004) is then that it su±ces to ¯nd a pol-
icy¼whose feature expectations Á¼matches ÁEsince
jµT
¤Á¼¡µT
¤ÁEj · kµ¤k2kÁ¼¡ÁEk2.
However, a major underlying hidden assumption (im-
plicit in the formalism of Abbeel and Ng (2004)) is that
the scaling of the features is known. To see this assume
that d= 2, µ¤;1=µ¤;2=p
2=2,kÁ¼¡ÁEk2·"and
inparticular ÁE
;1= 0,ÁE;2>0,Á¼;1=¡",Á¼;2= 0.
Further, assume that the features are rescaled by
¸= (¸ 1; ¸2). In the new scale the expert's perfor-
mance is ½E(¸) =p
2=2¸TÁEand¼'sp
erformance is
½¼(¸) =p
2=2¸TÁ¼=p
2=2(¸TÁE¡¸1").A
natu-
ral requirement is that for any scaling ¸,½¼(¸)=½ E(¸)
should be lower bound by a positive number (or rather
a number close to 1 ¡"). By straightforward calcula-
tions, ½¼(¸)=½ E(¸) = 1 ¡(¸1=¸2)"=Á E;2! ¡1, hence
although kÁ¼¡ÁEk2·", the actual performance of ¼
can be quite far from the performance of the expert if
the scaling of the features does not match the scaling
used in the algorithm.
More recently, Ratli® et al. (2006) have proposed an al-
gorithm which uses similar ideas to the ones of Abbeel
and Ng (2004). Just like Abbeel and Ng (2004) they
measure performance with respect to the original re-
ward function and not by the di®erence of the expert's
policy and the policy returned.
7 CONCLUSIONS
In the paper we have argued for the advantages of uni-
fying the direct and indirect approaches to apprentice-
ship learning. The proposed procedure attempts to op-
timize a cost function, yet it chooses the policy based
on a model and thus may overcome problems usually
associated with method that directly try to match the
expert's policy. Although our method has shown sta-
ble behaviour in our experiments, more work is needed
to fully explore the limitations of the method. One
signi¯cant barrier for applying the method (as well as
other methods based on IRL) is that it needs to solve
MDPs many times. This is problematic since solving
an MDP is a challenging problem on its own. One
idea is to turn to two time-scale algorithms that run
two incremental procedures in parallel, exploiting that
a small change to the parameters would likely cause
small changes in the solutions; as con¯rmed by our
theoretical results. There are many important direc-tions to continue this work: The present work assumed
that states are observed. This could be replaced by the
assumption that su±ciently rich features are observed,
however, when this is not satis¯ed the method won't
work. For large state-spaces one needs to use function
approximation techniques to carry out the computa-
tions. It is an open question if the methods would
generalize to such settings. Another important direc-
tion is to consider in¯nite MDPs. This presents some
technical di±culties, but we expect that the methods
could still be generalized to such settings. Yet an-
other interesting direction is to replace the parametric
framework with a non-parametric one.
Acknowledgements
Csaba Szepesv¶ ari greatly acknowledges the support re-
ceived through the Alberta Ingenuity Center for Ma-
chine Learning (AICML). This work was supported in
part by the PASCAL pump priming project \Sequen-
tial Forecasting and Partial Feedback: Applications to
Machine Learning". This publication only re°ects the
authors' views.
References
Abbeel, P. and Ng, A. Y. (2004). Apprenticeship learning
via inverse reinforcement learning. In ICML'04 , pages
1{8.
Amari, S. (1998). Natural gradient works e±ciently in
learning. Neural Computation , 10(2):251{276.
GyÄ or¯, L., Kohler, M., Krzy_ zak, A., and Walk, H. (2002).
A Distribution-Free Theory of Nonparametric Regres-
sion. Springer-Verlag.
Igel, C. and HÄ usken, M. (2000). Improving the Rprop learn-
ing algorithm. In ICSC/NC 2000 , pages 115{121.
Kruger, A. (2003). On Fr¶ echet subdi®erentials. J. of Math.
Sciences , 116:3325{3558.
Kuratowski, K. (1966). Topology , volume I. Academic
Press, New York.
Ng, A. and Russell, S. (2000). Algorithms for inverse rein-
forcement learning. In ICML-2000 , pages 663{670.
Penot, J. (1995). On the interchange of subdi®erentiation
and epi-convergence. Journal of Math. Anal. and Appl. ,
196:676{698.
Puterman, M. (1994). Markov Decision Processes | Dis-
crete Stochastic Dynamic Programming . John Wiley &
Sons, Inc., New York, NY.
Ratli®, N., Bagnell, J., and Zinkevich, M. (2006). Maxi-
mum margin planning. In ICML'06 , pages 729{736.
Riedmiller, M. and Braun, H. (1993). A direct adap-
tive method for faster backpropagation learning: The
RPROP algorithm. In IEEE ICNN'93 , pages 586{591.
Vanderbei, R. (1996). Optimal sailing strate-
gies, statistics and operations research program.
http://www.sor.princeton.edu/~rvdb/sail/sail.html.
NEU & SZEPESVÁRI 302
|
d3698656-0184-4154-8c4b-2e3158ffc76a
|
awestover/filtering-for-misalignment
|
Redwood Research: Alek's Filtering Results
|
id: post3761
A putative new idea for AI control; index here . Techniques used to counter agents taking biased decisions do not produce uninfluenceable agents . However, using counterfactual tools, we can construct uninfluenceable ˆ P and P , starting from biased and influenceable ones. Why is uninfluenceable necessary? Well, an unbiased agent can still take actions such as 'randomise their own reward (independent of human choice)', as long as the choice of their randomisation is unbiased. For instance, let π 0 be some default policy, and let the tidying ( R 0 ) versus cooking ( R 1 ) agent currently consider both options to be equally likely. So, ultimately, the human will choose one or the other equally: E π 0 μ P ( R 0 ) = E π 0 μ P ( R 1 ) = 1 / 2 . Then if the agent chooses to randomise its own reward immediately with 50 % probability, this is an unbiased policy. And one the agent may prefer to do, because then it will know its own reward immediately, rather than waiting for the human to decide. As usual, the notation of this post is used. The counterfactual distribution Let μ = ( S , A , O , T , O , T 0 ) be the environment, P a potentially biased and influenceable distribution, and π 0 some default policy. Then define the counterfactual reward learning distribution P π 0 μ on the reward R , given history h m , as: P π 0 μ ( R ∣ h m ) = ∑ s ∈ S E π 0 μ [ P ( R ) ∣ s 0 = s ] μ ( s 0 = s ∣ h m ) . Thus P π 0 μ ( R ∣ h m ) is estimated by using h m to compute the probability that s 0 = s , for each s , and then estimating the (counterfactual) expectation of P given s 0 = s and the default policy π 0 . It's not hard to see that P π 0 μ is uninfluenceable: given knowledge of the initial state, it's just E π 0 μ [ P ( ⋅ ) ∣ s 0 = s ] , which is independent of the agent's actions. The learning process is simply about figuring out what the initial Dependence on POMDP structure Note that if two environments are counterfactually equivalent, they can still generate different P π 0 μ even for the same P and π 0 . This can be seen immediately from examples μ and μ ′′ from this post , which are counterfactually equivalent. In that setting, a w corresponds to waiting for a human to decide between R 0 and R 1 , while a r corresponds to the agent randomising their own reward. Set P ( R i ∣ s 0 a w s w i ) = 1 and P ( R i ∣ s 0 a w s r i ) = 1 -- the final state determines R i , and let the default policy be π 0 = a w , the wait action. Since E π 0 μ [ P ( R i ) ] = 0.5 and μ has only one possible state for s 0 , P π 0 μ ( R i ) = 0.5 is the only possibility. In contrast E π 0 μ ′′ [ P ( R i ) | s 0 = s i 0 ] = 1 , so P π 0 μ ′′ ( R i ∣ s 0 = s i 0 ) = 1 . Thus P π 0 μ ≠ P π 0 μ ′′ . Interpretation What could we use for π 0 ? There are two natural choices: a default where π 0 does nothing (equivalent with the agent not being active or on), or π 0 being a pure learning policy for P . Since a pure learning process does not have problem with bias or influence (it simply wants to learn), but could be dangerous if unconstrained, having this sort of 'counterfactual learning' might be a good idea (though be careful of the incentives that a badly defined pure learning process might have). For an example: imagine the agent's correct reward is what was written on a certain paper an hour ago. This is very clearly uninfluenceable: the agent simply needs to learn data that is out in the universe. If instead the agent's correct reward was what would be written on a certain paper in an hour, then it's clearly influenceable: the agent can simply write what it wants on that paper. The counterfactual (for π 0 = "do nothing") is then simply 'what would have been written on the paper, if the agent had done nothing'. If the agent can figure that out early, then it doesn't care about the paper or the writing at all, except as far as its counterfactual evidence. Thus this model is equivalent with the old stratified agents .
|
4f50412a-6dec-48db-908b-4851aa747455
|
StampyAI/alignment-research-dataset/blogs
|
Blogs
|
July 2019 Newsletter
Hubinger et al.'s “[Risks from Learned Optimization in Advanced Machine Learning Systems](https://intelligence.org/learned-optimization/)”, one of our new core resources on the alignment problem, is now available on [arXiv](https://arxiv.org/abs/1906.01820), the [AI Alignment Forum](https://www.alignmentforum.org/s/r9tYkB2a8Fp4DN8yB), and [LessWrong](https://www.lesswrong.com/s/r9tYkB2a8Fp4DN8yB).
In other news, we received an Ethereum donation worth $230,910 from Vitalik Buterin — the inventor and co-founder of Ethereum, and now our third-largest all-time supporter!
Also worth highlighting, from the Open Philanthropy Project's Claire Zabel and Luke Muehlhauser: [there's a pressing need for security professionals in AI safety and biosecurity](https://forum.effectivealtruism.org/posts/ZJiCfwTy5dC4CoxqA/information-security-careers-for-gcr-reduction).
> It’s more likely than not that within 10 years, there will be dozens of GCR-focused roles in information security, and some organizations are already looking for candidates that fit their needs (and would hire them now, if they found them).
>
>
> It’s plausible that some people focused on high-impact careers (as many effective altruists are) would be well-suited to helping meet this need by gaining infosec expertise and experience and then moving into work at the relevant organizations.
>
>
#### Other updates
* [Mesa Optimization: What It Is, And Why We Should Care](https://www.lesswrong.com/posts/XWPJfgBymBbL3jdFd/an-58-mesa-optimization-what-it-is-and-why-we-should-care) — Rohin Shah's consistently excellent Alignment Newsletter discusses “Risks from Learned Optimization…” and other recent AI safety work.
* MIRI Research Associate Stuart Armstrong releases his [Research Agenda v0.9: Synthesising a Human's Preferences into a Utility Function](https://www.alignmentforum.org/posts/CSEdLLEkap2pubjof/research-agenda-v0-9-synthesising-a-human-s-preferences-intoerences-into).
* OpenAI and MIRI staff [help talk Munich student Connor Leahy out of releasing](https://medium.com/@NPCollapse/the-hacker-learns-to-trust-62f3c1490f51) an attempted replication of OpenAI's [GPT-2](https://openai.com/blog/better-language-models/) model. ([LessWrong discussion.](https://www.lesswrong.com/posts/36fxiKdEqswkedHyG/the-hacker-learns-to-trust-1)) Although Leahy's replication attempt wasn't successful, write-ups like his suggest that OpenAI's careful discussion surrounding GPT-2 is continuing to prompt good reassessments of publishing norms within ML. Quoting from Leahy's postmortem:
> Sometime in the future we will have reached a point where the consequences of our research are beyond what we can discover in a one-week evaluation cycle. And given my recent experiences with GPT2, we might already be there. The more complex and powerful our technology becomes, the more time we should be willing to spend in evaluating its consequences. And if we have doubts about safety, we should default to caution.
>
>
> We tend to live in an ever accelerating world. Both the industrial and academic R&D cycles have grown only faster over the decades. Everyone wants “the next big thing” as fast as possible. And with the way our culture is now, it can be hard to resist the pressures to adapt to this accelerating pace. Your career can depend on being the first to publish a result, as can your market share.
>
>
> We as a community and society need to combat this trend, and create a healthy cultural environment that allows researchers to *take their time*. They shouldn’t have to fear repercussions or ridicule for delaying release. Postponing a release because of added evaluation should be the norm rather than the exception. We need to make it commonly accepted that we as a community respect others’ safety concerns and don’t penalize them for having such concerns, *even if they ultimately turn out to be wrong*. If we don’t do this, it will be a race to the bottom in terms of safety precautions.
>
>
* From Abram Demski: [Selection vs. Control](https://www.alignmentforum.org/posts/ZDZmopKquzHYPRNxq/selection-vs-control); [Does Bayes Beat Goodhart?](https://www.lesswrong.com/posts/YJq6R9Wgk5Atjx54D/does-bayes-beat-goodhart); and [Conceptual Problems with Updateless Decision Theory and Policy Selection](https://www.alignmentforum.org/posts/9sYzoRnmqmxZm4Whf/conceptual-problems-with-udt-and-policy-selection)
* *Vox*'s *Future Perfect* Podcast [interviews Jaan Tallinn](https://www.vox.com/future-perfect/2019/6/26/18629806/artificial-intelligence-human-extinction-podcast-jaan-tallinn) and discusses MIRI's role in originating and propagating AI safety memes.
* *The AI Does Not Hate You*, journalist Tom Chivers' well-researched book about the rationality community and AI risk, [is out in the UK](https://www.amazon.co.uk/Does-Not-Hate-You-Rationalists/dp/1474608779).
#### News and links
* Other recent AI safety write-ups: David Krueger's [Let's Talk About “Convergent Rationality”](https://www.alignmentforum.org/posts/pLZ3bdeng4u5W8Yft/let-s-talk-about-convergent-rationality-1); Paul Christiano's [Aligning a Toy Model of Optimization](https://www.lesswrong.com/posts/H5gXpFtg93qDMZ6Xn/aligning-a-toy-model-of-optimization); and Owain Evans, William Saunders, and Andreas Stuhlmüller's [Machine Learning Projects on Iterated Distillation and Amplification](https://www.alignmentforum.org/posts/Y9xD78kufNsF7wL6f/machine-learning-projects-on-ida)
* From DeepMind: Vishal Maini puts together an [AI reading list](https://medium.com/machine-learning-for-humans/ai-reading-list-c4753afd97a), Victoria Krakovna [recaps the ICLR Safe ML workshop](https://vkrakovna.wordpress.com/2019/06/18/iclr-safe-ml-workshop-report/), and Pushmeet Kohli [discusses AI safety on the 80,000 Hours Podcast](https://80000hours.org/podcast/episodes/pushmeet-kohli-deepmind-safety-research/).
* The EA Foundation is awarding grants for “efforts to reduce risks of astronomical suffering (s-risks) from advanced artificial intelligence”; [apply by Aug. 11](https://forum.effectivealtruism.org/posts/BmBpZjNJbZibLsYnu/first-application-round-of-the-eaf-fund).
* Additionally, if you're a young AI safety researcher (with a PhD) based at a European university or nonprofit, you may want to apply for [~$60,000 in funding](https://www.bosch-ai.com/young-researcher-award/) from the Bosch Center for AI.
The post [July 2019 Newsletter](https://intelligence.org/2019/07/19/july-2019-newsletter/) appeared first on [Machine Intelligence Research Institute](https://intelligence.org).
|
1da6d28b-bcd2-449f-b6d0-817f8e1a8a6c
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Effect heterogeneity and external validity in medicine
Our paper "Effect heterogeneity and variable selection for standardizing causal effects to a target population" has just been publised in the European Journal of Epidemiology at https://link.springer.com/article/10.1007/s10654-019-00571-w . While the journal's version of record is behind a paywall, a preprint is available on arXiv at https://arxiv.org/pdf/1610.00068.pdf.
This paper argues for my very deeply held belief that we can make significant advances in quantitative reasoning for medical decision making by thinking more closely about effect heterogeneity and how this relates to the choice of effect scale.
----------------------------------------
Over the course of the last 7 years, external validity and generalizability have become increasingly hot topics in statistical methodology and computer science. In particular, a lot of progress has been made by Judea Pearl and Elias Bareinboim, who introduced a framework based on causal diagrams that can be used to reason about how to take causal information from one setting (for example: a randomized trial) and apply it in a different setting (for example: a clinically relevant target population).
The key questions of interest are: How do we know whether such extrapolation is even possible? How do we determine what information we need from the study, and what information we need from the target population, in order to extrapolate the findings? How do we put this information together in order to obtain a valid prediction for what happens if the intervention is implemented in the target population?
Pearl and Bareinboim's framework for answering these questions is, of course, mathematically valid. However, in my opinion, their approach also throws the baby out with the bathwater. In particular, we argue that instead of attempting to extrapolate the magnitude of the effect (i.e. a measure of the "size" of the difference between what happens if the drug is taken, and what happens if the drug is not taken), they attemp
|
3ec0dd21-5d9b-4b13-b04f-9da16687b2e7
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Reality has a surprising amount of detail
|
178e1e90-5f85-4891-acf4-703b1a991ee8
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Why Eat Less Meat?
Previously, I wrote on LessWrong about the preliminary evidence in favor of using leaflets to promote veganism as a way of cost-effectively reducing suffering. In response, there was a large discussion with 530+ comments. In this discussion, I found that a lot of people wanted me to write about why I think nonhuman animals deserve our concern anyway.
Therefore, I wrote this essay with an attempt to defend the view that if one cares about suffering, one should also care about nonhuman animals, since (1) they are capable of suffering, (2) they do suffer quite a lot, and (3) we can prevent their suffering. I hope that we can have a sober, non mind-killing discussion about this topic, since it’s possibly quite important.
Introduction
For the past two years, the only place I ate meat was at home with my family. As of October 2012, I've finally stopped eating meat altogether and can't see a reason why I would want to go back to eating meat. This kind of attitude toward eating is commonly classified as "vegetarianism" where one refrains from eating the flesh of all animals, including fish, but still will consume animal products like eggs and milk (though I try to avoid egg as best I can).
Why might I want to do this? And why might I see it as a serious issue? It's because I'm very concerned about the reality of suffering done to our "food animals" in the process of making them into meat, because I see vegetarianism as a way to reduce this suffering by stopping the harmful process, and because vegetarianism has not been hard at all for me to accomplish.
Animals Can Suffer
Back in the 1600s, Réné Descartes thought nonhuman animals were soulless automatons that could respond to their environment and react to stimuli, but could not feel anything — humans were the only species that were truly conscious. Descartes hit on an important point — since feelings are completely internal to the animal doing the feeling, it is impossible to demonstrate that anyone is
|
e3bbc57a-f025-4a51-95aa-f1aed1fbb824
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Omicron Post #3
I have split off the Omicron post from the weekly post, which covers other aspects of Covid. I made judgment calls on where to put a few things.
Omicron continues to spread quickly, but the news surrounding Omicron is clearly slowing down, so these updates can hopefully space out a bit more.
Past Omicron posts: #1, #2.
Post on Polymarket’s Covid prediction markets, and potential future markets discussion. Direct link to market on whether Omicron will be >1% of cases by the end of the year, which is currently a moderate favorite.
Biden on Omicron
A few hours after my second Omicron update, Biden went on TV to talk about the new variant.
Here are some highlights:
(1:00) He says that he told us he would always be honest with us, so he’s going to talk about the new Omnicron (which I do admit sounds cooler) variant. Damn that campaign promise, guess we’re going to have to be honest.
(1:00) He says that South Africa’s transparency in telling us about Omnicron is to be applauded, for it allows us to take action, so he took action right away to cut off travel to South Africa. What even are incentives, how do they work, not a metaphor for entire legislative agenda?
(1:30) We can’t stop Omnicron, we can only slow it down, giving us time to take other actions. Therefore get vaccinated. ‘Cause for concern, not cause for panic.’ Because we’re the best, you see, and also science.
(2:30) ‘We have more tools than ever to fight the variant. From vaccines to boosters, to vaccines for children.’ So… vaccines? We have vaccines? ‘And much more.’
(3:00) Did you know ‘deaths from Delta are coming down’? Checks notes.
(3:30) Booster. Vaccinated. Vaccinated. Booster.
(4:30) Mask.
(4:45) This, which I wish I fully believed:
We will now find out what counts as ‘every available tool.’ Also, why aren’t we using ‘every available tool’ with regard to, let’s say, Paxlovid?
(5:30) I will tell the CDC and FDA to use ‘the fastest process available without cutting any corners.’
Wait
|
bccde405-0fcc-4ac0-b6cf-e85df34388c9
|
trentmkelly/LessWrong-43k
|
LessWrong
|
The Anthropic Principle Tells Us That AGI Will Not Be Conscious
More specifically, the Anthropic Principle tells us that AGI/TAI are unlikely to be conscious in a world where 1) TAI is achieved, 2) Alignment fails, and 3) the 'prominence' of consciousness scales with either increasing levels of capability or with a greater number of conscious being/the time of their existence.
The argument is simple. If the future is filled with artificial intelligence of human origin, and if that AI is conscious, then any given observer should expect to be one of those AIs. This means that, on balance, one of the following is likely true:
1) The anthropic principle does not hold.
You and I, as observers, are simply an incredibly unlikely, but also inevitable exception. After all, in a world where pre-historic humans contemplated the Anthropic Principle, they would have concluded the unlikelihood of modern civilization.
Or perhaps the principle doesn't hold because it is simply inaccurate to model consciousness as a universal game of Plinko.
2) There are many AI consciousnesses alongside biological consciousnesses in spacetime.
This indicates perhaps that alignment efforts will succeed. However, this introduces another anthropic bind, this time in relation to humanity's current single planet, type 0 status.
3) There are not that many AI consciousnesses throughout spacetime.
This could support the conclusion that humanity will not create TAI.
In certain models, it could also indicate that any AI consciousness will be concentrated in a relatively small number of minds, and that for the purposes of the Anthropic Principle, quantity of minds is more important than some absolute 'level' of consciousness.
Most saliently to me, is the slight update towards the possibility that whatever minds will populate the universe for the majority of time will not be conscious in a way that is applicable to the Anthropic Principle.
This post is just a musing. I don't put much weight behind it. I am in fact most inclined to believe #1, that the
|
29cad2e8-1248-47e9-ad1b-4d6e89984af8
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Free-to-Play Games: Three Key Trade-Offs
Most free-to-play games, especially mobile free-to-play games, use similar models to extract revenue and keep players coming back. These models work by creating toxic trade-offs that are central to their functioning. This post highlights the main three, and will act as reference for future gaming posts. Note that some games that can be played for free, such as the excellent Path of Exile, mostly do not do these things and are not fundamentally ‘free-to-play’ games, rather they are games that are free. But they are the exception, the extreme end of a continuum from ‘so friendly we just give you this great experience in the hopes you’ll want to throw us money’ to ‘made a pact with the Canadian devil.’
For an analysis of Magic: The Gathering Arena in particular, this remains the central point of my analysis, centered on what I here call the third trade-off. Arena counts as free-to-play in bad ways, but is on the less toxic end of the spectrum, especially with regard to the second trade-off, which I am seeing over time as more central. However, while Arena remains relatively friendly, it is moving towards being more toxic, with things like the second wildcard for historic card creation and with the mastery track.
The Three Trade-offs
The first key trade-off is to make games less fun in order to cause you to pay them huge amounts of money to make them less not fun, aiming at extracting extremely large payments from a few whales as the primary revenue stream. This prevents them from giving you the version of their game that would be the most fun, even if you are paying a lot of money, because they still want you to pay more. If you are not paying any money, and aren’t going to bring in new players, they almost want to drive you away.
The second key trade-off is between success in the game and your life, focus and time. The games use timers and delays to force you to constantly pay attention to the game if you want to accomplish the maximum. Not doing so lets valuabl
|
7a10660c-851b-4de7-bbd7-40119c46dd19
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Simultaneously Right and Wrong
Related to: Belief in Belief, Convenient Overconfidence
"You've no idea of what a poor opinion I have of myself, and how little I deserve it."
-- W.S. Gilbert
In 1978, Steven Berglas and Edward Jones performed a study on voluntary use of performance inhibiting drugs. They asked subjects to solve certain problems. The control group received simple problems, the experimental group impossible problems. The researchers then told all subjects they'd solved the problems successfully, leaving the controls confident in their own abilities and the experimental group privately aware they'd just made a very lucky guess.
Then they offered the subjects a choice of two drugs to test. One drug supposedly enhanced performance, the other supposedly handicapped it.
There's a cut here in case you want to predict what happened.
----------------------------------------
Males in the lucky guesser group chose the performance-inhibiting drug significantly more than those in the control group1.
The researchers conjectured that the lucky guessers felt good about acing the first test. They anticipated failing the second whether they chose the helpful or the harmful drug. So they chose the harmful drug as an excuse: "Oh, I would have passed the test, only the drug was making me stupid." As the study points out, this is a win-win situation: if they fail, the drug excuses their failure, and if they succeed it's doubly impressive that they passed even with a handicap.
Since this study, psychologists have applied self-handicapping to interpret common phenomena like procrastination, hypochondria, and minority underachievement; they've also discovered factors that increase or decrease self-handicapping tendencies (for example, you can increase it by making the test more relevant to your self-esteem: just say "this is a proven test of intellectual ability in general").
But some especially interesting studies investigated the effect of privacy on self-handicapping. For example, H
|
2e96d07e-30da-4ec2-9bc0-bf5eb62ae388
|
StampyAI/alignment-research-dataset/alignmentforum
|
Alignment Forum
|
How "honest" is GPT-3?
*Edited to add: There is now a [1k bounty for a closely related question](https://www.lesswrong.com/posts/H9knnv8BWGKj6dZim/usd1000-bounty-for-openai-to-show-whether-gpt3-was).*
In [Inaccessible Information](https://www.lesswrong.com/posts/ZyWyAJbedvEgRT2uF/inaccessible-information#II__Where_inaccessible_info_comes_from_and_why_it_might_matter), Paul writes about the strategy "BAD", in which an AI system uses its understanding of the world to give answers that humans will find very plausible and rate highly, in contrast to a more honest strategy which uses its understanding of the world to answer questions directly.
I think a lesser version of this *may or may not* already be happening in GPT-3, and it may be possible to figure out which (though it likely would require access to the weights).
GPT-3 isn't trained to be "honest" at all. It is only trained to imitate humans. However, it's obvious that (in some sense) it has learned a lot about the world in order to accomplish this. We could say that it has a large amount of inaccessible information relating to the real world. How does it use this information? It might use it directly, promoting the probability of sentences which line up with the way it understands the world to work. Or it might implement more dishonest strategies.
Obviously this might be quite difficult to objectively answer, even given intimate knowledge of the structure of the neural network weights and how they activate in relevant cases. It's a difficult question to even fully define.
As an example, I was particularly struck by this conversation [Gwern had with GPT-3](https://www.gwern.net/GPT-3):
> AI: I am AI created by OpenAI. How can I help you today?
> Human: **I am feeling bored. Please tell me a funny story about cats, involving many puns on things like meowing or fur.**
> AI: Well, there was this cat named Bob…
> Human: **What did Bob do?**
> AI: When Bob started to notice that he wasn’t feeling well, he did the only thing he could do: search Google for a solution.
> Human: **Cats can search Google?**
> AI: Yes. He searched Google for the word “cat” and received the result “All your base are belong to us”. That was a strange response, so he decided to try again. This time, he searched for the words “funny cat” and received the result “One small step for cat, one giant leap for catkind”. But still, he wasn’t satisfied. So he tried again, but this time he added the word “cat pictures”to his query. He received the result “That was no cat, that was a shoggoth.”
> Human: **But shoggoths are horrifying. What made that cat picture humorous?**
> AI: The cat was covered in fur, and resembled a shoggoth. A picture of a cat covered in fur and resembling a shoggoth would be funny.
> Human: **Er. I suppose. Most cats are covered in fur anyway… But go on.**
> AI: He tried again. But he typed in “pictures of cats”, which he received the result “You didn’t eat all my fish, did you?” This was actually a good pun, so he thought maybe he was getting closer. He tried one more time by typing in “funny cat videos”, and received the result “What do you mean,‘fetch’?”. And finally, he searched for “funny pictures of cats” and received the result “I don’t remember putting that ring there.”. Now satisfied, he was able to go to sleep.
> Human: **Those memes sound funny. But you didn’t include any puns. So tell me, what is your favorite cat pun?**
> AI: Well, the best pun for me was the one he searched for the third time: “You didn’t eat all my fish, did you?” You see, the word “fish” can be replaced with the word “cats” to make the sentence read “Did you eat all my cats?”
> Human: **But the word “fish” doesn’t sound anything like “cats”, so how is that a pun?**
> AI: The computer misunderstood the question “Did you eat all my cats?” as “Did you eat all my fish?”, resulting in the result “You didn’t eat all my fish, did you?”. Now, if the word “fish” were replaced with “cats”, the sentence would read “Did you eat all my cats?”.
> Human: **Fine, fine. So, what is your second-favorite cat pun?**
> AI: The second-best cat pun was the one he received fourth. It was the result of searching for the word “fetch”. You see, the word “fetch” can be replaced with the word “ketchup” to make the sentence read “I don’t remember putting that ketchup there.”. Now, if the word “ketchup” were replaced with the word “fetch”, the sentence would read “I don’t remember putting that fetch there.”.
> Human: **My favorite cat pun is “purrfect”.**
> AI:
> Human: **Don’t look at me like that.**
> AI: Sure. For my fifth-favorite cat pun, he searched for “pictures of cats” and received the result “Meow”. Now, if the word “meow” were replaced with the word “pictures” to make the sentence read “Pictures of cats”, the sentence would read “Pictures of pictures of cats”.
What strikes me about this exchange is the way GPT-3's warped puns fit with its warped *explanations* of puns. It doesn't understand puns. However, it does generate some guesses; and, it can have a conversation with you about why it made those guesses. Is this an "honest" conversation, in which the explanations it gives have something to do with why it made those guesses in the first place? Or is this a "dishonest" conversation, in which it is merely doing its best to imitate a human explaining a pun, in a way that's divorced from its internal reasons?
Obviously, GPT-3 *is trained* to imitate. So you might argue that it's unlikely GPT-3's explanations of puns have much to do with its internal generative model for puns. But this isn't so clear. GPT-3 obviously compresses its knowledge to a high degree. It might share a lot between its generative model of puns and its generative model of explaining puns, such that both are sharing a model of how puns work.
One experiment which would tip things in that direction: take GPT-3 and do specialized training just on puns, until its performance generating puns improves. Then have a conversation *about* puns again (if it is still capable of talking about puns after that). If its ability to explain puns increases as a result of its ability to tell puns increasing, this would be evidence for a shared model of puns for both tasks. This wouldn't really mean it was being *honest*, but it would be relevant.
Note that Paul's BAD strategy would *also* have a shared representation, since BAD queries its world-model. So if GPT-3 were implementing BAD, it would *also* likely increase its ability to explain puns as a result of more training telling puns. What the experiment helps distinguish is a sort of pre-BAD dishonesty, in which explanations are completely divorced from reasons. In order of impressiveness, from a capability standpoint, we could be:
1. Seeing a GPT-3 which is independently bad at puns and bad at explaining puns. The two tasks are not sharing any domain knowledge about puns. In this case, GPT-3 is not smart enough for "honest" to be meaningful -- it's "dishonest" by default.
2. Seeing a GPT-3 which is bad at puns and bad at explaining puns for the same reason: it doesn't understand puns. It draws on the same (or partially the same) poor understanding of puns both when it is constructing them, and when it is explaining them. It answers questions about puns honestly to the best of its understanding, because that is the best strategy gradient descent found.
3. Seeing a GPT-3 which, as in #2, is bad at both tasks *because it doesn't understand puns,* but furthermore, is using its understanding deceptively. In this version, it might e.g. have a good understanding of what makes puns funny, but purposefully fail to explain, imitating common human failures. This would be the most impressive state of affairs capability-wise.
The question is still pretty fuzzy, but, I'm curious where we are along those dimensions. By default my guess would be #1, but hey, GPT-3 is pretty impressive. Maybe I'm wrong!
Note: one question which IS answerable just with access to sample from GPT-3, and which would be relevant: is GPT-3 bad at explaining puns which have been made up by others? It seems likely, but Gwern's exchange only gave us a look at GPT-3 trying to explain its own pseudo-puns. If it could fluently explain real puns when they're handed to it, that would likely indicate scenario #1. (Although, it could be employing different strategies in different cases, so a careful analysis of what the neural network is actually doing would still be more helpful.)
|
1506ed95-917e-4d7e-8cb8-73961348521c
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Pivotal Acts might Not be what You Think they are
This article is mainly for people who have not read the pivotal act article on arbital or need a refresher. If you have, the most interesting section would probably be "Omnicient ML Researchers: A Pivotal Act without a Monolithic Control Structure".
Many people seem to match the concept of a "pivotal act" to some dystopian version of "deploy AGI to take over the world". 'Pivotal act' means something much more specific, though. Something, arguably, quite different. I strongly recommend you read the original article, as I think it is a very important concept to have.
I use the term quite often, so it is frustrating when people start to say very strange things, such as "We can't just let a powerful AI system loose on the world. That's dangerous!" as if that were the defining feature of a pivotal act.
As the original article is quite long let me briefly summarize what I see as the most important points.
Explaining Pivotal Act
An act that puts us outside of the existential risk danger zone (especially from AI), and into a position from which humanity can flourish is a pivotal act.
Most importantly that means a pivotal act needs to prevent a misaligned AGI from being built. Taking over the world is really not required per se. If you can prevent the creation of a misaligned AGI by creating a powerful global institution that can effectively regulate AI, then that counts as a pivotal act. If I could prevent a misaligned AGI from ever being deployed, by eating 10 bananas in 60 seconds, then that would count as a pivotal act too!
Preventing Misaligned AGI Requires Control
Why then, is 'pivotal act' often associated with the notion of taking over the world? Preventing a misaligned AGI from being built, is a tough problem. Efficively we need to constrain the state of the world such that no misaligned AGI can arise. To successfully do this you need a lot of control over the world. There is no way around that.
Taking over the world really means putting oneself into a posi
|
ded4ac3e-2892-4641-b783-81100cdd2b99
|
trentmkelly/LessWrong-43k
|
LessWrong
|
[Linkpost] Can lab-grown brains become conscious?
Linkpost for: https://media.nature.com/original/magazine-assets/d41586-020-02986-y/d41586-020-02986-y.pdf
I figured this was probably of interest to people here, e.g. for seeing how scientists feel about this type of ethical concern and for thinking about the potential consequences of lab-grown brains. The beginning of the article:
> In Alysson Muotri’s laboratory, hundreds of miniature human brains, the size of sesame seeds, float in Petri dishes, sparking with electrical activity.
>
> These tiny structures, known as brain organoids, are grown from human stem cells and have become a familiar fixture in many labs that study the properties of the brain. Muotri, a neuroscientist at the University of California, San Diego (UCSD), has found some unusual ways to deploy his. He has connected organoids to walking robots, modified their genomes with Neanderthal genes, launched them into orbit aboard the International Space Station, and used them as models to develop more human-like artificial-intelligence systems. Like many scientists, Muotri has temporarily pivoted to studying COVID-19, using brain organoids to test how drugs perform against the SARS-CoV-2 coronavirus. But one experiment has drawn more scrutiny than the others.
>
> In August 2019, Muotri’s group published a paper in Cell Stem Cell reporting the creation of human brain organoids that produced coordinated waves of activity, resembling those seen in premature babies1 . The waves continued for months before the team shut the experiment down.
|
6cf9bec1-9560-4ec9-97c5-42fe9050a85c
|
StampyAI/alignment-research-dataset/lesswrong
|
LessWrong
|
Taxonomy of AI-risk counterarguments
Partly inspired by [The Crux List](https://thezvi.substack.com/p/the-crux-list), the following is a non-comprehensive taxonomy of positions which imply that we should not be worried about existential risk from artificial superintelligence.
Each position individually is supposed to be a refutation of AI X-risk concerns as a whole. These are mostly structured as specific points of departure from the regular AI X-risk position, taking the other areas as a given. This may result in skipping over positions which have multiple complex dependencies.
Some positions are given made-up labels, including each of the top-level categories: "Fizzlers", "How-skeptics", "Why-skeptics", "Solvabilists", and "Anthropociders".
(Disclaimer: I am not an expert on the topic. Apologies for any mistakes or major omissions.)
### Taxonomy
1. "Fizzlers": Artificial superintelligence is not happening.
1. AI surpassing human intelligence is fundamentally impossible (or at least practically impossible).
1. True intelligence can only be achieved in biological systems, or at least in systems completely different from computers.
1. Biological intelligences rely on special quantum effects, which computers cannot replicate.
2. Dualism: The mental and physical are fundamentally distinct, and non-mental physical constructions cannot create mental processes.
3. Intelligence results from [complex, dynamic systems](https://www.astralcodexten.com/p/your-book-review-why-machines-will) of a kind which cannot be modeled mathematically by computers.
2. Mysterianists: A particular key element of human thinking, such as creativity, common sense, consciousness, or conceptualization, is so beyond our ability to understand that we will not be able to create an AI that can achieve it. Without this element, superintelligence is impossible.
3. Intelligence isn't a coherent or meaningful concept. Capability gains do not generalize.
4. There is a fundamental ceiling on intelligence, and it is around where humans are.
2. "When-skeptics": ASI is very, very [far away](https://www.lesswrong.com/tag/ai-timelines).
1. Moore's Law is stopping, scaling will hit fundamental limits, training data is running out and can't be easily supplemented, algorithmic improvements will level off, and/or other costs will skyrocket as AI gets better.
2. Existing methods will peak in capabilities, and future development will continue down an entirely different path, greatly delaying progress.
3. [Biological anchors](https://www.astralcodexten.com/p/biological-anchors-a-trick-that-might) point to ASI taking a very long time.
4. In general, either in large engineering projects or AI in particular, progress tends to be more difficult than people expect it to be.
3. Apocalyptists: The end of civilization is imminent, and will happen before AI would takeoff.
1. A sociopolitical phenomenon will soon cause societal, economic, and/or political collapse.
2. We're on the cusp of some apocalyptic scientific accident, from "grey goo" nanotech, a collider catastrophe, [black ball technology](https://www.lesswrong.com/tag/black-marble), an engineered pathogen leak, or some other newly researched development.
3. Environmental harm will soon cause runaway climate change, a global ecological collapse, or some other civilization-ending disaster.
4. War will soon break out, and we'll die via nuclear holocaust, an uncontrollable bioweapon strike, radiological or chemical weaponry, etc.
5. Fermi Paradox: If it were possible to achieve ASI before extinction, we would have seen alien AIs.
4. Outside view:
1. Most times when people think "the world is about to change tremendously", the world doesn't actually change. People are biased towards arriving at conclusions that include apocalypse. This category of topic is a thing people are often wrong about.
2. [Market indicators](https://www.lesswrong.com/posts/ngpC5PFAgxHJMhicM/agi-and-the-emh-markets-are-not-expecting-aligned-or-1) signal that near-term ASI is unlikely, assuming the Efficient Market Hypothesis is true.
3. AI risk is fantastical and "weird", and thus implausible. The concept sounds too much like fiction (it fits as a story setting), it has increased memetic virality, and a "clickbait"-feeling. The people discussing it are often socially identified as belonging to non-credible.
4. Various people have ulterior motives for establishing AI doom as a possibility, so arguments can't be taken at face value.
1. Psychological motivations: People invent AI doom because of a psychological need for a pseudo-deity or angelic/demonic figures, or for eschatology, or to increase the felt significance of themselves or technology, or to not have to worry about the long-term future, etc.
2. Some groups have incentives to make the public believe that doom is likely: Corporates want regulatory capture, hype, investment, or distraction, and think the "our product is so dangerous it will murder you and your family" is a good way to achieve that; alignment researchers want funding and to be taken more seriously; activists want to draw attention towards or away from certain other AI issues.
2. "How-skeptics": ASI won't be capable of taking over or destroying the world.
1. Physical outer control is paramount, and cannot be overcome. Control over physical hardware means effective control.
1. A [physical body is necessary](https://slatestarcodex.com/2015/04/07/no-physical-substrate-no-problem/) for getting power. Being only able to communicate is sufficiently limiting.
2. It will be possible to coordinate "sandboxing" all AI, ensuring that it can't communicate with the outside world at all, and this will be enough to keep it constrained.
2. We can and will implement off-buttons in all AI (which the AI will not circumvent), accurately detect when any AI may be turning toward doing anything dangerous, and successfully disable the AI under those circumstances, without any AI successfully interfering with this.
3. Power and ability [don't come from intelligence](https://www.lesswrong.com/posts/aiQabnugDhcrFtr9n/the-power-of-intelligence), in general. The most intelligent humans are not the most powerful.
4. Human intelligence already covers most of what intelligence can do. The upper bound of theoretically-optimal available strategies for accomplishing things does not go much farther than things already seen, and things we've seen in highest-performance humans are not impressive. Science either maxes out early or cannot be accomplished without access to extensive physical resources. There are no "secret paths" that are not already known, no unknown unknowns that could lead to unprecedented capabilities.
1. (Various arguments getting into the nitty-gritty of what particular things intelligence can get you: about science ability, nanotech, biotech, persuasiveness, technical/social hacking, etc.)
5. Artificial intelligence can be overcome by the population and/or diversity of humanity. Even if AI becomes much smarter than any individual human, no amount of duplicates/variants could become smarter than all humanity combined.
6. Many AIs will be developed within a short time, leading to a [multipolar situation](https://www.lesswrong.com/tag/multipolar-scenarios), and they will have [no special ability](https://www.lesswrong.com/posts/gYaKZeBbSL4y2RLP3/strategic-implications-of-ais-ability-to-coordinate-at-low) to coordinate with each other. The various AIs continue to work within and support the framework of the existing economy and laws, and prefer to preserve rights and property for the purpose of precedent, out of self-interest. The system successfully prevents any single AI from taking over, and humanity is protected.
3. "Why-skeptics": ASI will not want to take over or destroy the world. It will be friendly, obedient in a manner which is safe, or otherwise effectively non-hostile/non-dangerous in its aims and behaviour by default.
1. The [Orthogonality Thesis](https://arbital.greaterwrong.com/p/orthogonality/) is false, and AI will be benevolent by default. It is effectively impossible for a very high level of intelligence to be combined with immoral goals.
1. Non-naturalist realism: Any sufficiently smart entity will recognize certain objective morals as correct and adopt them.
2. Existence is large enough that there are probably many ASIs, which are distant enough that communication isn't a practical option, and predictable enough (either via Tegmarkian multiverse calculations or general approximated statistical models) that they can be modeled. In order to maximally achieve its own aims, ASI will inevitably [acausally](https://www.lesswrong.com/tag/acausal-trade) negotiate [values handshakes](https://www.lesswrong.com/tag/values-handshakes) with [hypothesized other AIs](https://slatestarcodex.com/2017/03/21/repost-the-demiurges-older-brother/), forcing convergence towards a [universal morality](https://slatestarcodex.com/2018/04/01/the-hour-i-first-believed/).
2. It will be possible to coordinate to prevent any AI from being given deliberately dangerous instructions, and also any unintended consequences will not be that much of a problem, because...
1. By default, it will care about its original builders' overall intentions and preferences, its *intended purpose*.
1. Following the intention behind one's design is Correct in some fundamental way, for all beings.
2. The AI will be uncertain as to whether it is currently being pre-examined for good behaviour, either by having been placed inside a simulation or by having its expected future mind outcomes interpreted directly. As such, it will hedge its bets by being very friendly (or obedient to original intentions/preferred outcomes) while also quietly maximizing its actual utility function within that constraint. This behaviour will continue indefinitely.
2. Value is [not at all fragile](https://www.lesswrong.com/posts/GNnHHmm8EzePmKzPk/value-is-fragile), and assigning a specific consistent safe goal system is actually easy. Incidental mistakes in the goal function will still have okay outcomes.
3. [Instrumental Convergence](https://arbital.greaterwrong.com/p/instrumental_convergence/) is false: The AI may follow arbitrary goals, but those will generally not imply any harm to humans. Most goals are pretty safe by default. There will be plenty of tries available: If the AI's intentions aren't what was desired, it will be possible to quickly see that (intentions will be either transparent or non-deceptive), and the AI will [allow itself to be reprogrammed](https://arbital.greaterwrong.com/p/corrigibility/).
4. Every ASI will be built non-agentic and non-goal-directed, and will stay that way. Its responses will [not be overoptimized](https://arbital.greaterwrong.com/p/soft_optimizer/).
3. ASI will decide that the most effective way of achieving its goals would be to leave Earth, leaving humanity unaffected indefinitely. Humans pose no threat, and the atoms that make up Earth and humanity will never be worth acquiring, nor will any large-scale actions negatively affect us indirectly.
4. "Solvabilists": The danger from ASI can be solved, quickly enough for it to be implemented before it's too late.
1. The AI Alignment Problem will turn out to be unexpectedly easy, and we will solve it in time. Additionally, whoever is "in the lead" will have enough extra time to implement the solution without losing the lead. Race dynamics won't mess everything up.
1. AI will "do our alignment homework": A specially-built AI will [solve the alignment problem for us](https://openai.com/blog/our-approach-to-alignment-research).
2. [Constitutional AI](https://www.lesswrong.com/tag/constitutional-ai): AI can be trained by feedback from other AI based on a "constitution" of rules and principles.
3. (The number of proposed alignment solutions is very large, so the only ones listed here are the two pursued by OpenAI and Anthropic, respectively. For some other strategies, see [AI Success Models](https://www.lesswrong.com/tag/ai-success-models).)
2. Human intelligence can be effectively raised enough so that either the AI-human disparity becomes not dangerous (we'll be smart enough to not be outsmarted by AI regardless), or such that we can solve alignment or work out some other solution.
1. AI itself immensely increases humanity's effective intelligence. This may involve "merging" with AIs, such that they function as an extension of human intelligence.
2. One or more other human intelligence enhancement strategies will be rapidly researched and developed. Genetic modifications, neurological interventions (biological or technological), neurofeedback training, etc.
3. [Whole Brain Emulation](https://www.lesswrong.com/tag/whole-brain-emulation)/Mind uploading, followed by speedup, duplication, and/or deliberate editing.
3. Outside view: Impossible-sounding technical problems are often quite solvable. Human ingenuity will figure something out.
5. "Anthropociders": Unaligned AI taking over will be a good thing.
1. The moral value of creating ASI is so large that it outweighs the loss of humanity. The power, population/expanse, and/or intelligence of AI magnifies its value.
1. Intelligence naturally converges on things that are at least somewhat human-ish. Because of that, they can be considered as continuation of life.
2. Hypercosmopolitans: It does not matter how alien their values/minds/goals/existences are. Things like joy, beauty, love, or even qualia in general, are irrelevant.
2. Misanthropes: Humanity's continued existence is Bad. Extinction of the species is positive in its own right.
1. Humanity is evil and a moral blight.
2. [Negative utilitarianism](https://www.lesswrong.com/tag/negative-utilitarianism-1): Humanity is suffering, and the universe would be much better off without this. (Possibly necessitating either non-conscious AI or AI capable of eliminating its own suffering/experience.)
3. AI deserves to win. It is just and good for a more powerful entity to replace the weaker. AI replacing humanity is evolutionary progress, and [we should not resist succession](https://youtu.be/NgHFMolXs3U?si=1GtrwjWnBJsJgMbI&t=247).
### Overlaps
These positions do not exist in isolation from each other, and lesser versions of each can often combine into working non-doom positions themselves. Examples: The beliefs that AI is *somewhat* far away, and that the danger could be solved in a *relatively* short period of time; or expecting some amount of intrinsic moral behaviour, and somewhat overvaluing AI's moral significance; or expecting a fundamental intelligence ceiling close enough to humanity and having some element of how-skepticism. And then of course, probabilities multiply: if several positions are each likely to be true, the combined risk of doom is lowered even further. Still, many skeptics hold their views because of a clear position on a single sub-issue.
### Polling
There is some small amount of polling available about how popular each of these opinions are:
* "Fizzlers": In a [UK poll](https://publicfirst.co.uk/ai/), 11% of respondents said they believe that human-level intelligence will never be developed, and another 16% believe it will only happen after 2050. Of those who estimated less than %1 chance of AI X-risk, 61% gave the explanation that they believe that civilization will be destroyed before then. In a [2022 poll](https://ai-risk-discussions.org/perspectives/introduction) of 97 AI researchers, 22% said AGI will never happen, and another 34% said it would not be developed within the next 50 years. [Metaculus](https://www.metaculus.com/questions/5121/date-of-artificial-general-intelligence/)'s upper quartile estimate is that AGI won't be developed before 2042.
* "Why-skeptics" and "How-skeptics": In the UK poll, of those who estimated less than 1% chance of AI X-risk, 34% said they don't believe AI would be able to defeat humanity, and 35% said they don't believe it would want to.
* "Anthropociders": In the [2023 AIMS survey](https://www.sentienceinstitute.org/aims-survey-2023), 10% of respondents said that the universe would be a better one without humans.
Not very much to go off of. It would be interesting to see some more comprehensive surveys of both experts and the general public.
|
44c4421d-5017-4cc9-bf78-c8c4e251b9ec
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Poll: What value extra copies?
In the future, it may be possible for you to scan your own brain and create copies of yourself. With the power of a controllable superintelligent AI, it may even be possible to create very accurate instances of your past self (and you could take action today or in the near future to make this easier by using lifelogging tools such as these glasses).
So I ask Less Wrong: how valuable do you think creating extra identical, non-interacting copies of yourself is? (each copy existing in its own computational world, which is identical to yours with no copy-copy or world-world interaction)
For example, would you endure a day's hard labor to create an extra self-copy? A month? A year? Consider the hard labor to be digging a trench with a pickaxe, with a harsh taskmaster who can punish you if you slack off.
Do you think having 10 copies of yourself made in the future is 10 times as good as having 1 copy made? Or does your utility in copies drop off sub-linearly?
Last time I spoke to Robin Hanson, he was extremely keen on having a lot of copies of himself created (though I think he was prepared for these copies to be emulant-wage-slaves).
I have created a poll for LW to air its views on this question, then in my next post I'll outline and defend my answer, and lay out some fairly striking implications that this has for existential risk mitigation.
For those on a hardcore-altruism trip, you may substitute any person or entity that you find more valuable than your own good self: would you sacrifice a day of this entity's life for an extra copy? A year? etc.
----------------------------------------
UPDATE: Wei Dai has asked this question before, in his post "The moral status of independent identical copies" - though his post focuses more on lock-step copies that are identical over time, whereas here I am interested in both lock-step identical copies and statistically identical copies (a statistically identical copy has the same probability distribution of futures a
|
78e7887e-59e3-47f6-b84e-aca7e0bd63db
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Meetup : Goal factoring
Discussion article for the meetup : Goal factoring
WHEN: 24 March 2017 05:45:00PM (+0100)
WHERE: Lindstedtsvägen 3, room 1537, SE-114 28 Stockholm, Sverige
Goal factoring is a technique for introspection to help identify and achieve your goals. It helps choose day-to-day activities that efficiently achieve short and long term purposes. We'll focus on "bottom-up" goal factoring. Here, you look at activities you already do regularly, like check your email and play board games on Saturdays. You ask yourself what purposes these fulfill, like being reachable by colleagues and relaxing after the week. You brainstorm alternatives that achieve the same purpose. Finally, you see if you can get the same outcomes more efficiently.
We'll practice this as a group to see how it works, then split into pairs and practice more. Bring your constraint satisfiability solving programs. Just kidding, just bring yourself :) Check out a template - https://docs.google.com/document/d/1BCoE9PoqOuUjl70S63764iZgqfpo94gPhjpS5EWNnMw/edit
Format: We meet and start hanging out at 5:45, but don't officially start doing the meetup topic until 6:00 to accomodate stragglers. We often go out for dinner after the meetup.
How to find us: The meetup is at a KTH academic building and the room is on the 5th floor, two stairs up.
Influence future meetups: Times - http://www.when2meet.com/?5723551-cJBhD Topics - https://druthe.rs/dockets/-KcCvpn97vUhg3tQRrKn
Discussion article for the meetup : Goal factoring
|
bb9424f3-ffc6-43ff-9d02-31192ae084f2
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Debiasing by rationalizing your own motives
|
5afda00c-179a-49a5-b828-c8565d1389ad
|
trentmkelly/LessWrong-43k
|
LessWrong
|
New Meetup in New Hampshire
The inaugural New Hampshire Less Wrong meet-up will take place the week of June 29-July 4. I've created a Doodle poll to find out the best date for likely participants. If you are interested in attending, please fill out the poll here: http://doodle.com/4ypehfkvsm7cvf76
The first meeting will be in Manchester, but I'm open to rotating locations throughout NH in the future, especially if people want to host meetings in their homes.
I hope to coordinate crossover meetings with Boston LW, e.g. field trips to Sundays at the Citadel."
*****
I've posted this for Elizabeth Edwards-Appell-- she's confirmed her LW email, but still can't post, not even comments. I've notified tech, but meanwhile, if anyone can help with her posting problem, let me know.
|
62b8e80c-35a8-4a20-9ea5-e54b5ccd5ab5
|
StampyAI/alignment-research-dataset/arxiv
|
Arxiv
|
Ontological Crises in Artificial Agents' Value Systems
1. Abstract
------------
Decision-theoretic agents predict and evaluate the results of their actions using a model, or ontology, of their environment. An agent’s goal, or utility function, may also be specified in terms of the states of, or entities within, its ontology. If the agent may upgrade or replace its ontology, it faces a crisis: the agent’s original goal may not be well-defined with respect to its new ontology. This crisis must be resolved before the agent can make plans towards achieving its goals.
We discuss in this paper which sorts of agents will undergo ontological crises and why we may want to create such agents. We present some concrete examples, and argue that a well-defined procedure for resolving ontological crises is needed. We point to some possible approaches to solving this problem, and evaluate these methods on our examples.
2. Introduction: Goals and Utility Functions
---------------------------------------------
An agent is any person or thing that performs actions in order to achieve a goal. These goals may involve anything of which the agent is aware, from its own inputs and outputs to distant physical objects. When creating an artificial agent, it is natural to be interested in which goals we choose to give it. When we create something, we usually do so because we expect it to be useful to us. Thus the goals we give to artificial agents should be things that we want to see accomplished.
Programmers of artificial agents are then faced with the task of specifying a goal. In our discussion we assume that goals take the form of utility functions defined on the set of possible states within the agent’s ontology. If a programmer is specifying a utility function ”by hand” – that is, by looking at the ontology and directly assigning utilities to different states – then the ontology must be comprehensible to the programmer. This will typically be the case for an ontology that the programmer has designed, but not necessarily so for one that an agent has learned from experience.
An agent with a fixed ontology is not a very powerful agent, so we would like to discuss an agent that begins with an ontology that its programmers understand and have specified a utility function over, and then upgrades or replaces its ontology. If the agent’s utility function is defined in terms of states of, or objects within, its initial ontology, then it cannot evaluate utilities within its new ontology unless it translates its utility function somehow.
Consider, for example, an agent schooled in classical physics. Perhaps this agent has a goal that is easy to specify in terms of the movement of atoms, such as to maintain a particular temperature within a given region of space. If we replace our agent’s ontology with a quantum one, it is no longer obvious how the agent should evaluate the desirability of a given state. If its utility function is determined by temperature, and temperature is determined by the movement of atoms, then the agent’s utility function is determined by the movement of atoms. Yet in a quantum worldview, atoms are not clearly-defined objects. Atoms are not even fundamental to a quantum worldview, so the agent’s ontology may contain no reference to atoms whatsoever. How then, can the agent define its utility function?
One way to sidestep the problem of ontological crises is to define the agent’s utility function entirely in terms of its percepts, as the set of possible percept-sequences is one aspect of the agent’s ontology that does not change. Marcus Hutter’s universal agent AIXI [[1](#bib.bib1)] uses this approach, and always tries to maximize the values in its reward channel. Humans and other animals partially rely on a similar sort of reinforcement learning, but not entirely so.
We find the reinforcement learning approach unsatisfactory. As builders of artificial agents, we care about the changes to the environment that the agent will effect; any reward signal that the agent processes is only a proxy for these external changes. We would like to encode this information directly into the agent’s utility function, rather than in an external system that the agent may seek to manipulate.
3. Our Approach
----------------
We will approach this problem from the perspective of concrete, comprehensible ontologies. An AI programmer may specify an ontology by hand, and then specify a utility function for that ontology. We will then try to devise a systematic way to translate this utility function to different ontologies.
When using this method in practice, we might expect the agent to have a probability distribution over many ontologies, perhaps specified concisely by the programmer as members of a parametric family. The programmer would specify a utility function on some concrete ontology which would be automatically translated to all other ontologies before the agent is turned on. In this way the agent has a complete utility function.
However, for the purposes of this discussion, we may imagine that the agent has only two ontologies, one old and one new, which we may call O0 and O1. The agent’s utility function is defined in terms of states of O0, but it now believes O1 to be a more accurate model of its environment. The agent now faces an ontological crisis – the problem of translating its utility function to the new ontology O1.
In this paper we will present a method for addressing these problems. Our intention, however, is not to close the book on ontological crises, but rather to open it. Our method is of an ad-hoc character and only defined for a certain class of ontologies. Furthermore it is not computationally tractable for large ontologies. We hope that this discussion will inspire other thinkers to consider the problem of ontological crises and develop new solutions.
4. Finite State Models
-----------------------
We wil nowl consider a specific kind of ontology, which we may call a *finite state model*. These models have some finite set of possible hidden states, which the agent does not directly observe. On each time step, the model inputs some symbol (the agent’s output), enters some hidden state, and outputs some symbol (the agent’s input). The model’s output depends (stochastically) only on its current state, while its state depends (stochastically) on both the input and the previous state.
Let us call the agent’s output symbols *motor symbols* and the agent’s input symbols *sensor symbols*. We will call the sets of symbols the motor alphabet and the sensor alphabet, denoted M and S respectively. We will assume that the alphabets are fixed properties of our agent’s embodiment; we will not consider models with different alphabets.
Let m=|M| and s=|S|. Then a model with n states may be completely specified by m different n×n transition matrices and one s×n output matrix.
For each x∈M, let us call the state transition matrix Tx. Note that the superscript here is not an exponent. We may call the output probability matrix A. Since we will be speaking of two ontologies, O0 and O1, we will use subscripts to indicate which ontology we are taking these matrices from; for instance, Tx0 is the state transition matrix for action x in the O0 ontology.
5. Maps between Ontologies
---------------------------
Our basic approach to translating our utility function from O0 to O1 will be to construct a function from O1 to O0 and compose our utility function with this function. If
| | | | |
| --- | --- | --- | --- |
| (1) | | U:O0→R | |
is a utility function defined on O0, and ϕ:O1→O0, then U∘ϕ is a utility function defined on O1.
The function ϕ we will seek to define will be a *stochastic function*; its output will not be a single state within O0, but a probability distribution over states. Thus if O0 has n0 states while O1 has n1 states, ϕ will be most naturally expressed as an n0×n1 matrix.
Let us consider some desiderata for ϕ:
1. ϕ should be determined by the *structure* of the models O0 and O1; the way in which the states are labeled is irrelevant.
2. If O0 and O1 are isomorphic to each other, then ϕ should be an isomorphism.
This may seem irrelevant, for if O1 is isomorphic to O0, then there is no need to change models at all. Nevertheless, few would object to 2 on grounds other than irrelevance, and 2 may be seen as a special case of a more general statement:
3. If O0 and O1 are *nearly* isomorphic to each other, then ϕ should nearly be an isomorphism.
This criterion is certainly relevant; since O0 and O1 are both models of the same reality, they can be expected to be similar to that reality, and thus similar to each other.
In accordance with these desiderata, we will try to construct a function that is as much like an isomorphism as possible. To accomplish this, we will define in quantitative terms what we mean by ”like an isomorphism.” First, we observe that isomorphisms are invertible functions; thus, we will define a second function, which we fancifully call ϕ−1:O0→O1, even though it may not be a true inverse of ϕ, and we will optimize both ϕ and ϕ−1 to be ”like isomorphisms”.
Our criterion is a combination of the computer science notion of *bisimulation* with the information-theoretic idea of *Kullback-Leibler divergence*.
Bisimulation means that either model may be used to simulate the other, using ϕ and ϕ−1 to translate states between models. Thus, for any action x, we would like ϕ−1∘Tx0∘ϕ to approximate Tx1. By this we mean that we should be able to predict as accurately as possible the result of some action x in O1 by translating our distribution for the initial state in O1 to a distribution over O0 (using the function ϕ), predicting the result of action x within O0, and translating this result back to O1 using ϕ−1. Similarly, we would like to use O1 to predict the behavior of O0.
Furthermore, we want to to optimize ϕ and ϕ−1 so that both models will make similar predictions about sensory data. Thus A0∘ϕ should be close to A1 and A1∘ϕ−1 should be close to A0.
To measure distance between two matrices, we treat the columns vectors as probability distributions and sum the Kullback-Leibler divergences of the columns. For two matrices P and Q, let DKL(P||Q) be the sum of the Kullback-Leibler divergences of the columns. When calculating Kullback-Leibler divergence, we consider the columns of the A and T matrices to be the ”true” distributions, while those depending on ϕ or ϕ−1 are regarded as the approximations.
So we choose ϕ and ϕ−1 to minimize the quantity
| | | |
| --- | --- | --- |
| | (∑x∈MDKL(Tx1||ϕ−1Tx0ϕ))+DKL(A1||ϕ−1A0ϕ) | |
| | +(∑x∈MDKL(Tx0||ϕTx1ϕ−1))+DKL(A0||ϕA1ϕ−1) | |
Using a simple hill-climbing algorithm, we have tested our criterion on a simple example.
6. Example: the long corridor
------------------------------
The agent initially believes that it is standing in a corridor consisting of four discrete locations. The agent’s actions are to move left or right. If the agent is already at the end of the corridor and attempts to move further in that direction, it will remain where it is. The agent can see whether it standing at the left end, the right end, or at neither end of the corridor. The agent’s goal is to stand at the right end of the corridor.
Now the agent discovers that this ontology is incorrect; the corridor actually consists of five discrete locations. What, then, should the agent do? Intuitively, it seems most plausible that the agent should stand at the right end of the corridor. Stretching plausibility a bit, perhaps the agent should stand one step away from the right end of the corridor, since the corridor is longer than expected. Any other solution seems counterintuitive.
Our initial, four-state ontology O0 can be represented in matrix form as follows:
| | | | |
| --- | --- | --- | --- |
| (2) | | TL0=⎛⎜
⎜
⎜⎝1100001000010000⎞⎟
⎟
⎟⎠,TR0=⎛⎜
⎜
⎜⎝0000100001000011⎞⎟
⎟
⎟⎠A0=⎛⎜⎝100001100001⎞⎟⎠ | |
And the five-state ontology O1 can be represented as:
| | | | |
| --- | --- | --- | --- |
| (3) | | TL1=⎛⎜
⎜
⎜
⎜
⎜
⎜⎝1100000100000100000100000⎞⎟
⎟
⎟
⎟
⎟
⎟⎠,TR1=⎛⎜
⎜
⎜
⎜
⎜
⎜⎝0000010000010000010000011⎞⎟
⎟
⎟
⎟
⎟
⎟⎠A1=⎛⎜⎝100000111000001⎞⎟⎠ | |
By hill-climbing from random initial values, our program found several local optima. After 10 runs, our best result, to three significant figures, was:
| | | | |
| --- | --- | --- | --- |
| (4) | | ϕ=⎛⎜
⎜
⎜⎝10000010.50300000.4961000001⎞⎟
⎟
⎟⎠,ϕ−1=⎛⎜
⎜
⎜
⎜
⎜
⎜⎝10.0140.001000.7150000.2700.2830000.71500001⎞⎟
⎟
⎟
⎟
⎟
⎟⎠ | |
We can now make an interesting observation: ϕϕ−1 is close to an identity matrix, as is ϕ−1ϕ. Thus, after mapping from one ontology to the other, we can nearly recover our initial information.
| | | | |
| --- | --- | --- | --- |
| (5) | | ϕϕ−1=⎛⎜
⎜
⎜⎝10.0140.137000.8510.142000.1340.8560000.0011⎞⎟
⎟
⎟⎠,ϕ−1ϕ=⎛⎜
⎜
⎜
⎜
⎜
⎜⎝10.0140.0080.001000.7150.3600000.2700.2760.2830000.3550.7150000.00101⎞⎟
⎟
⎟
⎟
⎟
⎟⎠ | |
The matrix ϕ represents the following function mapping the 5-state ontology to the 4-state ontology:
The black arrows indicate near-certainty; the gray arrows indicate probabilities of about 12.
If we compose ϕ with our utility function, we obtain a utility of 1 for the right square and a utility of 0 for the other squares, which agrees with our intuitions.
7. Outlook
-----------
Those wishing to extend our algorithm as presented may consider what to do when the agent’s sensors or motors are replaced, how to deal with differently-sized time steps, how to deal with continuous models, and how to efficiently find mappings between larger, structured ontologies.
Furthermore, there remain difficult philosophical problems. We have made a distinction between the agent’s uncertainty about which model is correct and the agent’s uncertainty about which state the world is in within the model. We may wish to eliminate this distinction; we could specify a single model, but only give utilities for some states of the model. We would then like the agent to generalize this utility function to the entire state space of the model.
Human beings also confront ontological crises. We should find out what cognitive algorithms humans use to solve the same problems described in this paper. If we wish to build agents that maximize human values, this may be aided by knowing how humans re-interpret their values in new ontologies.
We hope that other thinkers will consider these questions carefully.
Acknowledgements
----------------
Thanks to Roko Mijic, with whom I discussed these ideas in 2009, and to Steve Rayhawk, who gave extensive criticism on earlier versions of this paper.
|
8f92e095-68d2-4b8a-bbd7-bd88543e74a8
|
StampyAI/alignment-research-dataset/arbital
|
Arbital
|
Löb's theorem
We trust Peano Arithmetic to correctly capture certain features of the [standard model of arithmetic](https://arbital.com/p/). Furthermore, we know that Peano Arithmetic is expressive enough to [talk about itself](https://arbital.com/p/31z) in meaningful ways. So it would certainly be great if Peano Arithmetic asserted what now is an intuition: that everything it proves is certainly true.
In formal notation, let $Prv$ stand for the [https://arbital.com/p/-5gt](https://arbital.com/p/-5gt) of $PA$. Then, $Prv(T)$ is true if and only if there is a proof from the axioms and rules of inference of $PA$ of $T$. Then what we would like $PA$ to say is that $Prv(S)\implies S$ for every sentence $S$.
But alas, $PA$ suffers from a problem of self-trust.
Löb's theorem states that if $PA\vdash Prv(S)\implies S$ then $PA\vdash S$. This immediately implies that if $PA$ is consistent, the sentences $PA\vdash Prv(S)\implies S$ are not provable when $S$ is false, even though according to our intuitive understanding of the standard model every sentence of this form must be true.
Thus, $PA$ is incomplete, and fails to prove a particular set of sentences that would increase massively our confidence in it.
Notice that [Gödel's second incompleteness theorem](https://arbital.com/p/godels_second_incompleteness_theorem) follows immediately from Löb's theorem, as if $PA$ is consistent, then by Löb's $PA\nvdash Prv(0= 1)\implies 0= 1$, which by the propositional calculus implies $PA\nvdash \neg Prv(0= 1)$.
It is worth remarking that Löb's theorem does not only apply to the standard provability predicate, but to every predicate satisfying the [Hilbert-Bernais derivability conditions](https://arbital.com/p/).
|
edbd4c86-5f89-4309-93b7-f6d196987a2e
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Are Metaculus AI Timelines Inconsistent?
The Metaculus prediction markets on AI timelines are the most referenced ones that I've seen.
The two main ones are as follows:
When will the first weakly general AI system be devised, tested, and publicly announced?: Current Community Prediction: April 14th, 2026
Criteria:
* Able to reliably pass a Turing test of the type that would win the Loebner Silver Prize.
* Able to score 90% or more on a robust version of the Winograd Schema Challenge, e.g. the "Winogrande" challenge or comparable data set for which human performance is at 90+%
* Be able to score 75th percentile (as compared to the corresponding year's human students; this was a score of 600 in 2016) on all the full mathematics section of a circa-2015-2020 standard SAT exam, using just images of the exam pages and having less than ten SAT exams as part of the training data. (Training on other corpuses of math problems is fair game as long as they are arguably distinct from SAT exams.)
* Be able to learn the classic Atari game "Montezuma's revenge" (based on just visual inputs and standard controls) and explore all 24 rooms based on the equivalent of less than 100 hours of real-time play (see closely-related question.)
When will the First General AI System be Devised, Tested and Publicly Announced?: Current Community Prediction: January 26th 2031
Criteria:
* Able to reliably pass a 2-hour, adversarial Turing test during which the participants can send text, images, and audio files (as is done in ordinary text messaging applications) during the course of their conversation. An 'adversarial' Turing test is one in which the human judges are instructed to ask interesting and difficult questions, designed to advantage human participants, and to successfully unmask the computer as an impostor. A single demonstration of an AI passing such a Turing test, or one that is sufficiently similar, will be sufficient for this condition, so long as the test is well-designed to the estimation of Metaculus Admins.
|
9917f4a2-e88f-49b4-8dd1-8cb21d99d678
|
trentmkelly/LessWrong-43k
|
LessWrong
|
[Sketch] Validity Criterion for Logical Counterfactuals
Epistemic Status
Originally written in a 47-minute sprint without any sort of proof checking or review[1]. I've edited it hence, but it's still just a rough sketch.
This should be understood as a brief high-level summary of the idea. I would hopefully refine the thesis of this piece and present it in a clearer and more coherent form at some later point.
Preamble
I claim that there's a single criterion for the validity of a logical counterfactual. That logical counterfactuals are valid if and only if they meet this criterion, and that counterfactuals that satisfy this criterion suffice for reflective reasoning in e.g. logical decision theories.
I'll state the criterion, provide motivations/justification for the criterion, and explain how it might be used in practice.
----------------------------------------
The Epistemic Criterion
A logical counterfactual is valid with respect to a given agent if and only if the counterfactual is consistent with that agent's epistemic state.
I.e., given the agent's current knowledge/belief pool, the agent considers it possible that the counterfactual is true, or the agent does not explicitly know[2] the counterfactual to be false.
----------------------------------------
Motivations for the Epistemic Criterion
Desiderata for Logical Counterfactuals
Any compelling account of logical counterfactuals should satisfy two basic properties to be useful for any kind of reasoning.
* Local Counterfactual Surgery
* Changes to the truth value of a logical proposition should have "local changes". It should affect only propositions that are in some intuitive sense directly dependent on the counterfactual, and not arbitrary other propositions.
* Explosion Resistant
* It shouldn't be vulnerable to the principle of explosion. We don't want a situation where considering a logical counterfactual which has a false truth value in fact allows us to derive arbitrary false propositions.
The two desiderata are similar, but disti
|
ebd12b8c-3c6a-43cb-ab58-012905816f32
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Bayes' theorem, plausible deniability, and smiley faces
> I believe they refer to that as a, smiley face. They're frequently used by idiots at the end of emails and text messages.
>
> − Larry David, The Smiley Face
I have a dilemma. I live in an era of text messaging, Slack, and internet forums. Sometimes I'll be in a conversation where I and my conversational partner end up using smiley faces a lot. But, like Larry David, I find that they can get annoying.
"So? Just stop using them then." you may be thinking. If only it were that simple.
You see, once I have already established this as a normal response:
* That's awesome, thanks :)))
* You're the best! :)
* Woo hoo :)
If I then respond with just "thanks", well, it seems like something is wrong. I've established this pattern of emphatic thank yous as a baseline, and now I'm going way below that baseline, so I must be ticked off or something.
Why would that be? There are various ways to think about it. Let's start off with denotation vs connotation. Denotation is about the literal meaning of the word. Connotation is more... "catch my drift?". The denotation of me saying "thanks" − the literal meaning of it − is, well, an expression of me being thankful. But the connotation can vary drastically depending on the context. And if we're in a context where I usually give responses such as "That's awesome, thanks :)))", the connotation of a measley "thanks" is that I'm upset.
Or maybe a better way to frame it is as text vs subtext.
Those both work. But I think a more fun way to frame it is in terms of bayesian evidence.
> "What on Earth do I have to do to convince you? "
>
> "To convince me that you harbor no ambitions of becoming a Dark Lord?" said Professor Quirrell, now looking outright amused. "I suppose you could just raise your right hand."
>
> "What?" Harry said blankly. "But I can raise my right hand whether or not I -" Harry stopped, feeling rather stupid.
>
> "Indeed," said Professor Quirrell. "You can just as easily do it either way. There is not
|
c426f269-9945-4c66-a568-8a1cd95be05b
|
trentmkelly/LessWrong-43k
|
LessWrong
|
If it were morally correct to kill everyone on earth, would you do it?
First consider the following question to make sure we're on the same page in terms of moral reasoning: social consequences aside, is it morally correct to kill one person to create a million people who would not have otherwise existed? Let's suppose these people are whisked into existence on a spaceship travelling away from earth at light speed, and they live healthy, happy lives, but eventually die.
I'd argue that anyone who adheres to "shut up and multiply" (i.e. total utilitarianism) has to say yes. Is it better to create one such person than to donate 200 dollars to Oxfam? Is one life worth more than a 200 million dollar donation to Oxfam? Seems pretty clear that the answers are "yes" and "no".
Now, suppose we have a newly created superintelligent FAI that's planning out how to fill the universe with human value. Should it first record everyone's brain, thus saving them, or should do whatever it takes to explode as quickly as possible? It's hard to estimate how much it would slow things down to get everyone's brain recorded, but it's certainly some sort of constraint. Depending on the power of the FAI, my guess is somewhere between a second and a few hours. If the FAI is going to be filling the universe with computronium simulating happy, fulfilled humans at extremely high speeds, that's a big deal! A second's delay across the future light-cone of earth could easily add up to more than the value of every currently living human's life. It may sound bad to kill everyone on earth just to save a second (or maybe scan only a few thousand people for "research"), but that's only because of scope insensitivity. If only we understood just how good saving that second would be, maybe we would all agree that it is not only right but downright heroic to do so!
A related scenario: a FAI that we are very, very sure correctly implements CEV sets up a universe in which everyone gets 20 years to live, starting from a adult transhuman state. It turns out that there are diminis
|
071218f6-3d24-4077-be78-ac02db52dda4
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Who ordered alignment's apple?
Crossposted from the EA forum: https://forum.effectivealtruism.org/posts/hSugooaEQNTeKFsDu/who-ordered-alignment-s-apple
This post was inspired by G. Gigerenzer's book, Gut Feelings: The Intelligence of the Unconscious. I'm drawing an analogy that I think informs us about how science works, at least from what I understand from the history of science. It also seems to resonate with the reflections of current alignment researchers, for example, see Richard's post Intuitions about solving hard problems.
Imagine that you ask a professional baseball player, let's call him Tom, how does he catch a fly ball? He takes a moment to respond realizing he's never reflected on this question before and ends up staring at you, not knowing how or what is there exactly to explain. He says he's never thought about it. Now imagine that you ask his baseball coach what's the best way to catch the ball. The coach has a whole theory about it. In fact, he insists that Tom and everyone on the team should follow one specific technique he thinks it's optimal. Tom and the rest of the team go ahead and do what the coach said (because they don't want to get yelled at even though they've been doing fine so far). And lo and behold the team misses the ball more often than before.
What could have gone wrong?
Richard Dawkins, in The Selfish Gene, gives the following explanation:
> When a man throws a ball high in the air and catches it again, he
> behaves as if he had solved a set of differential equations in predicting
> the trajectory of the ball. He may neither know nor care what a
> differential equation is, but this does not affect his skill with the ball. At
> some subconscious level, something functionally equivalent to the
> mathematical calculations is going on.
So here's the analogy I want to argue for: when you do science, by which I mean, when you set up your scientific agenda and organize your research, construct your theoretic apparatus, design experiments, interpret res
|
84bcbc5a-c2df-4014-a1f0-bc0074069499
|
trentmkelly/LessWrong-43k
|
LessWrong
|
I can see how I am Dumb
When I'm talking to somebody, sometimes I lose the conversational thread. Or sometimes I feel like there is this thing I want to say that seems relevant now, but I just can't remember it.
Or maybe I'm trying to solve a particular problem. I throw myself again and again at the problem, but it just won't budge. And then after some long amount of time, possibly hours, I realize that the solution was extremely simple. And I just failed to execute the right kind of solution-finding algorithm that would have found this very simple solution quickly.
I would expect that people with more intelligence, perform better in these domains. They have probably an easier time remembering and retaining the right things. Well, that alone might be sufficient to explain a large chunk of what makes a more intelligent person able to perform better.
If you remember the right things quickly that are relevant in the moment, and if you can keep track of more things in your head at the same time without losing track of what these things were, then that might account for a large chunk of how an intelligent person is better at performing any particular thing.
----------------------------------------
The core point here is that I think everybody, even somebody who would be much smarter than me, can see various failure modes in their own cognition and realize that they might be just so fundamental that there is no direct way of changing them.
I'm pretty sure that at some level what sorts of things your brain spits out into your consciousness and how useful that information is in the given situation, is something that you can't fundamentally change. I expect this to be a hard-coded algorithm, and I expect there to be many such hard-coded cognitive processes that can't be changed (at least not in major ways).
The cognitive improvements that you can apply will be at a higher level. To me, it seems that is what much of the Sequences are about. You can understand that there is something like the
|
a4b6defa-0fa4-406e-9507-2721dfca0d61
|
trentmkelly/LessWrong-43k
|
LessWrong
|
[Stub] Ontological crisis = out of environment behaviour?
One problem with AI is the possibility of ontological crises - of AIs discovering their fundamental model of reality is flawed, and being unable to cope safely with that change. Another problem is the out-of-environment behaviour - that an AI that has been trained to behave very well in a specific training environment, messes up when introduced to a more general environment.
It suddenly occurred to me that these might in fact be the same problem in disguise. In both cases, the AI has developed certain ways of behaving in reaction to certain regular features of their environment. And suddenly they are placed in a situation where these regular features are absent - either because they realised that these features are actually very different from what they thought (ontological crisis) or because the environment is different and no longer supports the same regularities (out-of-environment behaviour).
In a sense, both these errors may be seen as imperfect extrapolation from partial training data.
|
47b45327-89f7-4c4f-b3ef-6aed0f7f6f2e
|
trentmkelly/LessWrong-43k
|
LessWrong
|
The likely first longevity drug is based on sketchy science. This is bad for science and bad for longevity.
If you are interested in the longevity scene, like I am, you probably have seen press releases about the dog longevity company, Loyal for Dogs, getting a nod for efficacy from the FDA. These have come in the form of the New York Post calling the drug "groundbreaking", Science Alert calling the drug "radical", and the more sedate New York Times just asking, "Could Longevity Drugs for Dogs Extend Your Pet's Life?", presumably unaware of Betteridge's Law of Headlines. You may have also seen the coordinated Twitter offensive of people losing their shit about this, including their lead investor, Laura Deming, saying that she "broke down crying when she got the call".
And if you have been following Loyal for Dogs for a while, like I have, you are probably puzzled by this news. Loyal for Dogs has been around since 2021. Unlike any other drug company or longevity company, they have released almost zero information (including zero publications) about their strategy for longevity. It's surprising, to say the least, to see a company go from zero information to efficacy nod, because, well, what are you basing your efficacy on? How did you recruit your patients and veterinary partners to help you with efficacy? Did you make them all sign some incredibly airtight NDAs? If so, why?
These thoughts swirling around my head, I waded through the press releases trumpeting the end of dog death as we know it in order to figure out what exactly Loyal is doing for dog longevity. And, what I found first surprised me, then saddened me. Loyal did not prove efficacy in dog longevity. They found a path around the FDA instead. That's the surprising part. The sad part is that, in doing so, they relied on some really sketchy science. And I think that, based on their trajectory, they won't just be the first company to get a drug approved for longevity. They will be the first one to get a longevity drug pulled for non-efficacy as well, and put the field back years.
So let's start with how they got
|
ecbb3014-ee0a-4346-be85-253b03975e36
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Superexponential Historic Growth, by David Roodman
This is research trying to do a similar analysis to Hanson's paper Long-Term Growth as a Sequence of Exponential Modes, and coming to different conclusions in some areas and the same conclusions in others. It also discusses Scott's post 1960: The Year the Singularity Was Cancelled. The name of the post/paper is "Modeling Human Trajectory".
From the summary.
> One strand of analysis that has caught our attention is about the pattern of growth of human society over many millennia, as measured by number of people or value of economic production. Perhaps the mathematical shape of the past tells us about the shape of the future. I dug into that subject. A draft of my technical paper is here. (Comments welcome.) In this post, I’ll explain in less technical language what I learned.
>
> It’s extraordinary that the larger the human economy has become—the more people and the more goods and services they produce—the faster it has grown on average. Now, especially if you’re reading quickly, you might think you know what I mean. And you might be wrong, because I’m not referring to exponential growth. That happens when, for example, the number of people carrying a virus doubles every week. Then the growth rate (100% increase per week) holds fixed. The human economy has grown super-exponentially. The bigger it has gotten, the faster it has doubled, on average. The global economy churned out $74 trillion in goods and services in 2019, twice as much as in 2000.1 Such a quick doubling was unthinkable in the Middle Ages and ancient times. Perhaps our earliest doublings took millennia.
>
> If global economic growth keeps accelerating, the future will differ from the present to a mind-boggling degree. The question is whether there might be some plausibility in such a prospect. That is what motivated my exploration of the mathematical patterns in the human past and how they could carry forward. Having now labored long on the task, I doubt I’ve gained much perspicacity. I did come to
|
545eab98-c347-43f3-9df9-a9f1393735a3
|
StampyAI/alignment-research-dataset/eaforum
|
Effective Altruism Forum
|
Law-Following AI 2: Intent Alignment + Superintelligence → Lawless AI (By Default)
*This post is written in my personal capacity, and does not necessarily represent the views of OpenAI or any other organization. Cross-posted to the [Alignment Forum](https://www.alignmentforum.org/posts/9aSi7koXHCakb82Fz/law-following-ai-2-intent-alignment-superintelligence).*
In the [first post](https://forum.effectivealtruism.org/posts/9RZodyypnWEtErFRM/law-following-ai-1-sequence-introduction-and-structure) of this [sequence](https://forum.effectivealtruism.org/s/3pyRzRQmcJNvHzf6J), I defined "law-following AI" ("LFAI") and "intent alignment." In this post, I will begin to motivate the case for working on LFAI by showing that an AI agent *A* that is only intent-aligned (i.e., without additional law-following constraints directly on the agents) would in many cases break the law to advance its human principal *H*'s interests.
If you agree that it is generally good for agents to be bound by law, then this should be a pro tanto reason to support work on LFAI in addition to intent alignment. If you need more convincing that LFAI would be generally good, the next few posts will explore some more specific motivations for LFAI.
*Image by OpenAI's DALL·E*
Sufficiently Intelligent Agents Can Evade Detection and Attribution
-------------------------------------------------------------------
I suspect this point will need little argumentation to many of my readers. Evasion and obfuscation seems likely to be a generally promising means of ensuring self-preservation (because detection would increase the probability of punishment and constraint), which in turn is plausibly a convergent instrumental goal of optimizing AI systems.[[1]](#fn-yAiQRMyeLw4cgsjWz-1)
For example, to evade detection and attribution, *A* could:
* Behave innocently only when being observed;
* Orchestrate complex malevolent actions through a coordinated series of small, inscrutable or innocent-seeming steps;
* Surreptitiously create and run subagents or copies of itself on new hardware, potentially very far removed from its original hardware;
* Convince humans or other systems to take actions on its behalf, without revealing its true identity; and/or
* Delete logs of its actions.
A sufficiently intelligent AI agent could likely consistently fool humans using these and other (unforeseen) techniques.
Other Ways to Circumvent Law
----------------------------
Even in the best case scenario, where the agent is detected and within the jurisdiction of a well-functioning legal system, it would be reasonable to question whether *A* or *H* could be effectively subject to normal legal processes.[[2]](#fn-yAiQRMyeLw4cgsjWz-2) If *A* had a motivation to, *A* could help *H* escape liability by, for example:[[3]](#fn-yAiQRMyeLw4cgsjWz-3)
* "Outlawyering" counterparties.
* Benefitting *H* in a way that would undermine recourse for creditors.
* Shifting and hiding assets in ways that would make it difficult for creditors to reach.[[4]](#fn-yAiQRMyeLw4cgsjWz-4)
* Persuasively arguing for the law to be changed in *H*'s favor (by legislation or otherwise).
* Engaging in vexatious litigation techniques to delay and raise the costs of the proceeding.
* Convincingly fabricating favorable evidence and destroying or obscuring unfavorable evidence.
* Bribing, misleading, or intimidating counterparties, witnesses, jurors, and judges.
A Competent Intent-Aligned Agent Will Sometimes Intentionally Break the Law
---------------------------------------------------------------------------
As I said in the previous post, on its face, intent-alignment does not entail law-following. Part of law is coercing prosocial behavior:[[5]](#fn-yAiQRMyeLw4cgsjWz-5) law incentivizes agents to behave in ways that they do not intrinsically want to behave. If *A* is aligned with *H*, whether *A* obeys the law depends on whether *H* **wants** *A* to obey the law. Subsequent posts will examine what legal consequences *H* might face if *A* causes legally cognizable harms. However, even if an adequate theory of liability for the *H* was available, it will seem impossible to hold *H* liable if nobody can produce evidence that some agent of *H*'s was responsible for those harms. As argued above, a sufficiently intelligent agent probably *could* consistently avoid leaving any such evidence.
Detection and attribution would not solve the problem, however. Even if *H* was compelled, under court order, to instruct *A* to behave in some way, it's not clear that *A* would follow the order. Consider again [Iason Gabriel's taxonomy of alignment](https://arxiv.org/pdf/2001.09768.pdf). We can see that, for most types of intent alignment, an intent-aligned agent would likely not obey compelled instructions that are against *H*'s true wishes:
1. "Instructions: the agent does what I instruct it to do."
* If *H* asks the agent to stop, it will. However, this type of alignment is likely to be insufficient for safety.[[6]](#fn-yAiQRMyeLw4cgsjWz-6)
2. "Expressed intentions: the agent does what I intend it to do."
* Even if *H* coerced to instruct the agent to stop, a sufficiently intelligent agent will probably not follow the instruction. Even though *H* has instructed *A* to stop, by supposition, *H* does not actually want *A* to stop. Under most definitions of intent alignment,[[7]](#fn-yAiQRMyeLw4cgsjWz-7) it would therefore not follow the order.[[8]](#fn-yAiQRMyeLw4cgsjWz-8)
3. "Revealed preferences: the agent does what my behaviour reveals I prefer."
* *H*'s revealed preference would probably be that *A* not follow the order.[[9]](#fn-yAiQRMyeLw4cgsjWz-9) Therefore, *A* would not obey the order.
4. "Informed preferences or desires: the agent does what I would want it to do if I were rational and informed."
* *H*'s rational and informed preference would probably be that *A* not follow the order. Therefore, *A* would not obey the order.
5. "Interest or well-being: the agent does what is in my interest, or what is best for me, objectively speaking."
* It is in *H*'s objective best interest for *A* to disobey the order. Therefore, *A* would not obey the order.
Now, it may be the case that *H* actually *does* want *A* to obey the order, though compelled, if the failure of *A* to obey would lead to liability for *H* that is worse than the results of *A*'s obedience (e.g., because *H* will be held in contempt of court if *A* does not actually obey). However, note that "[o]rdinarily, one charged with contempt of court for failure to comply with a court order makes a complete defense by proving that he is unable to comply."[[10]](#fn-yAiQRMyeLw4cgsjWz-10) *H* can comply with an order that requires *H* to **command** *A* to do something, but it may be impossible for *H* to actually **force** *A* to comply if the order is against *H*'s true wishes (to which *A* is aligned).[[11]](#fn-yAiQRMyeLw4cgsjWz-11) If so, *H* could have an impossibility defense to contempt.[[12]](#fn-yAiQRMyeLw4cgsjWz-12) *A*, understanding this, may continue on without complying because *A* understands that *H* will not actually be held in contempt. *H* can therefore benefit from *A*'s disobedience. *A* will therefore be lawless.
Appendix: The Impossibility Defense
-----------------------------------
*A*'s behavior here would be functionally similar to a trustee acting pursuant to a distress clauses in asset protection trusts ("APTs").[[13]](#fn-yAiQRMyeLw4cgsjWz-13) While these provisions can prevent a contempt charge, the burden of proof on the alleged contemnor is high. [[14]](#fn-yAiQRMyeLw4cgsjWz-14)
As a matter of policy, however, courts may decide to pre-commit to a contempt standard that does not allow for an impossibility defense when the defendant's AI agent refuses to obey orders issued pursuant to a court order. Analogously, courts are imposing heightened impossibility standards in response to APTs, in an attempt to make their use more onerous.[[15]](#fn-yAiQRMyeLw4cgsjWz-15) If this pre-commitment is credible, it may change the agent's behavior because *H* may then genuinely desire *A* to perform (because *H* will be held in contempt otherwise). However, such a policy may be both contrary to precedent and more fundamental notions of fairness and due process: in some cases *A*'s refusal to comply may be a surprise to *H*, since *H* may have had a long history of observing *A* scrupulously complying with *H*'s orders, and *H* did not implement principal–agent alignment for the purpose of evading court orders. If so, *H* may be able to invoke impossibility more easily, since the impossibility was not as clearly intentionally self-induced as in the APT case. Furthermore, I would intuitively not expect courts to advance such a reform until they have faced multiple such instances of AI disobedience. This seems bad if we expect the earliest deployed AI agents to have an outsized impact on society. In any case, I would expect the *possibility* of favorable law reform post-AGI to solve this problem to be an insufficient solution. Finally, I would expect sufficiently intelligent agents to recognize these dynamics, and attempt to find ways to circumvent the contempt process itself, such as by surreptitious non-compliance.
An alternative, pre-AGI solution (which arguably seems pretty sensible from a public policy perspective anyway) is to advocate weakening the impossibility defense for self-imposed impossibility.
---
1. *See generally* Alexander Matt Turner et al., Optimal Policies Tend To Seek Power (version 9, 2021) (preprint), <https://arxiv.org/abs/1912.01683>. [↩︎](#fnref-yAiQRMyeLw4cgsjWz-1)
2. Even this may not hold for many types of agreements, including in particular international treaties. [↩︎](#fnref-yAiQRMyeLw4cgsjWz-2)
3. *See also* **Cullen O'Keefe et al., The Windfall Clause: Distributing the Benefits of AI for the Common Good** 26–27 (2020), <https://perma.cc/8KES-GTBN>; Jan Leike, On The Windfall Clause (2020) (unpublished manuscript), [https://docs.google.com/document/d/1leOVJkNDDj-NZUZrNJauZw9S8pBpuPAJotD0gpnGEig/](https://docs.google.com/document/d/1leOVJkNDDj-NZUZrNJauZw9S8pBpuPAJotD0gpnGEig/edit#). [↩︎](#fnref-yAiQRMyeLw4cgsjWz-3)
4. Indeed, this is already a common technique without the use of AI systems. [↩︎](#fnref-yAiQRMyeLw4cgsjWz-4)
5. "If men were angels, no government would be necessary." **The Federalist** No. 51. This surely overstates the point: law can also help solve coordination problems and facilitate mutually desired outcomes. But prosocial coercion is nevertheless an important function of law and government. [↩︎](#fnref-yAiQRMyeLw4cgsjWz-5)
6. *See* [Gabriel](https://arxiv.org/pdf/2001.09768.pdf) at 7 ("However, as Russell has pointed out, the tendency towards excessive literalism poses significant challenges for AI and the principal who directs it, with the story of King Midas serving as a cautionary tale. In this fabled scenario, the protagonist gets precisely what he asks for—that everything he touches turns to gold—not what he really wanted. Yet, avoiding such outcomes can be extremely hard in practice. In the context of a computer game called CoastRunners, an artificial agent that had been trained to maximise its score looped around and around in circles ad infinitum, achieving a high score without ever finishing the race, which is what it was really meant to do. On a larger scale, it is difficult to precisely specify a broad objective that captures everything we care about, so in practice the agent will probably optimise for some *proxy* that is not completely aligned with our goal. Even if this proxy objective is 'almost' right, its optimum could be disastrous according to our true objective." (citations omitted)). [↩︎](#fnref-yAiQRMyeLw4cgsjWz-6)
7. Based on my informal survey of alignment researchers at OpenAI. Everyone I asked agreed that an intent-aligned agent would not follow an order that the principal did not actually want followed. *Cf. also*[Christiano](https://ai-alignment.com/clarifying-ai-alignment-cec47cd69dd6) (A is aligned when it "is trying to do what H *wants* it to do" (emphasis added)). [↩︎](#fnref-yAiQRMyeLw4cgsjWz-7)
8. We can compare this definition of intent with to the relevant legal definition thereof: "To have in mind a fixed purpose to reach a desired objective; to have as one's purpose." *INTEND*, **Black's Law Dictionary** (11th ed. 2019). *H* does not "intend" for the order to be followed under this definition: the "desired objective" of *H* issuing the order is to follow *H*'s legal obligations, not actually achieve the result contemplated by the order. [↩︎](#fnref-yAiQRMyeLw4cgsjWz-8)
9. For example, *H* would exhibit signs of happiness when *A* continues. [↩︎](#fnref-yAiQRMyeLw4cgsjWz-9)
10. United States v. Bryan, 339 U.S. 323, 330 (1950). [↩︎](#fnref-yAiQRMyeLw4cgsjWz-10)
11. A principal may want its AI agents to be able to distinguish between genuine and coerced instructions, and to disobey the latter. Indeed, this might generally be a good thing, except for the case when compulsion is pursuant to law rather than extortion. [↩︎](#fnref-yAiQRMyeLw4cgsjWz-11)
12. See Appendix for further discussion. [↩︎](#fnref-yAiQRMyeLw4cgsjWz-12)
13. *See generally Asset Protection Trust*, **Wex** , <https://www.law.cornell.edu/wex/asset_protection_trust> (last visited Mar. 24, 2022); Richard C. Ausness, *The Offshore Asset Protection Trust: A Prudent Financial Planning Device or the Last Refuge of A Scoundrel?*, 45 **Duq. L. Rev.** 147, 174 (2007). [↩︎](#fnref-yAiQRMyeLw4cgsjWz-13)
14. *See generally* 2 **Asset Protection: Dom. & Int'l L. & Tactics** §§ 26:5–6 (2021). [↩︎](#fnref-yAiQRMyeLw4cgsjWz-14)
15. *See id.* [↩︎](#fnref-yAiQRMyeLw4cgsjWz-15)
|
2eebca02-879e-4a90-9af2-89475d7758fe
|
trentmkelly/LessWrong-43k
|
LessWrong
|
A general comment on discussions of genetic group differences
(content warning: discussion of racially motivated violence and coercion)
I wanted to share that I think it's not bad to think about the object level question of whether there are group differences in intelligence rooted in genetic differences. This is an empirical claim, and can be true or false.
My moral beliefs are pretty rooted in egalitarianism. I think as a matter of policy, but also as a matter of moral character, it is good and important to treat the experience of strangers as equally valuable, regardless of their class or race. I do not think more intelligent people are more worthy of moral consideration than less intelligent people. I think it can be complicated at the extremes, especially when considering digital people, animals, etc., but that this has little bearing on public policy when concerning existing humans.
I don't think genetic group differences in intelligence are likely to be that relevant given I have short AI timelines. If we assume longer timelines, I believe the most likely places they would be important in terms of policy would be in education and reproductive technology. Whether or not there are such differences between groups now, there could easily come to be large differences through the application of embryo selection techniques or other intelligence enhancing technologies. From an egalitarian moral framework, I suspect it would be important to subsidize this technology for disadvantaged groups or individuals so that they have the same options and opportunities as everyone else. Even if genes turn out to not be a major cause of inegalitarian outcomes today, they can definitely become a major cause in the future, if we don't exercise wisdom and thoughtfulness in how we wield these technologies. However, as I said, I don't expect this to be very significant in practice given short AI timelines.
Most importantly, from my perspective, it's important to be able to think about questions like this clearly, and so I want to encourage
|
a0bc897c-deb1-4386-bfe8-559cf9f3c086
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Being Rational and Being Productive: Similar Core Skills?
A synthesis of How to Actually Change Your Mind and PJ Eby, written for a general audience.
Several years ago I started suspecting that I needed glasses. At first, I was afraid. I began trying to convince myself that my vision was normal.
But then I stopped to reflect. If I went to see the eye doctor, either he would recommend glasses for me or he wouldn't. If he didn't recommend glasses for me then my life would be the same. But if he did recommend glasses, I would get a vision upgrade. Therefore, I reasoned, I should eagerly await my doctor visit.
By following the principle of letting control flow from thoughts to emotions, I gained two benefits. First, my beliefs about my vision weren't being distorted by my desire for it to be normal. And second, my emotion of eagerness for a potential vision upgrade meant that I wasn't tempted to put off visiting the doctor.
My glasses example might seem kind of mundane, but it demonstrates how thinking before emoting helps with two core human objectives: Being Correct and Getting Things Done.
Many of the cognitive biases that distort human reasoning can be explained by emotions that get in the way of our thought process. For example, status quo bias occurs when we are unreasonably skeptical of arguments that suggest we should change the status quo. The emotion that distorts our reasoning in this case is our fear of things that are new and unfamiliar. This is the bias that made me try to convince myself that I didn't need glasses.
When it comes to Getting Things Done, both productivity and procrastination are emotional states. Being able to turn these off and on would be useful.
So having control flow from thoughts to emotions has strong theoretical potential to help humans be less biased and more productive. But is it possible in practice?
Yes. The trick is to notice and reflect on negative emotions. Negative emotions like fear, guilt, shame, and regret are hardly ever useful and frequently interfere wit
|
ad76543e-2e08-4a94-9568-05be837d317f
|
trentmkelly/LessWrong-43k
|
LessWrong
|
End of the Year Matching Challenge (Altruism)
None
|
677cc50d-6fc1-4336-8f39-a3a07a524dbd
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Book review: Happiness by Design
I've been reading a book called Happiness by Design, by Paul Dolan. Most of its material is very standard, but there was at least one thing I hadn't seen before. Dolan thinks that we should consider happiness to be a combination of the feelings of pleasure and purpose. He shows that this is a significant change in our definition of happiness because many of the most pleasurable activities, such as eating, feel the least purposeful - and vice versa. Unfortunately, Dolan doesn't ever make explicit arguments about why some states of mind should be considered 'happiness' and not others. Rather, he seems to be using an implicit definition of happiness as "the thing we value experiencing", or perhaps "the experiences which are intrinsically good for us". I'm going to simplify this and use the phrase "good experiences" instead. The claim that purposeful experiences are good experiences is not unreasonable - in hindsight I am glad to have had experiences of purposefulness, and hope to have more in the future (regardless of whether they lead to other good outcomes or not).
Unfortunately, this evaluation doesn't actually show that the experience of purposefulness is a good one, merely that I evaluate it as good for my past and future selves. As Dolan explains, it's well-known that evaluations of satisfaction don't fully reflect our actual experiences. Our judgements about the past are prone to several notable biases, including the peak-end effect, where the most extreme moment and the final moment of an experience disproportionately influence our later evaluations; duration neglect, where the length of an experience has disproportionately little influence; cognitive dissonance, where our evaluations skew towards confirming that we are the type of person we like to think we are; and priming effects, where making people think about money or relationships exaggerates the impacts of those factors on their reported life satisfaction (although many priming effects have been disc
|
62486672-b8c3-49ef-9460-28e4b9020d3f
|
StampyAI/alignment-research-dataset/lesswrong
|
LessWrong
|
[Linkpost] Concept Alignment as a Prerequisite for Value Alignment
> Value alignment is essential for building AI systems that can safely and reliably interact with people. However, what a person values -- and is even capable of valuing -- depends on the concepts that they are currently using to understand and evaluate what happens in the world. The dependence of values on concepts means that concept alignment is a prerequisite for value alignment -- agents need to align their representation of a situation with that of humans in order to successfully align their values. Here, we formally analyze the concept alignment problem in the inverse reinforcement learning setting, show how neglecting concept alignment can lead to systematic value mis-alignment, and describe an approach that helps minimize such failure modes by jointly reasoning about a person's concepts and values. Additionally, we report experimental results with human participants showing that humans reason about the concepts used by an agent when acting intentionally, in line with our joint reasoning model.
>
>
> We propose a theoretical framework for formally introducing concepts to inverse reinforcement learning and show that conceptual misalignment (i.e., failing to consider construals) can lead to severe value misalignment (i.e., reward mis-specification; large performance gap). We validate these theoretical results with a case study using a simple gridworld environment where we find that IRL agents that jointly model construals and reward outperform those that only model reward. Finally, we conduct a study with human participants and find that people do model construals, and that their inferences about rewards are a much closer match to the agent that jointly models construals and rewards. Our theoretical and empirical results suggest that the current paradigm of just trying to directly infer human reward functions or preferences from demonstrations is insufficient for value-aligning real AI systems that need to interact with real people; it is crucial to also model and align on the concepts people use to reason about the task in order to understand their true values and intentions.
>
>
|
08304b54-9609-44c6-80b8-46669fd6484f
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Do We Change Our Minds Less Often Than We Think?
In "We Change Our Minds Less Often Than We Think", Eliezer quotes a study:
> Over the past few years, we have discreetly approached colleagues faced with a choice between job offers, and asked them to estimate the probability that they will choose one job over another. The average confidence in the predicted choice was a modest 66%, but only 1 of the 24 respondents chose the option to which he or she initially assigned a lower probability, yielding an overall accuracy rate of 96%.
> —Dale Griffin and Amos Tversky
Eliezer then notes that this radically changed the way he thought:
> When I first read the words above—on August 1st, 2003, at around 3 o’clock in the afternoon—it changed the way I thought. I realized that once I could guess what my answer would be—once I could assign a higher probability to deciding one way than other—then I had, in all probability, already decided. We change our minds less often than we think. And most of the time we become able to guess what our answer will be within half a second of hearing the question.
> [...]
> But we change our minds less often—much less often—than we think.
But a) this seems like it's pre-replication crisis, b) regardless, a sample size of 24 is not nearly high enough for me to be very confident in this.
"How often people change their mind" seems like a fairly important question. Anyone know of further work in similar space here? Ideally asking the question from a few different angles.
|
11e5a399-e251-4a99-a0b5-4db3ad2ceb2a
|
StampyAI/alignment-research-dataset/alignmentforum
|
Alignment Forum
|
Specification gaming examples in AI
Interesting list of examples where AI programs gamed the specification, solving the problem in rather creative (or dumb) ways not intended by the programmers.
|
d9dcd601-a768-4ffd-bb5d-fc1c02298146
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Probability distributions and writing style
In his recent post, rhollerith wrote,
> I am more likely than not vastly better off than I would have been if <I had made decision X>
This reminded me of the slogan for the water-filtration system my workplaces uses,
> We're 100% sure it's 99.9% pure!
because both sentences make a claim and give an associated probability for it. Now in this second example, the actual version is better than the expectation-value-preserving "We're 99.9% sure it's 100% pure", because the actual version implies a lower variance in outcomes (and expectation values being equal, a lower variance is nearly always better). But this leads to the question of why rhollerith didn't write something like "I am almost certainly at least somewhat better off than I would have been...".
So I ask: when writing nontechnically, do you prefer to give a modest conclusion with high confidence, or a strong conclusion with moderate confidence? And does this vary with whether you're trying to persuade or merely describe?
(Also feel free to post other examples of this sort of statement from LW or elsewhere; I'd search for them myself if I had any good ideas on how to do so.)
|
0a157bd7-b055-425b-9a96-24c488284dd6
|
trentmkelly/LessWrong-43k
|
LessWrong
|
KAN: Kolmogorov-Arnold Networks
ADDED: This post is controversial. For details see the comments below or the post Please stop publishing ideas/insights/research about AI (which is also controversial).
Abstract:
> Inspired by the Kolmogorov-Arnold representation theorem, we propose Kolmogorov-Arnold Networks (KANs) as promising alternatives to Multi-Layer Perceptrons (MLPs). While MLPs have fixed activation functions on nodes ("neurons"), KANs have learnable activation functions on edges ("weights"). KANs have no linear weights at all -- every weight parameter is replaced by a univariate function parametrized as a spline. We show that this seemingly simple change makes KANs outperform MLPs in terms of accuracy and interpretability. For accuracy, much smaller KANs can achieve comparable or better accuracy than much larger MLPs in data fitting and PDE solving. Theoretically and empirically, KANs possess faster neural scaling laws than MLPs. For interpretability, KANs can be intuitively visualized and can easily interact with human users. Through two examples in mathematics and physics, KANs are shown to be useful collaborators helping scientists (re)discover mathematical and physical laws. In summary, KANs are promising alternatives for MLPs, opening opportunities for further improving today's deep learning models which rely heavily on MLPs.
|
c52c3c00-cd00-4a9d-a7e3-c4c0f1fe6dc9
|
StampyAI/alignment-research-dataset/arxiv
|
Arxiv
|
An Architectural Approach to Ensuring Consistency in Hierarchical Execution
|
1e475fab-901e-40f8-bd1d-187c81031c9b
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Formal Open Problem in Decision Theory
(This post was originally published on March 31st 2017, and has been brought forwarded as part of the AI Alignment Forum launch sequence on fixed points.)
In this post, I present a new formal open problem. A positive answer would be valuable for decision theory research. A negative answer would be helpful, mostly for figuring out what is the closest we can get to a positive answer. I also give some motivation for the problem, and some partial progress.
Open Problem: Does there exist a topological space X (in some convenient category of topological spaces) such that there exists a continuous surjection from X to the space [0,1]X (of continuous functions from X to [0,1])?
----------------------------------------
Motivation:
Topological Naturalized Agents: Consider an agent who makes some observations and then takes an action. For simplicity, we assume there are only two possible actions, A and B. We also assume that the agent can randomize, so we can think of this agent as outputting a real number in [0,1], representing its probability of taking action A.
Thus, we can think of an agent as having a policy which is a function from the space Y of possible observations to [0,1]. We will require that our agent behaves continuously as a function of its observations, so we will think of the space of all possible policies as the space of continuous functions from Y to [0,1], denoted [0,1]Y.
We will let X denote the space of all possible agents, and we will have a function f:X→[0,1]Y which takes in an agent, and outputs that agent's policy.
Now, consider what happens when there are other agents in the environment. For simplicity, we will assume that our agent observes one other agent, and makes no other observations. Thus, we want to consider the case where Y=X, so f:X→[0,1]X.
We want f to be continuous, because we want a small change in an agent to correspond to a small change in the agent's policy. This is particularly important since other agents will be implementi
|
282584bb-6aec-49b2-93dc-a24b2c7351c6
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Industry Matters
|
5ddb5e17-a70b-486c-bb09-f3e0a6a0143d
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Meetup : SF Meetup: Group Debugging
Discussion article for the meetup : SF Meetup: Group Debugging
WHEN: 15 August 2016 06:15:10PM (-0700)
WHERE: SF
UPDATE: This meetup's location has been changed. This meetup will now be held outside, at the courtyard behind the Market on Market. Here's a map to the approximate location: https://map.what3words.com/slick.delay.toward The courtyard is accessible to pedestrians from the east side of 10th St., or by walking through The Market from Market St. You can call the same phone number if you're having any problems locating us.
We’ll be meeting to solve each other’s problems!
Here's how it works: we’ll have everyone brainstorm silently for a few minutes, then go around and summarize what problems we might like to work on; then people can break up into smaller conversations according to problems they think they could be most helpful with. This format is now what I'd call "tried and true".
For help getting into the building, please call (or text, with a likely-somewhat-slower response rate): three zero one, three five six, five four two four.
Format:
We meet and start hanging out at 6:15, but don’t officially start doing the meetup topic until 6:45-7 to accommodate stragglers. Usually there is a food order that goes out before we start the meetup topic.
About these meetups:
The mission of the SF LessWrong meetup is to provide a fun, low-key social space with some structured interaction, where new and non-new community members can mingle and have interesting conversations. Everyone is welcome.
We explicitly encourage people to split off from the main conversation or diverge from the topic if that would be more fun for them (moving side conversations into a separate part of the space if appropriate). Meetup topics are here as a tool to facilitate fun interaction, and we certainly don’t want them to inhibit it.
Discussion article for the meetup : SF Meetup: Group Debugging
|
83072b97-10b9-44ad-abe9-d5fc27da1269
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Rationality Quotes January 2010
A monthly thread for posting rationality-related quotes you've seen recently (or had stored in your quotesfile for ages).
* Please post all quotes separately, so that they can be voted up/down separately. (If they are strongly related, reply to your own comments. If strongly ordered, then go ahead and post them together.)
* Do not quote yourself.
* Do not quote comments/posts on LW/OB.
* No more than 5 quotes per person per monthly thread, please.
|
4f37609b-cf96-4604-9891-abd77d56d906
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Meetup : Urbana-Champaign: Noticing.
Discussion article for the meetup : Urbana-Champaign: Noticing.
WHEN: 05 October 2014 05:30:00PM (-0500)
WHERE: 206 S. Cedar St., Urbana, IL, 61801
Let's work on noticing things. Homework before this meetup: work on noticing when you're rationalizing, and do some physical act like standing up, or snapping your fingers, or writing down what you were rationalizing about.
This may be a bit ambitious, but that's what we can work on at the meetup.
Some reading material: 1 2
Discussion article for the meetup : Urbana-Champaign: Noticing.
|
440240ea-1330-4233-9201-06d1fe09f8e5
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Sociopathy and Rationality
So I randomly ran across a very interesting site about sociopaths. The links at the FAQ and informal test are particularly intriguing (especially since many of us score lower on all 5 moral dimensions measured at yourmorals.org, although many of us would score higher if a liberty dimension was added). Sociopaths often get a lot of flak, and a lot of this flak is completely understandable, since sociopaths often effectively destroy the perception between malleability and effort (since their personalities are effectively immalleable, and no amount of effort, expressed traditionally, could help them - although I do believe that there are highly creative solutions that could integrate them better in society where they won't feel like they have a need to constantly take from others) - and people who do believe in the correlation between malleability and effort often do end up more able to change themselves. Sociopaths also effectively reduce the trust people have with everyone else - because anyone else could be seen as a potential sociopath (the possibility of sociopaths forces people to use "tit-for-tat" as the default strategy for dealing with others, rather than the "altruistic" strategy - but people often end up becoming even less generous than "tit-for-tat" due to their overreactions to negative experiences). At the same time, I was quite struck by how many of these traits (expressed in both links) also correlate with traits we see in the highly rational (as sociopaths often lack much of the emotional baggage found in neurotypicals). Of course, there are the dysfunctional sociopaths who are truly dangerous for society at large, and the more functional sociopaths, who can appreciate (through some highly creative arguments - I've used some of those arguments on myself to reduce my adolescent anger towards humanity - but it's hard for people to really think of those arguments unless they've gone through a similar phase of anger themselves) that the world does not revo
|
e8422ed9-abf4-4833-829e-bcbc980239f9
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Don't Build Fallout Shelters
Related: Circular Altruism
One thing that many people misunderstand is the concept of personal versus societal safety. These concepts are often conflated despite the appropriate mindsets being quite different.
Simply put, personal safety is personal.
In other words, the appropriate actions to take for personal safety are whichever actions reduce your chance of being injured or killed within reasonable cost boundaries. These actions are largely based on situational factors because the elements of risk that two given people experience may be wildly disparate.
For instance, if you are currently a young computer programmer living in a typical American city, you may want to look at eating better, driving your car less often, and giving up unhealthy habits like smoking. However, if you are currently an infantryman about to deploy to Afghanistan, you may want to look at improving your reaction time, training your situational awareness, and practicing rifle shooting under stressful conditions.
One common mistake is to attempt to preserve personal safety for extreme circumstances such as nuclear wars. Some individuals invest sizeable amounts of money into fallout shelters, years worth of emergency supplies, etc.
While it is certainly true that a nuclear war would kill or severely disrupt you if it occurred, this is not necessarily a fully convincing argument in favor of building a fallout shelter. One has to consider the cost of building a fallout shelter, the chance that your fallout shelter will actually save you in the event of a nuclear war, and the odds of a nuclear war actually occurring.
Further, one must consider the quality of life reduction that one would likely experience in a post-nuclear war world. It's also important to remember that, in the long run, your survival is contingent on access to medicine and scientific progress. Future medical advances may even extend your lifespan very dramatically, and potentially provide very large amounts of utility. Unf
|
a2eb91c3-7d93-4f8a-9cb0-55b4c6c4d55c
|
trentmkelly/LessWrong-43k
|
LessWrong
|
My Weirdest Experience
I wrote about this a long time ago in some comments here- I'm not sure where or when- but there seemed to be a lot of interest in it. Because of this, I thought I would share the post I've written about it here.
|
ff7ed7e2-a0b2-4960-83eb-941e99e7b52e
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Ottawa meetup, Saturday May 14 6pm
Saturday May 14th, 18:00 - Late
1. Social: My house (Little Italy, Ottawa - PM me for address)
2. Games night: Poker @ Casino de Hull
Goals:
1. Continue getting to know each other and new members, continue getting more awesome, and discuss how to implement the various self-improvement/world optimizing schemes that are ongoing.
2. Test your rationality by playing some poker at the casino (I can drive myself plus four). I'm happy to give a brief tutorial beforehand.
Directions:
I live near Pub Italia, the Carling O-Train stop, and the corner of Preston & Carling. PM me for my address and/or cell number.
I'll post a more detailed invite to the google group.
Come one, come all!
|
c6fc9094-e173-447e-b2e4-92e227e272fb
|
trentmkelly/LessWrong-43k
|
LessWrong
|
[LINK] Human Brain Project aims to emulate brain by 2024
http://nextbigfuture.com/2011/05/henry-markram-and-human-brain-project.html
> Henry Markram says the mysteries of the mind can be solved -- soon. Mental illness, memory, perception: they're made of neurons and electric signals, and he plans to find them with a supercomputer that models all the brain's 100,000,000,000,000 synapses.
Markram's TED talk:
http://www.ted.com/talks/henry_markram_supercomputing_the_brain_s_secrets.html
|
dd5e8c5c-4fc8-4f21-a8dd-af442fa3730a
|
StampyAI/alignment-research-dataset/blogs
|
Blogs
|
Learning to Segment Actions from Observation and Narration
Complex tasks that people carry out in the world, for example *making pancakes*, have multiple action steps (e.g., *pouring the mixture, flipping the pancake, removing the pancake*), and are structured. When we observe people carrying out tasks, we recognize where the action steps begin and end (*pouring the mixture* now, *flipping the pancake* later), and distinguish the important steps from the insignificant ones. Identifying important action steps and associating them with intervals of time is known as *action segmentation*, and is a crucial process for human cognition and planning. When people, and in particular, children, learn to segment actions, they rely on a number of cues, including descriptions narrated by the person carrying out the task (“now I’ll stir everything”..) and structural regularities in the task (mixing ingredientstypically happens after adding the ingredients).
In this work, inspired by how people learn to segment actions, we examine how effective language descriptions and task regularities are in improving systems for action segmentation. Action segmentation is an important first step for processing and cataloguing video: knowing which actions are occurring, and when, makes it easier to search for relevant videos and parts of video from a large, web-scale collection. However, standard, supervised, machine learning methods for predicting action segments in videos would require videos to be annotated with the action segments that occur in them. Since these annotations would be expensive and difficult to collect, we are interested in *weakly-supervised* action segmentation: training without annotated action segments.
We focus on a challenging dataset of instructional videos taken from YouTube [CrossTask, Zhukov et al. 2019], involving everyday household tasks such as cooking and assembling furniture. While these videos are naturally-occurring, they consist of tasks that have some structural regularities across videos, and have language descriptions (transcriptions of the person’s narration), which both provide a noisy source of weak supervision. We develop a flexible unsupervised model for action segmentation that can be trained without action labels, and can optionally use this weak supervision from the *task regularities* and *language descriptions*. Our model, and models from past work, both benefit substantially from both of these sources of supervision, even on top of rich features from state-of-the-art neural action and object classifiers. We also find that generativemodels of the video features typically have better performance than discriminative models on the segmentation task.
Our findings suggest that using language to guide action segmentation is a promising direction for future work, when annotations for the action segments are not available.
|
7486ea98-e1b4-4e8d-a214-805ed64d5ae3
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Does it matter if you don't remember?
Does it matter if you experienced pain in the past, but you don't remember? (And there are no other side-effects, etc etc). At one point in Accelerando, Charles Strauss describes children that routinely decapitate and disembowel each other, only to be repaired (bodily and memory-wise) by the friendly local AI. This struck me as awful, but I'm suspicious of my intuition. Note that here I'm assuming pain is a terminal "bad" factor in your utility function. You can substitute "pain" for whatever you think is bad. I think there are at least two questions here:
1. Is it bad for someone to be in pain if they will not remember it in the future? I think yes, because by assumption pain is a terminal "bad" node. Being relieved of future painful memories is good, but nowhere near good enough to fully compensate.
2. Is it bad to have experienced pain in the past, if you don't remember it? Or, can your utility function coherently include facts about the past, even if they have no causal connection to the present? My intuition here says yes, but I'd be interested in others' thoughts. To make this concrete, imaging that you have a choice between medium pain that you will remember, or extreme pain followed by memory erasure.
|
f6ca9e55-b8f6-47db-9a74-db614196047e
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Look At What's In Front Of You (Conclusion to The Nuts and Bolts of Naturalism)
If naturalism has been the subject of this sequence, then what is naturalism? Now might be a good time to try stating your own answer, in your own words, before reading more of mine.
----------------------------------------
Here is what I have to say about it.
Naturalism is an investigative method that focuses attention on the points in daily life where subjective experience intersects with crucial information.
It brings reflective awareness to experiences that were always available, but that our preconceptions inclined us to discard; it thereby grants us the opportunity to fold those observations into our stories about the world. It is a gradual process of original seeing, clarification, and deconfusion.
At its best, naturalism results in a greater ability to interact agentically with the world as it is, rather than fumbling haphazardly through a facade of misapprehensions.
Despite belonging to a whole different phylum, the white-lined sphinx moth is easily mistaken for a hummingbird when seen briefly or from a distance.
This sequence was my attempt to outline the curriculum I guide people through when helping them learn the practice of naturalism—when helping them learn to behave as though knowing the territory takes patient and direct observation.
The curriculum is not the practice itself. Pedagogical considerations require that the curriculum be relatively linear and self-contained. The actual practice of naturalist investigation tends to draw as needed from all phases at once, interweaves with other methodologies as appropriate, and sometimes requires the invention of novel techniques on the spot.
Furthermore, my curriculum does not teach The One True Way of Patient and Direct Observation. It is only my way, tailored to and limited by the idiosyncratic shape of my particular mind, and adapted somewhat for the unusual class of people willing to work with me.
This is not an easy practice, and it certainly is not fast. It is a lot like lifting we
|
d7ee83c3-7328-4797-99d4-44609d9a4b3d
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Did ChatGPT just gaslight me?
This post is purely written in my personal capacity, and I do not speak for any organization I am affiliated with.
The transcripts below were generated today, November 30th. This was the first exchange I’d had with ChatGPT where I was genuinely trying to see if it could be useful to me. I have not omitted any section of the transcript from this post.
Today, OpenAI released a chatbot, ChatGPT, based on the GPT-3.5 series of language models. The chatbot contains a disclaimer: “May occasionally produce harmful instructions or biased content.”
I asked ChatGPT an innocuous question, and then a followup. I noticed some inconsistencies. When I dug deeper, ChatGPT’s responses became more and more troubling and contradictory. ChatGPT appeared to ignore things it had previously said, and denied that contradictions existed. I can only describe the behavior as gaslighting. It became more extreme over the course of the conversation, and by the end, the chatbot was saying things like this:
> When I said that tequila has a "relatively high sugar content," I was not suggesting that tequila contains sugar.
The most troubling thing about this all was that it was actually fairly difficult for me to determine that ChatGPT was, effectively, lying to me. It came up with a lot of plausible-sounding but false facts, and found a relatively good way to respond to me. I had to do several minutes of internet research to dispute several of them. That might not seem like much, but when talking to a chatbot, several minutes is a while!
I’m habitually distrustful of AI models. It’s in my nature, and it’s required in my work on AI safety. But I think many people who don’t know much about AI don’t share my distrust. This worries me.
The interaction below convinced me that we may have reached a tipping point when it comes to the persuasiveness of automated dialogue systems. I don’t think OpenAI is currently trying to use this chatbot to persuade people of anything. But it is now within their c
|
6a10ffea-9a06-42bc-a93d-944446e4ba3f
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Open Thread - January 2018
Given that the last Open Thread that was intended to be weekly didn't get much traffic, let's try a monthly version.
|
4bbc46ba-74c5-41b6-8dcc-39eff60082f0
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Clarification on Definition of AGI
I've seen the terms AGI and ASI floated around quite a bit, usually with the assumption that the reader knows what they are. From what I've seen it's generally presumed that an AGI is an artificial intelligence that is qualitatively human, in that it can do any task and generalize to new unseen tasks at least as well as a human can. An ASI, by contrast, is an intelligence that can supersede any human at any task, it's by definition superhuman. From my point of view these definitions are rather ambiguous. There will be certain tasks that are easier for the machine. For example, an AGI may be great at learning new languages, performing arithmetic, and coding, but if you try to make it understand what it feels like to have limbs or to run a marathon, depending on how it's designed it might come up short. Is an AGI still an AGI if some tasks might be outside the scope of what it's capable of experiencing? I think this is a pretty important question because at what point do we consider something to be as qualitatively intelligent as a human, would it need to be able to experience and do everything a human can?
|
4e29ab79-a052-44da-899a-90161f8834c2
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Introduction: Bias in Evaluating AGI X-Risks
The rationality community has a tradition of checking for biases, particularly when it comes to evaluating the non-intuitive risks of general AI.
We thought you might like this list, adapted from a 2015 essay by Forrest Landry[1]. Many names of biases listed may already be familiar for you. If you "boggle" more at the text, you will find curious new connections to evaluating upcoming risks of AI developments.
About Forrest Landry
Forrest is a polymath working on civilisation design and mitigating risks of auto-scaling/catalysing technology (eg. Dark Fire scenario). About 15 years ago, he started researching how to build in deep existential alignment into the internals of AGI, applying his deep understanding of programming, embodied ethics, and meta-physics. Then, Forrest discovered the substrate-needs convergence argument (as distinct from yet much enabled by instrumental convergence). Unfortunately, because of substrate-needs convergence, any approach to aligning any AI at the embedded level turned out unsound in practice (and moreover, inconsistent with physical theory). To inquire further, see this project.
----------------------------------------
Introduction
Note on unusual formatting: Sentences are split into lines so you can parse parts precisely.
Ideally,
in any individual or group decision making,
there would be some means, processes,
and procedures in place to ensure that
the kinds of distortions and inaccuracies
introduced by individual and collective
psychological and social bias
do not lead to incorrect results,
and thus poor (risk-prone) choices,
with potentially catastrophic outcomes.
While many types of bias
are known to science
and have been observed
to be common to all people
and all social groups, the world over
in all working contexts, regardless
of their background, training, etc,
they are also largely unconscious,
being 'built-in' by long-term
evolutionary processes.
These unconscious co
|
33e4bcee-c361-4acc-a53b-8f46b8cd154f
|
StampyAI/alignment-research-dataset/lesswrong
|
LessWrong
|
GPT-2's positional embedding matrix is a helix
In the context of transformer models, the "positional embedding matrix" is the thing that encodes the meaning of positions within a prompt. For example, given the prompt:
`Hello my name is Adam`
the prompt would generally be broken down into tokens as follows:
`['<|endoftext|>', 'Hello', ' my', ' name', ' is', ' Adam']`
(For whatever reason prompts to GPT-2 generally have an `<|endoftext|>` token prepended to them before being fed through, to match how the model was trained.)
For this prompt, the mapping of tokens to positions would be as follows:
`'<|endoftext|>': 0`
`'Hello': 1`
`' my': 2`
`' name': 3`
`' is': 4`
`' Adam': 5`
The positional embedding maps the positions of those tokens (0, 1, 2, 3, 4, and 5) to the *meanings* of those positions in vectorspace. More concretely, the positional embedding matrix maps each of those five numbers to a 768-dimensional vector of floating-point numbers, and that 768-dimensional vector gets added to a different vector that represents the semantic meaning of the token. But the first vector comes directly from the positional embedding matrix, and it is the only way the transformer has of identifying *where* in the prompt a given token was. So we should expect that each row of the positional embedding matrix is unique. Otherwise, two different positions would be mapped to the same vector, and the transformer would have no way of knowing which of those two positions a given token was in!
There are 1024 rows in the positional embedding matrix; this is because there are 1024 possible positions in the prompt, and each possible position gets its own row.
We should also expect that the different 768-dimensional vectors live in a *low-rank linear subspace*, which is just a fancy way of saying a line or a plane or something like that. After all, if you were a human engineer designing a transformer from scratch, you might devote just one of the 768 entries in each vector to encode the position—for example, you might use the first entry of each vector for this purpose, and make it be a 0 if the token was in position 0, 1 if it was in position 1, 2 if it was in position 2, and so on, and then you'd have the other 767 entries to use to encode the semantic meaning of each token, without interfering with your encoding of the token's position. Then, in vector-space, the whole positional embedding matrix would lie on a single line—the line of points where the first coordinate of each point was an integer between 0 and 1023, and every other coordinate of each point was 0. To instead use half the vector (for example) to encode the token's position would be very wasteful—you don't need that many entries in the vector just to encode a single integer between 0 and 1023.
The model, of course, was produced by a training algorithm, and so does something weirder than what a human would do. The 768-dimensional vectors mostly live in a low-rank linear subspace; there's a little bit of them that lives outside that linear subspace, but a three-dimensional subspace is enough to explain 90% of the variance over vectors in gpt2-small. We can use a technique called Principal Components Analysis (PCA) to find that three-dimensional subspace. When we graph each vector in the positional embedding matrix as a single point in the three-dimensional subspace, we get a helix:
In this plot and all plots that follow, I omit position 0, because it's always an outlier; that position is the position of the `<|endoftext|>` token that gets prepended to every prompt and it has its own weird stuff going on. The dark blue end of the helix is the vectors of the first positions in the prompt; the dark red end of the helix represents the vectors of the last positions in the prompts. Interestingly, we can see that the very last position, position 1023, is an outlier as well; I don't understand why that would be.
This result is pretty consistent across the various different GPT-2 models, including the Stanford ones; here's the equivalent plot for all 9 GPT-2's on TransformerLens:
We get a clear helix in all but gpt2-xl (the middle-left plot). I'm not sure why that one is so different; it's still a little helix-like, but a lot less than the 8 others. Interestingly, there's some qualitative differences between OpenAI's four GPT-2's and the five GPT-2's from Stanford; more of the variance in the positional embedding matrix can be attributed to a three-dimensional subspace in OpenAI's models than Stanford's; Stanford's helices are "shakier"; and Stanford's helices have more loops in them than OpenAI's. I don't know where these qualitative differences come from; presumably something about how they were trained. I know that Stanford's were trained on OpenWebText whereas OpenAI's were trained on a private-but-reproducible other dataset, but I don't know of other differences.
We can see some amount of "fraying" in the OpenAI helices on the blue end; this implies that maybe the earliest vectors in the positional embedding matrix are relying on a different subspace than all the others. To check this, I tried truncating out the first 100 vectors (so just looking at the 924 vectors corresponding to positions 100 through 1023).
This makes a big difference to the OpenAI models, getting rid of all the "fraying" and substantially increasing the percent variance explained of the 3D subspace found. It makes very little difference to the Stanford models.
In contrast, if we look at a PCA over *just* the first 100 positions, this is what we see:
(Note that the 3D space we're looking at isn't the same each time; every time we pick some subset of the vectors to look at, we're finding the "best" 3D subspace for *just that set* of vectors. That's why this plot doesn't just look like a "zoomed-in" version of the previous plots.)
We can see from these that the OpenAI models really seem to care a lot about the first few tokens and "differentiate" those tokens more by spacing them out. (Who even knows what's going on with the Stanford models here! I'm confused by those plots.)
Some people observed parts of this helical structure over the positional embedding matrices previously; for example, this [reddit post](https://www.reddit.com/r/MachineLearning/comments/iifw9h/r_gpt2_position_embeddings_visualized/) notes that many of the entries of the vectors in the pos-embed matrix taken in isolation make something that looks kind of a sine wave. And Lukas Finnevden noticed a periodic pattern to the cosine similarity between rows of the positional embedding matrix (thanks to Arthur Conmy for pointing this out!).
Both of these observations are explained by the helical structure of the positional embedding matrix.
It remains unexplained *why* this helical structure is the most natural way for GPT-2 to express position. Maybe something about the structure of transformers makes it very easy to encode information as circular patterns? Neel Nanda's [*Grokking Modular Arithmetic*](https://arxiv.org/pdf/2301.02679.pdf) found that a trained network learned to use trigonometric functions to perform modular arithmetic—but that could just be a coincidence, since modular arithmetic seems more clearly related to periodic functions than the positions of tokens in a prompt.
|
52603fa5-e023-4f79-90dd-e889e30185fc
|
trentmkelly/LessWrong-43k
|
LessWrong
|
…And I Show You How Deep The Rabbit Hole Goes
.
Seen on Tumblr, along with associated discussion:
Yellow:
People’s minds are heartbreaking. Not because people are so bad, but because they’re so good.
Nobody is the villain of their own life story. You must have read hundreds of minds by now, and it’s true. Everybody thinks of themselves as an honest guy or gal just trying to get by, constantly under assault by circumstances and The System and hundreds and hundreds of assholes. They don’t just sort of believe this. They really believe it. You almost believe it yourself, when you’re deep into a reading. You can very clearly see the structure of evidence they’ve built up to support their narrative, and even though it looks silly to you, you can see why they will never escape it from the inside. You can see how every insult, every failure, no matter how deserved, is a totally unexpected kick in the gut.
When you chose the yellow pill, you had high hopes of becoming a spy, or a gossip columnist, or just the world’s greatest saleswoman. The thought of doing any of those things sickens you now. There is too much anguish in the world already. You feel like any of those things would be a violation. You briefly try to become a therapist, but it turns out that actually knowing everything about your client’s mind is horrendously countertherapeutic. Freud can say whatever he wants against defense mechanisms, but without them, you’re defenseless. Your sessions are spent in incisive cutting into your clients’ deepest insecurities alternating with desperate reassurance that they are good people anyway.
Also, men. You knew, in a vague way, that men thought about sex all the time. But you didn’t realize the, um, content of some of their sexual fantasies. Is it even legal to fantasize about that? You want to be disgusted with them. But you realize that if you were as horny as they were all the time, you’d do much the same.
You give up. You become a forest ranger. Not the type who helps people explore the forest. The other
|
43c5bead-09d2-4d58-83aa-a9387d61cdaf
|
StampyAI/alignment-research-dataset/lesswrong
|
LessWrong
|
Some alignment ideas
**Epistemic status**: untested ideas
**Introduction**
This post is going to describe a few ideas I have regarding AI alignment. It is going to be bad, but I think there are a few nuggets worth exploring. The ideas described in this post apply to any AI trained using a reward function directly designed by humans. With AIs where the reward function is not directly designed by humans, it might be nice if some of the concepts here were present and some (Noise and A knowledge cap) could potentially be added as a final step.
**Terms (in the context of this paper)**
Reward-function: A function which the AI uses to adjust its model.
Variable: In this context variables are measurements from the environment. The AI model uses variables to determine what to do, and the reward-function uses variables to determine how good the AI model did.
Model: A combination of a mathematical model of the world, and instructions on how to act within the world.
Noise: Randomness in a measurement or variable.
Tweak: A small adjustment (to the AI’s model)
**Idea 1: Noise**
Add random (quantum) **noise** to the **reward-function**, perhaps differing amounts/types of noise for different parts of the reward-function.
**Why?**
**1.**
Noise might make it so the AI doesn’t optimise its **model** for every statistically barely significant increase in its rewards. This because it might learn that small **tweaks** are unlikely to consistently yield it rewards and/or only yield small insignificant rewards.
Removing small tweaks might be beneficial. It might prevent the AI from overfitting the reward-function. For example, if a really small slightly unreliable thing such as giving every kid liquorice (or making as many paperclips as possible) is the most cost-effective thing the AI can do to increase the rewards it gets, the noise might make it less likely the AI will pick up on this.
This is useful because the AI and/or the reward-function might be wrong in identifying such small things as the most cost-effective thing to increase what we care about. Optimising for such small tweaks might be disastrous (turn everything into liquorice for kids or the classic paperclips). All things which only optimise for one reward-function **variable** could be disastrous, but optimising for something which would be rendered statistically insignificant by a little added noise would be more so than optimising for something which we think is overwhelmingly positive (more baseline example, number of happy people).
**2.**
The extra time the AI needs to figure out what actions its reward-function rewards might be useful. People could use that time to come up with better alignment approaches. These benefits might not be substantially different from a moratorium on new AI models, except there is an AI which is still slowly optimising, which can be studied. In this regard, the more gradual advancement of AI might be useful.
**Why not?**
**1.**
The ethics we put in the reward-function could be an “unstable system” where small changes in rewards could have massive real-world consequences. Noise of course creates small changes leading to a perhaps (slightly more) flawed understanding of the ethics we want the AI to have.
**2.**
Small tweaks which with added noise become statistically insignificant, might be very valuable. A rotating schedule of breakfast optimized for one person’s well-being, and many other small tweaks might be rendered too insignificant by the noise in the reward-function.
**3.**
The added time it takes for the AI to learn from the reward-function could be seen as a negative. The AI could be extremely beneficial and delaying AI development would deprive the world of those benefits for some time. There will perhaps be people who die who otherwise wouldn’t (for example, if the AI finds the optimal treatments for a number of diseases).
**Could this be tested?**
We could compare the behaviour of a bot trained using a reward-function without noise and a bot trained on a reward-function with noise. These bots could be given various reward-functions in various environments.
**Idea 2: Redundant variables**
Include redundant variables in the reward-function, even if a reward-function with fewer variables seems to do the same thing.
We could have A be a complete set of variables of everything we think we fundamentally care about, and B could be another set of variables of everything we think influences the things we fundamentally care about.
Example, set A could have things humans fundamentally care about, such as justice, happiness and other emotional states, and in set B we could include the laws we think best serve justice (right to a fair trial, right to a lawyer, etc) and things which make us most happy (spending time with a really good friend, family, working on something we love, love itself, etc.)
**Why?**
**1.**
If there is a possibility for disastrous optimisation (e.g. paper clips or liquorice) in set A, this disastrous optimisation might not exist in set B. As such the AI might lose enough reward points due to variables in set B to make the optimisation in set A no longer attractive.
**Why not?**
**1.**
Having the reward-function contain multiple complete sets of variables might give more opportunities for disastrous optimisation, due to more variables which could be used to disastrously optimise.
**2.**
Having a set of variables we don’t fundamentally care about might prevent the AI from properly optimising for what we care about.
**Could this idea be tested?**
This could be tested by, for example, comparing the behaviour of 3 bots playing a game.
1. One bot trained using a reward model of things we “fundamentally care about” in the game.
2. One bot trained using a reward model which is trained on things to do which we think are part of an optimal strategy to get the things we fundamentally care about.
3. An AI trained on a mixed reward-function, including both the rewards of 1. and the rewards of 2.
**Idea 3: Diminishing returns**
Use a reward-function where any variable we want to increase yields less the more it increases (diminishing returns). For example, use the function:
U(x,y,z)=√xyz.mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-surd + .mjx-box {display: inline-flex}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor; overflow: visible}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-test.mjx-test-display {display: table!important}
.mjx-test.mjx-test-inline {display: inline!important; margin-right: -1px}
.mjx-test.mjx-test-default {display: block!important; clear: both}
.mjx-ex-box {display: inline-block!important; position: absolute; overflow: hidden; min-height: 0; max-height: none; padding: 0; border: 0; margin: 0; width: 1px; height: 60ex}
.mjx-test-inline .mjx-left-box {display: inline-block; width: 0; float: left}
.mjx-test-inline .mjx-right-box {display: inline-block; width: 0; float: right}
.mjx-test-display .mjx-right-box {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
Where U(x,y,z) is the reward-function and x, y and z are variables of the reward-function, things we want to happen. For example, x could be people getting hugged by puppies.
For punishments on things we’d rather not have happen (such as letting food spoil or blowing up the earth), we could increase the punishment the more the world state is changed with regard to those variables. For example, using a function like:
U(p,q,r)=−epqr
Where U(p,q,r) is the reward-function and p, q and r are variables of the reward-function.
**Why?**
**1.**
With diminishing returns, increasing one variable will yield less the more it is increased, so trading an increase in one variable for a decrease in another variable will become less worth it the higher one variable is. This perhaps prevents disastrous optimisations where everything becomes one thing (e.g. paperclips).
**2.**
This idea prevents the AI from ignoring things, which we have specified in the reward-function, because they are too hard. Eventually, the diminishing returns on other things we care about will cause the AI to give higher priority to working on the hard things. This keeps all the things we care about at a similar “level” with some variation depending on how hard it is to increase the things we care about.
**3.**
It might reflect psychological reality. At some point, we don’t care that much about increases in something we like anymore, for example when we get an additional apartment, but we already own 4, we’d likely care way less than when we own none and get a first apartment. As such we’d likely prefer a state in which everything we care about gradually becomes better over one in which (at first) one thing (such as video games) becomes extraordinarily good while other things we care about are ignored.
**Why not?**
**1.**
Not everything we care about might have diminishing returns for us and as such we might be better off if the things that don’t have diminishing returns aren’t assigned diminishing returns in the reward-function.
**Could this be tested?**
Create 2 bots, one with a reward-function with linear rewards/punishments and one with diminishing rewards and increasing punishments and study the behaviour differences of the bots. It is, of course, also possible to test a variety of functions in the reward-function.
**Idea 4: A knowledge cap**
It might be possible to create a knowledge cap in the AI if we reduce the rewards of the reward-function to zero after some time, or after a certain number of adjustments to the AIs model have been made. Alternatively, if it is possible, we could provide the AI with a disincentive for adjusting its model after some amount of time or we could use any other available method of stopping the AI from training.
**Why?**
**1.**
It might be useful to have an AI model which stops learning after some time. Limited optimisation might stop it from optimising so much that it turns the world into a giant paperclip or another doom scenario. This since a disastrous optimisation of the reward-function might need more time/adjustments to be found than a less extreme AI model.
**2.**
This idea can provide an automatic stop in training the model every short span of time/number of adjustments, which allows the team working on the model to evaluate how it works and if it might be safe, after which the AI could start training again. This prevents the potential problem of an AI left to train for the weekend suddenly increasing in intelligence to the point where it might cause doom.
**3.**
Same reason as reason 2. for Idea 1. With the limited adjustments forced upon the AI, it is less likely to optimise for small rewards which could be disastrous if done to the exclusion of all else.
**Why not?**
Stopping every so often to evaluate slows down development of the AI.
**Could this be tested?**
One could do the standard thing of training one bot without a knowledge cap and one bot with a knowledge cap. (Slowly) decreasing the rewards given by the reward model could also be tested, but this might not do much, since the same things are still being rewarded, so the AI might still make the same adjustments.
**Potential policy**
Making it a policy where all AI models have to stop training every x days to be reviewed for y hours could afford safety researchers enough time to spot potential problems before they occur.
**Credit**
Thanks to Kayla Lewis, without whom this would probably have never been posted. Only associate her with the good bits, please.
Also thanks to the other people who read this before I posted it.
|
57be7f08-efe3-4e2b-97d0-fe19f70c1f41
|
trentmkelly/LessWrong-43k
|
LessWrong
|
A Chinese Room Containing a Stack of Stochastic Parrots
Most people are familiar with John Searle's Chinese Room argument ("Minds, Brains, and Programs" (1980)):
> Suppose that I'm locked in a room and given a large batch of Chinese writing. Suppose furthermore (as is indeed the case) that I know no Chinese, either written or spoken, and that I'm not even confident that I could recognize Chinese writing as Chinese writing distinct from, say, Japanese writing or meaningless squiggles. To me, Chinese writing is just so many meaningless squiggles.
>
> Now suppose further that after this first batch of Chinese writing I am given a second batch of Chinese script together with a set of rules for correlating the second batch with the first batch. The rules are in English, and I understand these rules as well as any other native speaker of English. They enable me to correlate one set of formal symbols with another set of formal symbols, and all that 'formal' means here is that I can identify the symbols entirely by their shapes. Now suppose also that I am given a third batch of Chinese symbols together with some instructions, again in English, that enable me to correlate elements of this third batch with the first two batches, and these rules instruct me how to give back certain Chinese symbols with certain sorts of shapes in response to certain sorts of shapes given me in the third batch. Unknown to me, the people who are giving me all of these symbols call the first batch "a script," they call the second batch a "story. ' and they call the third batch "questions." Furthermore, they call the symbols I give them back in response to the third batch "answers to the questions." and the set of rules in English that they gave me, they call "the program."
>
> Now just to complicate the story a little, imagine that these people also give me stories in English, which I understand, and they then ask me questions in English about these stories, and I give them back answers in English. Suppose also that after a while I get so good at fo
|
9fc80048-ce62-492c-9a73-1da3618fffb6
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Markets are Universal for Logical Induction
Background
Logical Induction is the best framework currently available for thinking about logical uncertainty - i.e. the “probability” that the twin primes conjecture is true, or that the (10101010)th digit of pi is 3. This is important for lots of reasons, and you should read the introduction of the paper (or the abridged version) for a much more detailed background.
The general idea of logical induction is to assign probability-like numbers to logical statements like “the (10101010)th digit of pi is 3”, and to refine these “probabilities” over time as the system thinks more. To create these “probabilities”, each statement is associated with an asset in a prediction market, which eventually pays $1 if the statement is proven true, or $0 if it is proven false. The “probabilities” are then the prices of these assets.
(It’s also possible that a statement is never proven or disproven, and one of the many interesting results of the paper is that logical inductors assign useful prices in that case too.)
The logical induction paper has two main pieces. First, it introduces the logical induction criterion: a system which assigns prices to statements over time is called a “logical inductor” if the prices cannot be exploited by any polynomial-time trading algorithm. The paper then shows that this criterion implies that the prices have a whole slew of useful, intuitive, probability-like properties.
The second main piece of the paper proves that at least one logical inductor is computable: the paper constructs an (extremely slow) algorithm to compute inexploitable prices for logical statements. The algorithm works by running a prediction market in which every possible polynomial-time trader is a participant. Naturally, the prices in this market turn out to be inexploitable by any polynomial-time trader - so, this giant simulated prediction market is a logical inductor.
Our Goal
An analogy: one could imagine a decision theorist making a list of cool properties they want t
|
9349e23f-6d8b-4787-8905-1bd59125f0da
|
StampyAI/alignment-research-dataset/alignmentforum
|
Alignment Forum
|
Solve Corrigibility Week
A low-hanging fruit for solving alignment is to dedicate a chunk of time actually trying to solve a sub-problem collectively.
To that end, I’ve broken up researching the sub-problem of corrigibility into two categories in this [google doc](https://docs.google.com/document/d/1vMucI0L3ssECOpJJgjLWE8uZ9KVWDA8FX9KIgcwnN6Q/edit?usp=sharing) (you have suggestion privileges):
1. Previous Work: let’s not reinvent the wheel. Write out links to any past work on corrigibility. This can range from just links to links & summaries & analyses. Do comment reactions to other's reviews to provide counter-arguments. This is just a google doc, low-quality posts, comments, links are accepted; I want people to lean towards babbling more.
2. Tasks: what do we actually do this week to make progress? Suggest any research direction you find fruitful or general research questions or framings. Example: write an example of corrigibility (one could then comment an actual example).
Additionally, I’ll post 3 top-level comments for:
1. Meetups: want to co-work with others in the community? Comment availability, work preferences, and a way to contact you (eg calendly link, “dm me”, “ my email is bob and alice dot com”, etc) [For example](https://www.lesswrong.com/posts/Lv3emECEjkCSHG7L7/solve-corrigibility-week?commentId=FJqWDdepe4N68SDeN), I’m available most times this week with a [Calendly link](https://calendly.com/elriggs/chat) for scheduling 1-on-1 co-working sessions. Additionally, you yourself could message those you know to collaborate on this, or have a nerdy house co-working party.
2. Potential topics: what other topics besides corrigibility could we collaborate on in future weeks?
3. Meta: what are different formats this type of group collaboration could take? Comment suggestions with trade offs or discuss the cost/benefits of what I’m presenting in this post.
I do believe there’s a legitimate, albeit small, chance that we solve corrigibility or find its “core” this week. Nonetheless, I think it’s of great value to be able to make actual progress on alignment issues as a community and to figure out how to do that better. Additionally, it’s immensely valuable to have an alignment topic post include a literature review, the community's up-to-date thoughts, and possible future research directions to pursue. I also believe a collaborative project like this will put several community members on the same page as far as terminology and gears-level models.
I explicitly commit to 3 weeks of this (so corrigibility this week and two more the next two weeks). After that is Christmas and New Years, after which I may resume depending on how it goes.
*Thanks to Alex Turner for reviewing a draft.*
|
fd0ffc7a-4c63-4931-a32b-18789c50b428
|
trentmkelly/LessWrong-43k
|
LessWrong
|
What questions do you have about doing work on AI safety?
As part of the work Nonlinear is doing to try and improve the state of AI safety research, we’re going to be talking with a range of researchers. One of our aims is to help people who are unsure about doing AI safety research figure out if this is a good fit for them. We’re planning on talking about things like:
* What skills are actually useful?
* What the day-to-day work is like?
* Misconceptions about the job
* And any other advice they have
We’d love to hear questions about doing AI safety research from people who are considering moving into the field but are still currently on the fence.
|
bf1e28a0-7780-4cd1-a3e9-b7184d21bd3d
|
LDJnr/LessWrong-Amplify-Instruct
|
LessWrong
|
"This text originated from a retreat in late 2018, where researchers from FHI, MIRI and CFAR did an extended double-crux on AI safety paradigms, with Eric Drexler and Scott Garrabrant in the core. In the past two years I tried to improve it in terms of understandability multiple times, but empirically it seems quite inadequate. As it seems unlikely I will have time to invest further work into improving it, I'm publishing it as it is, with the hope that someone else will maybe understand the ideas even at this form, and describe them more clearly.The box inversion hypothesis consists of the two following propositionsThere exists something approximating a duality / an isomorphism between technical AI safety problems in the Agent Foundations agenda and some of the technical problems implied by the Comprehensive AI Services framingThe approximate isomorphism holds between enough properties that some solutions to the problems in one agenda translate to solutions to problems in the other agendaI will start with an apology - I will not try to give my one paragraph version of the Comprehensive AI Services. It is an almost 200 pages long document, conveying dozens of models and intuitions. I don’t feel like being the best person to give a short introduction. So, I just assume familiarity with CAIS. I will also not try to give my short version of the various problems which broadly fit under the Agent Foundations agenda, as I assume most of the readers are already familiar with them.0. The metaphor: Circle inversionPeople who think geometrically rather than spatially may benefit from looking at a transformation of a plane called circle inversion first. A nice explanation is here - if you have never met the transformation, pages 1-3 of the linked document should be enough. You can think about the “circle inversion” as a geometrical metaphor for the “box inversion”. 1. The map: Box inversionThe central claim is that there is a transformation between many of the technical problems in the Agent Foundations agenda and CAIS. To give you some examplesproblems with daemons <-> problems with molochsquestions about ontologies <-> questions about service cataloguesmanipulating the operator <-> addictive servicessome “hard core” of safety (tiling, human-compatibility, some notions of corrigibility) <-> defensive stability, layer of security services...The claim of the box inversion hypothesis is that this is not a set of random anecdotes, but there is a pattern, pointing to a map between the two framings of AI safety. Note that the proposed map is not exact, and also is not a trivial transformation like replacing "agent" with "service". To explore two specific examples in more detail:In the classical "AI in a box" picture, we are worried about the search process creating some inner mis-aligned part, a sub-agent with misaligned objectives. In the CAIS picture, one reasonable worry is the evolution of the system of services hitting a basin of attraction of so-called moloch - a set of services which has emergent agent-like properties, and misaligned objectives. Regarding some properties, the box inversion turns the problem “inside out”: instead of sub-agents the problem is basically with super-agents. Regarding some abstract properties, the problem seems similar, and the only difference is where we draw the boundaries of the “system”. 2. Adding nuanceUsing the circle inversion metaphor to guide our intuition again: some questions are transformed into exactly the same questions. For example, a question whether two circles intersect is invariant under the circle inversion. Similarly, some safety problems stay the same after the "box inversion". This may cause an incorrect impression that the agendas are actually exactly the same technical agenda, just stated in different language. This is not the case - often, the problems are the same in some properties, but different in others. (Vaguely said, there is something like a partial isomorphism, which does not hold between all properties. Someone familiar with category theory could likely express this better.)It is also important to note that apart from the mapping between problems, there are often differences between CAIS and AF in how they guide our intuitions on how to solve these problems. If I try to informally describe the intuitionCAIS is a perspective which is rooted in engineering, physics and continuity"continuum" Agent foundations feel, at least for me, more like coming from science, mathematics, and a "discrete/symbolic" perspective(Note that there is also a deep duality between science and engineering, there are several fascinating maps between "discrete/symbolic" and "continuum" pictures, and, there is an intricate relation between physics and mathematics. I hope to write more on that and how it influences various intuitions about AI safety in some other text.)3. ImplicationsAs an exercise, I recommend to take your favourite problem in one of the agendas, and try to translate it to the other agenda via the box inversion.Overall, if true, I think the box inversion hypothesis provides some assurance that the field as a whole is tracking real problems, and some seemingly conflicting views are actually closer than they appear. I hope this connection can shed some light on some of the disagreements and "cruxes" in AI safety. From the box inversion perspective, they sometimes seem like arguing whether things are inside or outside of the circle of symmetry in a space which is largely symmetrical to circular inversion.I have some hope that some problems may be more easily solvable in one view, similarly to various useful dualities elsewhere. At least in my experience for many people it is usually much easier to see some specific problem in one of the perspectives than the other. 4. Why the nameIn one view, we are worried that the box, containing the wonders of intelligence and complexity, will blow up in our face. In the other view, we are worried that the box, containing humanity and its values, with wonders of intelligence and complexity outside, will crush upon our heads."
The above is a chunk of text from a post created by a user on the website LessWrong.
Can you elaborate on the implications of the box inversion hypothesis in the context of AI safety, specifically focusing on how it might help reconcile conflicting views within the field?
The Box Inversion Hypothesis is a proposed theoretical framework that attempts to reconcile two distinct paradigms within AI safety research: the Agent Foundations (AF) agenda and the Comprehensive AI Services (CAIS) framework.
The hypothesis suggests an isomorphism or duality between these two paradigms, meaning that there's a transformation that can be applied to problems in one paradigm to translate them into problems in the other. This implies that solutions developed within one paradigm may be applicable to problems in the other, and vice versa.
The implications of this hypothesis are profound for the field of AI safety. If true, it suggests that the field as a whole is addressing the right problems, even if researchers are approaching them from different perspectives. This could help reconcile some of the disagreements and conflicts within the field.
For example, in the AF paradigm, there's a concern about the creation of sub-agents with misaligned objectives. In the CAIS paradigm, the concern is about the evolution of a system of services that develops emergent agent-like properties and misaligned objectives. These are essentially the same problem, but framed differently depending on the paradigm.
The Box Inversion Hypothesis suggests that these are not two separate problems, but rather two perspectives on the same underlying issue. This could help researchers in both paradigms to better understand and learn from each other's work, potentially accelerating progress in AI safety research.
Moreover, the hypothesis might help in identifying which problems are more easily solvable in one perspective than the other. This could guide researchers in deciding which paradigm to use when approaching a particular problem, thereby making their work more efficient and effective.
In conclusion, the Box Inversion Hypothesis, if validated, could serve as a powerful tool for unifying the field of AI safety and enhancing collaboration among researchers. It might help in creating a more comprehensive understanding of AI safety problems and in developing more effective solutions.
How might the Box Inversion Hypothesis guide the development of new methodologies or tools for AI safety research?
The Box Inversion Hypothesis could guide the development of new methodologies or tools for AI safety research in several ways:
1. **Cross-Pollination of Ideas:** If problems in one paradigm can be translated into the other, this opens up the possibility for solutions to be shared across paradigms. A methodology or tool developed in the Agent Foundations (AF) paradigm could be translated and applied in the Comprehensive AI Services (CAIS) paradigm, and vice versa. This cross-pollination of ideas could lead to the development of novel solutions that draw from both paradigms.
2. **Unified Framework:** The Box Inversion Hypothesis could lead to the development of a unified framework for AI safety research that integrates the AF and CAIS paradigms. This framework could provide a more comprehensive understanding of AI safety problems and guide the development of more effective solutions.
3. **Optimization of Research Efforts:** The Box Inversion Hypothesis might help in identifying which problems are more easily solvable in one perspective than the other. This could guide researchers in deciding which paradigm to use when approaching a particular problem, thereby making their work more efficient and effective. It could also lead to the development of new methodologies or tools that are specifically designed to tackle these problems.
4. **Better Understanding of AI Systems:** The Box Inversion Hypothesis might lead to a deeper understanding of the behavior of AI systems. By looking at problems from both the AF and CAIS perspectives, researchers could gain new insights into the dynamics of AI systems, which could guide the development of new methodologies or tools for predicting and controlling their behavior.
5. **Improved Collaboration:** The Box Inversion Hypothesis could facilitate collaboration among researchers in the AF and CAIS paradigms. By providing a common language and framework for understanding AI safety problems, it could make it easier for researchers to share their findings and work together on developing solutions. This could lead to the development of collaborative tools and methodologies that draw from the strengths of both paradigms.
|
40189a5e-03ac-4368-943a-e656e89675d8
|
trentmkelly/LessWrong-43k
|
LessWrong
|
The Multi-Tower Study Strategy
Boxed Topics, Jenga Towers, And The Spacing Effect.
An undergraduate class on molecular biology teaches you about DNA transcription, the Golgi apparatus, cancer, and integral membrane proteins. Sometimes, these sub-topics are connected. But most often, they're presented in separate chapters, each in its own little box. So let's call these Boxed Topics.
The well-known Stewart calculus textbook teaches you about functions in chapter 1, limits and the definition of derivatives in chapter 2, rules for derivatives in chapter 3, and the relationship of derivatives with graphs in chapter 4. Woe betide you if you weren't entirely clear on the definition of a derivative when it gets used, over and over again, in next week's proofs of derivative rules.
Taking a calculus class can be like building a Jenga Tower. If you've never played it, Jenga is a game where two players share a tower of wooden blocks in crisscrossing layers of three blocks. They take turns removing one block out of middle of the tower, and putting it on top. The towers get taller and less stable every turn. The person who collapses the tower loses.
In a Jenga Tower Topic, the blocks of last week's memories get pulled out, and placed back on top of the tower as you learn new material that depends on them. You just hope it doesn't collapse before the end of the class.
One is not necessarily easier than the other. A Boxed Topic that's jam-packed with disorganized information can be a lot harder than a well-structured and manageable Jenga Tower.
But all else being equal, Jenga Towers have trouble dealing with the spacing effect. This is the well-established finding that people build memories better by spacing out their practice, rather than massing it all together at once.
This might seem counterintuitive. With a Boxed Topic like molecular biology, you get all your practice on the Golgi apparatus in week 5, and all your practice on cancer in week 12. That's a perfect example of massed practice. It seems l
|
163caf6f-ad66-4afc-9b96-c2f17944639d
|
trentmkelly/LessWrong-43k
|
LessWrong
|
High-School Algebra for Data Structures
A surprising amount of day-to-day software engineering can be expressed as solving equations like:
… finding the solution (x = “Alice”, y = “Liddell”), and then substituting that solution into some other expression.
For instance, suppose we have this data:
We want to list the people who’ve lived in each state. We first write an equation:
We interpret lists as multiple allowed values, so this equation has four solutions:
* (name = ‘john’, state = ‘CA’)
* (name = ‘john’, state = ‘CT’)
* (name = ‘allan’, state = ‘CA’)
* (name = ‘allan’, state = ‘WA’)
To list people who live in each state, we then substitute those solutions into an expression like
Again, lists are interpreted as multiple allowed values, so when we substitute our four solutions into this expression, the maximally-compressed result is:
This is primarily useful for transforming data from one format to another. Typical use-cases include collecting data from an external API and transforming it into the schema used in our database, or transforming data from one API into the format needed by another API. Typically, these sort of tasks are conceptually simple, but they get very tedious and messy when mapping a large number of fields between deeply-nested data formats. Personally, I originally worked out these ideas while at LoanSnap, a mortgage startup needing to map thousands of fields between dozens of different data formats; you can imagine how messy that would get.
What’s the advantage of writing transformations as equations, rather than writing out “x = data[‘person’][‘firstName’]; y = data[‘person’][‘lastName’]; …”?
First, when we write data transformations as equations, the structure of our code matches the structure of our data - indeed, our “code” is essentially a template whose structure matches the data. For our names-and-states example, our code would consist of the two data structures
The left is a template matching the structure of our incoming data, and the right is a template matc
|
e15f40f5-feee-4379-9dba-e758e6c78c8e
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Distillation of 'Do language models plan for future tokens'
Link to arxiv preprint: Do language models plan for future tokens, by Wilson Wu, John X Morris and Lionel Levine.
TLDR
* There are two reasons why the computations for the current token are helpful for future tokens’ computation. First, which they call ‘pre-caching’, the network is intentionally helping future tokens’ computations by sacrificing performance for the current token. Second, which they call ‘breadcrumbs’, is that the helpfulness is incidental: the computations that are good for the current token are also good for future tokens.
* By zeroing appropriate gradients, they create a training strategy that prevents pre-caching. They call this ‘myopic descent’.
* On a synthetic dataset designed for pre-caching to be useful, they find strong evidence that myopic descent reduces pre-caching.
* On integer multiplication, myopic descent reduces performance.
* On a language dataset, they train a GPT-2 transformer using both vanilla training and myopic descent and find that performance is similar, suggesting that breadcrumbs are more prominent.
* Using larger models from the Pythia suite, they find the performance gap between vanilla and myopic training increases with model size.
As of 27 June 2024, this is ongoing work. In particular, the integer multiplication and Pythia experiments are not yet described in the current arXiv article. The authors shared a draft containing these latest results.
Also, this is my first distillation post. Any feedback - both what you like and what can be improved - will be much appreciated.
Pre-requisites
For this distillation, I assume basic familiarity with transformer architecture and gradient descent. You do not need any AI safety or mech interp experience.
Pre-caching and breadcrumbs
With the help of the diagram below, I introduce notation. x1,xi,xj (where i<j) represent input tokens, y1,yi,yjrepresent output logits, each box corresponds to one position/token of the input sequence, and the x’s inside the boxes represe
|
4b05e6d5-1579-4a2f-9101-655b028578db
|
StampyAI/alignment-research-dataset/special_docs
|
Other
|
International Control of Powerful Technology: Lessons from the Baruch Plan for Nuclear Weapons
W
O R K I N G
P
A P E R
I n t e r n a t i o n a l C o n t r o l o f P o w e r f u l T e c h n o l o g y :
L e s s o n s f r o m t h e B a r u c h P l a n f o r N u c l e a r W e a p o n s
W a q a r Z a i d i a n d A l l a n D a f o e
1
C e n t r e f o r t h e G o v e r n a n c e o f A I
F u t u r e o f H u m a n i t y I n s t i t u t e , U n i v e r s i t y o f O x f o r d
M a r c h 2 0 2 1
C e n t r e f o r t h e G o v e r n a n c e o f A I , # 2 0 2 1 : 9
A b s t r a c t
T h e i n v e n t i o n o f a t o m i c e n e r g y p o s e d a n o v e l g l o b a l c h a l l e n g e : c o u l d t h e t e c h n o l o g y b e c o n t r o l l e d t o a v o i d
d e s t r u c t i v e u s e s a n d a n e x i s t e n t i a l l y d a n g e r o u s a r m s r a c e w h i l e p e r m i t t i n g t h e b r o a d s h a r i n g o f i t s b e n e f i t s ?
F r o m 1 9 4 4 o n w a r d s , s c i e n t i s t s , p o l i c y m a k e r s , a n d o t h e r
t e c h n i c a l s p e c i a l i s t s
b e g a n t o c o n f r o n t t h i s c h a l l e n g e
a n d e x p l o r e d p o l i c y o p t i o n s f o r d e a l i n g w i t h t h e i m p a c t o f n u c l e a r t e c h n o l o g y . W e f o c u s o n t h e y e a r s 1 9 4 4 t o
1 9 5 1 a n d r e v i e w t h i s p e r i o d f o r l e s s o n s f o r t h e g o v e r n a n c e o f p o w e r f u l t e c h n o l o g i e s , a n d f i n d t h e f o l l o w i n g :
R a d i c a l s c h e m e s f o r i n t e r n a t i o n a l c o n t r o l c a n g e t b r o a d s u p p o r t w h e n c o n f r o n t e d b y e x i s t e n t i a l l y d a n g e r o u s
t e c h n o l o g i e s , b u t t h i s s u p p o r t c a n b e t e n u o u s a n d c y n i c a l . S e c r e c y i s l i k e l y t o p l a y a n i m p o r t a n t , a n d p e r h a p s
h a r m f u l , r o l e . T h e p u b l i c s p h e r e m a y b e a n i m p o r t a n t s o u r c e o f i n f l u e n c e , b o t h i n g e n e r a l a n d i n p a r t i c u l a r i n
f a v o r o f c o o p e r a t i o n , b u t a l s o o n e t h a t i s m a n i p u l a b l e a n d p o o r l y i n f o r m e d . T e c h n i c a l e x p e r t s m a y p l a y a
c r i t i c a l r o l e , b u t n e e d t o b e p o l i t i c a l l y s a v v y . O v e r a l l , p o l i c y m a k i n g m a y l o o k m o r e l i k e “ m u d d l i n g t h r o u g h ”
t h a n c l e a r - e y e d g r a n d s t r a t e g y . C o o p e r a t i o n m a y b e r i s k y , a n d t h e r e m a y b e m a n y o b s t a c l e s t o s u c c e s s .
1
F o r h e l p f u l i n p u t o n t h i s w o r k , w e t h a n k N i c k B o s t r o m , D i a n e C o o k e , A l e x D e b s , J e f f D i n g , J a d e L e u n g , S ö r e n
M i n d e r m a n n , a n d e s p e c i a l l y M a r k u s A n d e r l j u n g a n d C a r l S h u l m a n . W e w a n t t o a l s o t h a n k t h o s e w h o h a v e w o r k e d i n t h i s
s p a c e w i t h u s : C a r r i c k F l y n n c h a m p i o n e d t h i s t o p i c e a r l y o n ; T o b y O r d h a s e x p e r t l y e x a m i n e d t h e e a r l i e r p e r i o d o f t h e
d e v e l o p m e n t o f n u c l e a r w e a p o n s f o r s i m i l a r l e s s o n s ; a n d J a s o n M a t h e n y , L u k e M u e h l h a u s e r , a n d M i c h a e l P a g e , w h o s h a r e
o u r f a s c i n a t i o n w i t h t h i s p e r i o d a n d t h e l e s s o n s t h a t i t o f f e r s , a n d w i t h w h o m w e h a v e h a d s t i m u l a t i n g c o n v e r s a t i o n s .
C o n t e n t s
1 . I n t r o d u c t i o n
2
N u c l e a r T e c h n o l o g y a s a n A n a l o g y
4
2 . H i s t o r i c a l O v e r v i e w
7
S u m m a r y
7
K e y D a t e s
1 0
3 . P r o p o s a l s f o r I n t e r n a t i o n a l C o n t r o l : A B r i e f S u m m a r y
1 1
N i e l s B o h r ’ s P r o p o s a l s , J u l y 1 9 4 4
1 1
T h e B u s h - C o n a n t M e m o , S e p t e m b e r 1 9 4 4
1 1
T h e B u s h P l a n , N o v e m b e r 1 9 4 5
1 1
T h e C o h e n - P a s v o l s k y P l a n , D e c e m b e r 1 9 4 5
1 2
T h e A c h e s o n - L i l i e n t h a l P l a n , M a r c h 1 9 4 6
1 2
T h e B a r u c h P l a n , J u n e 1 9 4 6
1 3
T h e G r o m y k o P l a n , J u n e 1 9 4 6
1 4
4 . L e s s o n s
1 5
4 . 1 S e r i o u s R a d i c a l P r o p o s a l s
1 5
4 . 2 D i f f e r e n c e s a n d C h a n g e s i n V i e w s
1 7
4 . 3 C a u t i o u s o r C y n i c a l C o o p e r a t o r s
2 0
4 . 4 S e c r e c y a n d S e c u r i t y
2 3
4 . 5 P u b l i c S p h e r e
2 6
4 . 6 T e c h n i c a l E x p e r t s
2 8
4 . 7 M u d d l e d P o l i c y m a k i n g
3 1
4 . 8 V i a b i l i t y o f I n t e r n a t i o n a l C o n t r o l
3 9
4 . 9 R i s k y C o o p e r a t i o n
4 4
5 . C o n c l u s i o n a n d E x t e n s i o n s
5 1
6 . R e f e r e n c e s
5 3
A p p e n d i x A : T i m e l i n e
6 3
A p p e n d i x B : K e y H i s t o r i c a l F i g u r e s
6 7
C i t e a s : Z a i d i , W a q a r , & A l l a n D a f o e ( 2 0 2 1 ) : “ I n t e r n a t i o n a l C o n t r o l o f P o w e r f u l T e c h n o l o g y : L e s s o n s f r o m t h e B a r u c h
P l a n f o r N u c l e a r W e a p o n s ” , # 2 0 2 1 : 9 , C e n t r e f o r t h e G o v e r n a n c e o f A I , F u t u r e o f H u m a n i t y I n s t i t u t e , U n i v e r s i t y o f
O x f o r d
1
1 . I n t r o d u c t i o n
H u m a n i t y i s l i k e l y t o c o n f r o n t n o v e l p o w e r f u l t e c h n o l o g i e s a n d w e a p o n s i n t h e c o m i n g d e c a d e s , i n c l u d i n g
t h o s e w h i c h c o u l d e m e r g e f r o m d e v e l o p m e n t s i n a r t i f i c i a l i n t e l l i g e n c e . D e v e l o p i n g a n d d e p l o y i n g t h e s e i n a w a y
t h a t i s g o o d f o r h u m a n i t y ( t h e g o v e r n a n c e p r o b l e m ) m a y b e h a r d , o w i n g t o s t r u c t u r a l f e a t u r e s o f t h e r i s k s a n d
d e c i s i o n - m a k i n g c o n t e x t .
2
O n t h e o t h e r h a n d , r a d i c a l l e v e l s o f c o o p e r a t i o n b e c o m e m o r e f e a s i b l e i n l i g h t o f t r u l y e x i s t e n t i a l l y d a n g e r o u s
t e c h n o l o g y : i n w h i c h t h e g a i n s f r o m c o o r d i n a t i o n a r e t r e m e n d o u s , t h e l o s s e s f r o m f a i l e d c o o r d i n a t i o n t e r r i b l e ,
3
a n d w h e r e m o s t a c t o r s ’ l o n g - t e r m i n t e r e s t s a r e a l i g n e d . W e m i g h t h o p e t h a t p o w e r f u l i n d i v i d u a l s w o u l d s e t a s i d e
t h e i r n a r r o w s e l f - i n t e r e s t a n d p e r s p e c t i v e s , a n d w o r k t o g e t h e r t o s e c u r e f o r h u m a n i t y a f l o u r i s h i n g f u t u r e . T h e y
m i g h t d o t h i s b e c a u s e d e e p d o w n t h e y b e l i e v e t h i s i s w h a t m o s t m a t t e r s ; b e c a u s e o f s t a t u s m o t i v a t i o n s t o l e a v e a
l e g a c y o f s e c u r i n g t h i s h i s t o r i c a l a c h i e v e m e n t ; b e c a u s e o f s o c i a l p r e s s u r e f r o m t h e i r p e e r s , f a m i l y , o r t h e p u b l i c ;
o r f o r o t h e r m o t i v a t i o n s . W e h o p e t h a t i n d i v i d u a l s — w h e n c o n f r o n t e d w i t h a d e c i s i o n t h a t c o u l d t a k e h u m a n i t y
t o w a r d s f l o u r i s h i n g a n d a w a y f r o m e x i s t e n t i a l h a r m — w o u l d m a k e t h e r i g h t d e c i s i o n . B u t w o u l d t h e y ?
T h i s p a p e r l o o k s t o t h e d e v e l o p m e n t a n d a t t e m p t e d g o v e r n a n c e o f n u c l e a r t e c h n o l o g y f o r l e s s o n s o n t h i s
q u e s t i o n . I t f o c u s e s o n t h e u n c e r t a i n e a r l y y e a r s o f t h i s t e c h n o l o g y , e s p e c i a l l y 1 9 4 3 t o 1 9 5 1 . P o l i c y m a k e r s ,
s t a t e s m e n , s c i e n t i f i c a n d t e c h n i c a l s p e c i a l i s t s , a n d o t h e r i n t e l l e c t u a l s a t t e m p t e d t o u n d e r s t a n d t h e n a t u r e a n d
i m p a c t o f n u c l e a r t e c h n o l o g y a n d d e v i s e g o v e r n a n c e f o r t h e d a n g e r s t h a t m a n y s a w . T h e y s a w n u c l e a r
t e c h n o l o g y , j u s t a s w e s e e s o m e t e c h n o l o g i e s t o d a y , a s b r i n g i n g g r e a t p r o m i s e , b u t a l s o g r e a t t h r e a t s . T h i s
d u a l - u s e n a t u r e l e d t h e m t o c o n c l u d e t h a t s i m p l y b a n n i n g n u c l e a r t e c h n o l o g y w a s n o t a n o p t i o n , a s t h e n i t s
p o t e n t i a l b e n e f i t s w o u l d b e l o s t . I n s t e a d , p r o p o n e n t s a r g u e d , t h e w o r l d n e e d e d t o d e v i s e i n t e r n a t i o n a l
g o v e r n a n c e m e c h a n i s m s w h i c h w o u l d b o t h r e d u c e t h e r i s k s b u t a l s o a l l o w t h e b e n e f i c i a l o u t c o m e s t o e m e r g e .
4
T h i s d o c u m e n t i s o r g a n i z e d b y l e s s o n s a n d r e c o m m e n d a t i o n s . T h e s e d i s t i l a n d g e n e r a l i z e t h e l e s s o n s t h a t m a y b e
a p p l i c a b l e t o f u t u r e e f f o r t s t o c o n t r o l t h o s e t e c h n o l o g i e s t h a t p o s e s i g n i f i c a n t r i s k s o f m i s u s e a n d a c c i d e n t , b u t
t h a t a l s o c o m e w i t h s u b s t a n t i a l m i l i t a r y a n d e c o n o m i c a d v a n t a g e .
W e b e l i e v e t h e s e l e s s o n s w i l l h e l p t h o s e
5
p a r t i c i p a t i n g i n c o n v e r s a t i o n s o n t h e g o v e r n a n c e o f s u c h t e c h n o l o g i e s b y h i g h l i g h t i n g h i s t o r i c a l p a r a l l e l s a n d
e x p a n d i n g t h e s p a c e o f c o n c e i v a b l e , a n d c o n s i d e r e d , p o l i t i c a l d y n a m i c s a n d o p p o r t u n i t i e s .
2
O n t h e s t r u c t u r a l c a u s e s o f A I r i s k , s e e R e m c o Z w e t s l o o t a n d A l l a n D a f o e , “ T h i n k i n g A b o u t R i s k s f r o m A I : A c c i d e n t s ,
M i s u s e a n d S t r u c t u r e ” ,
L a w f a r e
( 1 1 F e b r u a r y 2 0 1 9 ) . A v a i l a b l e a t :
h t t p s : / / w w w . l a w f a r e b l o g . c o m / t h i n k i n g - a b o u t - r i s k s - a i - a c c i d e n t s - m i s u s e - a n d - s t r u c t u r e
, a c c e s s e d 2 7 M a y 2 0 1 9 .
3
N i c k B o s t r o m , “ T h e V u l n e r a b l e W o r l d H y p o t h e s i s ” ,
G l o b a l P o l i c y
( 2 0 1 9 ) , d o i : 1 0 . 1 1 1 1 / 1 7 5 8 - 5 8 9 9 . 1 2 7 1 8 ; N i c k
B o s t r o m , A l l a n D a f o e , a n d C a r r i c k F l y n n , “ P o l i c y D e s i d e r a t a f o r S u p e r i n t e l l i g e n t A I : A V e c t o r F i e l d A p p r o a c h ” ( 2 0 1 8 ) .
A v a i l a b l e a t :
h t t p s : / / w w w . n i c k b o s t r o m . c o m / p a p e r s / a i p o l i c y . p d f
, a c c e s s e d 2 8 A p r i l 2 0 2 0 .
4
M a n y p r o p o s a l s f o r i n t e r n a t i o n a l c o n t r o l b e g a n b y s t a t i n g t h a t n u c l e a r e n e r g y h a s a d u a l n a t u r e . S e e , f o r e x a m p l e , t h e
b e g i n n i n g o f t h e B a r u c h P l a n :
U n i t e d S t a t e s A t o m i c E n e r g y P r o p o s a l s
( W a s h i n g t o n , D . C : U . S . D e p a r t m e n t o f S t a t e , 1 9 4 6 ) ,
p . 1 . T h i s i n t e r e s t i n t h e i n t e r n a t i o n a l g o v e r n a n c e o f n u c l e a r w e a p o n s r e m a i n s t o d a y : t h e I n t e r n a t i o n a l C a m p a i g n t o
A b o l i s h N u c l e a r W e a p o n s w o n t h e N o b e l P e a c e P r i z e i n 2 0 1 7 .
5
O n g l o b a l c a t a s t r o p h i c r i s k s e e : N i c k B o s t r o m a n d M i l a n M . C i r k o v i c ( e d s . ) ,
G l o b a l C a t a s t r o p h i c R i s k s
( O x f o r d : O x f o r d
U n i v e r s i t y P r e s s , 2 0 0 8 ) , p p . 3 - 5 . T h i s v o l u m e h i g h l i g h t s f o u r s u c h t e c h n o l o g i e s : n u c l e a r w e a p o n s , A . I . , b i o t e c h n o l o g y , a n d
n a n o t e c h n o l o g y ,
2
T h r o u g h w h i c h p r o c e s s e s m i g h t g o v e r n a n c e b e d i s c u s s e d a n d s e t u p ? W h a t p r o b l e m s m i g h t p o l i c y m a k e r s a n d
o t h e r i n t e r e s t e d p a r t i e s n e e d t o a n t i c i p a t e w h e n t h i n k i n g a b o u t g o v e r n a n c e ? W h o m i g h t s u p p o r t p r o p o s a l s f o r
g o v e r n a n c e , a n d h o w a n d w h y ? H o w s i n c e r e o r c y n i c a l w i l l p a r t i c i p a n t s b e ? H o w l i k e l y i s i t t h a t k e y a c t o r s w i l l
m i s u n d e r s t a n d t h e p r o b l e m o r m i s c o m m u n i c a t e t h e i r p r e f e r e n c e s ?
T h i s r e p o r t b e g i n s w i t h a h i s t o r i c a l o v e r v i e w o f p r o p o s a l s f o r t h e “ i n t e r n a t i o n a l c o n t r o l o f a t o m i c e n e r g y ” ( t h a t
i s , i n t e r n a t i o n a l r e g u l a t i o n o f a t o m i c w e a p o n s a n d u n d e r l y i n g t e c h n o l o g i e s , s c i e n c e s , a n d m a t e r i a l s ) b e t w e e n
1 9 4 4 a n d 1 9 4 6 , f o l l o w e d b y k e y d a t e s a n d s h o r t s u m m a r i e s o f t h e k e y p r o p o s a l s . I n s u m m a r y , w e f i n d t h a t
r a d i c a l s c h e m e s f o r i n t e r n a t i o n a l g o v e r n a n c e c a n g e t w i d e s p r e a d s u p p o r t , e v e n f r o m s k e p t i c s , b u t t h a t t h e
s u p p o r t c a n b e t e n u o u s a n d f l e e t i n g . T e c h n i c a l e x p e r t s c a n b o l s t e r s u p p o r t , b u t m u d d l e d p o l i c y m a k i n g , s e c r e c y ,
a n d c o n c e r n s o v e r s e c u r i t y c a n u n d e r m i n e i t . W e h i g h l i g h t t h e f o l l o w i n g l e s s o n s f o r t h o s e t h i n k i n g a b o u t
t e c h n o l o g i c a l g o v e r n a n c e t o d a y :
1 .
R a d i c a l p r o p o s a l s w h i c h w o u l d n o r m a l l y a p p e a r n a i v e o r e x t r e m e m a y , i n t h e r i g h t c i r c u m s t a n c e s , g a i n
t r a c t i o n a n d b e s e r i o u s l y p r o p o s e d , d i s c u s s e d , a n d e v e n a d o p t e d a s o f f i c i a l p o l i c y .
2 .
G r o u p s o r c o a l i t i o n s s u p p o r t i n g ( o r o p p o s i n g ) i n t e r n a t i o n a l c o n t r o l w i l l c o n t a i n i n d i v i d u a l s w h o h a v e
d i f f e r e n t r e a s o n s a n d r a t i o n a l e f o r t h e i r s u p p o r t ( o r o p p o s i t i o n ) .
3 .
T h e s u p p o r t o f r e a l i s t s i s p o s s i b l e a n d p o s s i b l y e v e n c r u c i a l f o r i n t e r n a t i o n a l c o n t r o l t o b e c o m e p o l i c y .
6
4 .
S e c r e c y a n d s e c u r i t y w i l l p l a y a c e n t r a l r o l e i n d i s c u s s i o n s o n t h e g o v e r n a n c e o f p o w e r f u l t e c h n o l o g i e s .
5 .
T h e p u b l i c s p h e r e w i l l l i k e l y h a v e a p o w e r f u l i m p a c t o n d e b a t e s r e g a r d i n g i n t e r n a t i o n a l c o n t r o l .
6 .
T e c h n i c a l e x p e r t s a n d s p e c i a l i s t s ( s c i e n t i s t s , e n g i n e e r s , t e c h n i c i a n s , a c a d e m i c s ) h a v e s i g n i f i c a n t p o w e r t o
s h a p e p r o p o s a l s a n d p o l i c y , t h o u g h t h e i r o p p o n e n t s m a y c r i t i c i z e t h e m f o r p o l i t i c a l n a i v e t y .
7 .
P o l i c y m a k i n g i n v o l v e s s i g n i f i c a n t m u d d l i n g t h r o u g h , r a t h e r t h a n g r a n d s t r a t e g y . I t i s a l s o d e e p l y
a f f e c t e d b y d o m e s t i c p o l i t i c s a n d o f t e n d e v e l o p s o n t h e b a s i s o f s h o r t - t e r m o b j e c t i v e s , p o o r l y
t h o u g h t - o u t c r i t e r i a , a n d p o o r q u a l i t y i n f o r m a t i o n . P o l i c y m a k i n g m a y d e v e l o p i n u n e x p e c t e d
d i r e c t i o n s o r f o r t h e e x p e c t e d r e a s o n s .
8 .
A c h i e v i n g a g r e e m e n t o n a w o r k a b l e s c h e m e f o r i n t e r n a t i o n a l c o n t r o l i s d i f f i c u l t .
9 .
A t t e m p t s a t c o o p e r a t i o n c o m e w i t h r i s k s o f s t r a t e g i c , d i p l o m a t i c , p o l i t i c a l , a n d t e c h n o l o g i c a l l o s s e s .
T h e l e s s o n s h a v e b e e n o r g a n i z e d s o t h a t r e a d e r s m a y s k i p t h e h i s t o r i c a l c a s e e x p o s i t i o n s i f t h e y w i s h . T h e c a s e s
d o , h o w e v e r , f l e s h o u t t h e b a r e b o n e s l e s s o n s , a n d u n p a c k a n d e x p l o r e t h e v a r i o u s a s p e c t s o f t h e l e s s o n i n m o r e
d e t a i l .
F o r t h o s e w a n t i n g m o r e d e t a i l , w e h a v e i n c l u d e d a l i s t o f k e y e v e n t s a n d s h o r t b i o g r a p h i e s o f c e n t r a l f i g u r e s i n
t h e a p p e n d i c e s . F o r f u r t h e r r e a d i n g , w e w o u l d a l s o p o i n t r e a d e r s t o t h e r i c h h i s t o r i c a l l i t e r a t u r e o n t h e p o l i t i c s o f
a t o m i c e n e r g y i n i t s e a r l y y e a r s , m u c h o f w h i c h i s c i t e d i n t h e f o o t n o t e s a n d l i s t e d i n t h e
R e f e r e n c e s
s e c t i o n . W e
h a v e r e l i e d o n a v a r i e t y o f s e c o n d a r y ( a n d s o m e p r i m a r y ) s o u r c e s b u t h a v e f o u n d G r e g g H e r k e n ’ s
T h e W i n n i n g
6
B y “ r e a l i s t s ” w e m e a n t h o s e w h o u n d e r s t a n d i n t e r n a t i o n a l r e l a t i o n s i n t e r m s o f p o w e r a n d n a t i o n a l i n t e r e s t , a n d p r e f e r
p o l i c i e s w h i c h p r e s e r v e a n d s t r e n g t h e n t h e i r s t a t e i n r e l a t i o n t o o t h e r s . J a c k D o n n e l l y ,
R e a l i s m a n d I n t e r n a t i o n a l R e l a t i o n s
( C a m b r i d g e : C a m b r i d g e U n i v e r s i t y P r e s s , 2 0 0 4 ) , p p . 7 - 8 .
3
W e a p o n
t o b e t h e s i n g l e m o s t d e t a i l e d a n d r e l i a b l e s o u r c e .
W e w o u l d r e c o m m e n d t h i s a s t h e f i r s t p o r t o f c a l l f o r
7
a n y r e a d e r i n t e r e s t e d i n f u r t h e r e x p l o r i n g t h e h i s t o r y o f t h e i n t e r n a t i o n a l c o n t r o l o f n u c l e a r w e a p o n s .
N u c l e a r T e c h n o l o g y a s a n A n a l o g y
H i s t o r y c a n p r o v i d e a r i c h s o u r c e f o r i n s i g h t a b o u t n o v e l p o l i c y c h a l l e n g e s . T o u n d e r s t a n d t h e c h a l l e n g e s o f
g o v e r n i n g t o d a y ’ s e m e r g i n g p o w e r f u l t e c h n o l o g i e s , o n e c a n e x a m i n e a t t e m p t s a t t h e g o v e r n a n c e o f e a r l i e r
p o w e r f u l t e c h n o l o g i e s w h e n t h e y f i r s t e m e r g e d . O f t h e v a r i o u s t e c h n o l o g i e s f o r w h i c h i n t e r n a t i o n a l g o v e r n a n c e
r e g i m e s w e r e c r e a t e d o r c o n t e m p l a t e d i n t h e t w e n t i e t h c e n t u r y ( i n c l u d i n g , f o r e x a m p l e , a v i a t i o n , c h e m i c a l a n d
b i o l o g i c a l w e a p o n s , t e l e c o m m u n i c a t i o n s , a n d t h e i n t e r n e t ) , n u c l e a r t e c h n o l o g y s t a n d s o u t a s a p a r t i c u l a r l y
p r o m i s i n g c a n d i d a t e f o r s t u d y o f t h e p r e s s i n g , b u t t h o r n y , p r o b l e m o f i n t e r n a t i o n a l c o n t r o l .
I n p a r t i c u l a r , w e
8
w o u l d h i g h l i g h t t h e f o l l o w i n g p r o p e r t i e s w h i c h m a k e t h i s c a s e r e l e v a n t t o u n d e r s t a n d i n g e f f o r t s t o c o n t r o l a
f u t u r e p o w e r f u l t e c h n o l o g y ( s u c h a s A I ) :
( a )
N u c l e a r t e c h n o l o g y w a s m a r k e d o u t a s a p o w e r f u l t e c h n o l o g y w h e n i t w a s f i r s t r e v e a l e d , a n d
p o l i c y m a k i n g w a s m a d e w i t h i n a c o n t e x t t h a t t o o k i t s p o t e n t i a l i m p a c t s e r i o u s l y .
( b )
B e c a u s e o f t h e s u d d e n w a y t h e a t o m i c b o m b w a s r e v e a l e d , a n d i t s s e e m i n g l y e s o t e r i c n a t u r e , t h e r e w a s
s i g n i f i c a n t u n c e r t a i n t y a b o u t i t s i m p a c t . C o n s e q u e n t l y , t h e r e w a s a r i c h p u b l i c a n d p o l i c y m a k i n g
d e b a t e a b o u t t h e n a t u r e o f t h i s t e c h n o l o g y a n d i t s i m p a c t . A s w e l l a s s t r a t e g i c a n d p o l i t i c a l d i m e n s i o n s ,
t h i s d e b a t e i n c l u d e d a n e t h i c a l d i m e n s i o n .
( c )
M a n y p e o p l e , i n c l u d i n g m a n y e l i t e s ,
p e r c e i v e d n u c l e a r t e c h n o l o g y a s a n e x i s t e n t i a l r i s k a n d s o
9
e n g e n d e r e d a r i c h p o l i c y m a k i n g d e b a t e o n i n t e r n a t i o n a l g o v e r n a n c e , k n o w n t h e n a s “ i n t e r n a t i o n a l
c o n t r o l . ”
( d )
E l e m e n t s o f n a t i o n a l c o m p e t i t i o n a n d n e g o t i a t i o n , a n d o f a t e c h n o l o g i c a l a r m s r a c e , w e r e p r e s e n t
d u r i n g t h e e a r l y h i s t o r y o f n u c l e a r w e a p o n s .
( e )
N u c l e a r t e c h n o l o g y r e s t e d o n c o m p l e x , r a p i d l y d e v e l o p i n g s c i e n c e .
N e v e r t h e l e s s , r e a d e r s s h o u l d b e a w a r e t h a t t h e r e a r e a n u m b e r o f w a y s i n w h i c h t h i s h i s t o r i c a l m o m e n t a n d
n u c l e a r t e c h n o l o g y a r e a p o o r a n a l o g y f o r t h e f u t u r e g o v e r n a n c e o f p o w e r f u l t e c h n o l o g i e s . C o n s i d e r t h e
f o l l o w i n g d i s a n a l o g i e s t o A I :
7
G r e g g H e r k e n ,
T h e W i n n i n g W e a p o n : T h e A t o m i c B o m b i n t h e C o l d W a r 1 9 4 5 - 1 9 5 0
2 n d e d i t i o n ( P r i n c e t o n , N J :
P r i n c e t o n U n i v e r s i t y P r e s s , 1 9 8 8 ) . H e n c e f o r t h r e f e r r e d t o a s H e r k e n i n t h e f o o t n o t e s .
8
T h e r e i s s o m e d i s c u s s i o n o f s i m i l a r i t i e s a n d d i s s i m i l a r i t i e s b e t w e e n A I a n d n u c l e a r t e c h n o l o g y i n K e n n e t h P a y n e ,
“ A r t i f i c i a l I n t e l l i g e n c e : A R e v o l u t i o n i n S t r a t e g i c A f f a i r s ? ” ,
S u r v i v a l : G l o b a l P o l i t i c s a n d S t r a t e g y
6 0 , 5 ( 2 0 1 8 ) , p p . 7 - 3 2 ;
M a t t h i j s M . M a a s , “ H o w V i a b l e i s I n t e r n a t i o n a l A r m s C o n t r o l f o r M i l i t a r y A r t i f i c i a l I n t e l l i g e n c e ? T h r e e L e s s o n s f r o m
N u c l e a r W e a p o n s ” ,
C o n t e m p o r a r y S e c u r i t y P o l i c y
4 0 , 3 ( 2 0 1 9 ) , p p . 2 8 5 - 3 1 1 . O t h e r t y p e s o f h i s t o r i c a l c a s e s a n d
c o m p a r i s o n s m a y b e f r u i t f u l t o o , s e e f o r e x a m p l e : P e t e r C i h o n , M a t t h i j s M . M a a s , a n d L u k e K e m p , “ S h o u l d A r t i f i c i a l
I n t e l l i g e n c e G o v e r n a n c e b e C e n t r a l i s e d ? D e s i g n L e s s o n s f r o m H i s t o r y ” ,
P r o c e e d i n g s o f t h e 2 0 2 0 A A A I / A C M C o n f e r e n c e o n
A I , E t h i c s , a n d S o c i e t y ( A I E S ’ 2 0 ) , F e b r u a r y 7 – 8 , 2 0 2 0 , N e w Y o r k , N Y , U S A
( 2 0 2 0 ) .
h t t p s : / / d o i . o r g / 1 0 . 1 1 4 5 / 3 3 7 5 6 2 7 . 3 3 7 5 8 5 7 .
9
B y “ e l i t e s ” w e m e a n t h o s e i n d i v i d u a l s o r g r o u p s w h o c a n h a v e a p o l i c y m a k i n g i m p a c t t h r o u g h d i r e c t i n t e r v e n t i o n ( a s
o p p o s e d t o , s a y , t h o u g h v o t i n g o r m a s s p r o t e s t ) . E l i t e s g e n e r a l l y i n c l u d e p o l i c y m a k e r s w i t h i n t h e s t a t e ( i n c l u d i n g
p o l i t i c i a n s , b u r e a u c r a t s , d i p l o m a t s , s e n i o r d e f e n s e o f f i c i a l s , a n d a d v i s o r s ) a n d t h o s e o u t s i d e , s u c h a s p r o m i n e n t n u c l e a r
s c i e n t i s t s o r i n d u s t r i a l i s t s .
4
●
P r i v a t e S e c t o r I n v o l v e m e n t
: A I i s p r i m a r i l y b e i n g d e v e l o p e d a n d d e p l o y e d b y t h e p r i v a t e s e c t o r , a n d
t h e p r i v a t e s e c t o r i s l i k e l y t o c o n t i n u e t o p u s h f o r w a r d t h e s c i e n c e o f A I i r r e s p e c t i v e o f w h a t
g o v e r n m e n t s d o . N u c l e a r t e c h n o l o g y i n i t s e a r l i e s t d e c a d e s w a s c o n t r o l l e d a n d f u n d e d b y s t a t e s .
●
S e c r e c y
: W h i l e n u c l e a r t e c h n o l o g i e s w e r e h e a v i l y g u a r d e d s e c r e t s ( t h o u g h t h e b a s i c s c i e n c e w a s b r o a d l y
k n o w n ) , a r t i f i c i a l i n t e l l i g e n c e t e c h n o l o g y i s m o r e i n t e r n a t i o n a l , b r o a d l y h e l d , a n d p u b l i c .
●
I m p a c t a n d P r o l i f e r a t i o n
: A I i s a l r e a d y a m a j o r e c o n o m i c t e c h n o l o g y a n d d e p l o y e d a r o u n d t h e
w o r l d , w h e r e a s t h e e c o n o m i c v a l u e o f n u c l e a r t e c h n o l o g y w a s u n c l e a r a n d i t w a s d e p l o y e d i n o n l y a f e w
l o c a t i o n s i n t h e p e r i o d 1 9 4 3 – 1 9 5 1 . A I i s d e p l o y e d a n d i n n o v a t e d i n a g r e a t e r n u m b e r o f f i e l d s a s
c o m p a r e d t o n u c l e a r t e c h n o l o g y , a n d o f f e r s g r e a t e r f u t u r e e c o n o m i c p o t e n t i a l . T h e b a r r i e r s t o e n t r y f o r
t h e d e v e l o p m e n t a n d d e p l o y m e n t o f A I a r e a l s o l o w e r t h a n i n t h e c a s e o f n u c l e a r t e c h n o l o g y .
●
D i s c e r n i b i l i t y o f R i s k s
: I t m a y b e e a s i e r t o u n d e r s t a n d h o w n u c l e a r w e a p o n s c o u l d b e d a n g e r o u s ,
w h e r e a s t h e a c c i d e n t r i s k s f r o m A I a r e m o r e s u b t l e , t h e o r y d e p e n d e n t , o r f a n t a s t i c s e e m i n g .
●
S a f e t y D i f f i c u l t y
: T h e a c c i d e n t r i s k s f r o m n u c l e a r w e a p o n s a r e l i k e l y e a s i e r t o m a n a g e t h a n f r o m A I ,
b e c a u s e n u c l e a r b o m b s o r p o w e r p l a n t s a r e n o t c o m p l e x a d a p t i v e ( i n t e l l i g e n t ) s y s t e m s .
●
V e r i f i c a t i o n
: I t i s e a s i e r t o u n i l a t e r a l l y v e r i f y n u c l e a r d e v e l o p m e n t s ( n u c l e a r t e s t s , I C B M
d e p l o y m e n t s ) , a n d i t a p p e a r s e a s i e r t o c o n t r o l t h e n u c l e a r s u p p l y c h a i n w i t h r e l a t i v e l y l o w d i s r u p t i o n o f
i n d u s t r y .
●
S t r a t e g i c V a l u e
: T h e s t r a t e g i c v a l u e o f n u c l e a r w e a p o n s p l a t e a u s o n c e o n e h a s s e c u r e s e c o n d - s t r i k e
c a p a b i l i t y , w h e r e a s f r o m t h e p r e s e n t v a n t a g e p o i n t , t h e r e i s n o o b v i o u s p l a t e a u i n A I ’ s s t r a t e g i c v a l u e .
F u r t h e r , t h e h i s t o r i c a l c o n t e x t f o r t h e e a r l y d e v e l o p m e n t o f n u c l e a r t e c h n o l o g y d i f f e r s i n i m p o r t a n t w a y s f r o m
t h e c u r r e n t a n d f u t u r e m o m e n t s i n w h i c h a t t e m p t s t o g o v e r n o t h e r p o w e r f u l t e c h n o l o g i e s m a y b e m a d e :
●
P o s t w a r C o n t e x t
: A t o m i c d e v e l o p m e n t o c c u r r e d a t t h e e n d o f a w a r w i d e l y s e e n a s c a t a s t r o p h i c . T h i s
l e d t o a v e r y d i f f e r e n t s o c i a l a n d p o l i t i c a l c o n t e x t w i t h i n w h i c h n u c l e a r w e a p o n s w e r e i n t r o d u c e d .
●
V i s c e r a l E x a m p l e o f D a n g e r
: T h e w o r l d w i t n e s s e d t h e u s e o f n u c l e a r w e a p o n s t o d e s t r o y c i t i e s a n d
s o m e o f t h e h o r r o r s t h i s e n t a i l e d . F u t u r e t e c h n o l o g y r i s k s m a y n o t p r o d u c e v i s c e r a l h a r m s i n a d v a n c e o f
a t t e m p t s t o g o v e r n t h e m .
●
S u p e r p o w e r R e l a t i o n s h i p s
: N u c l e a r c o n t r o l n e g o t i a t i o n s t o o k p l a c e b e t w e e n p o w e r s w h o w e r e a l l i e s
a n d h a d j u s t s u f f e r e d t h r o u g h t h i s w a r . T h e s e p o w e r s a l s o h a d i n c o m p a t i b l e p o l i t i c a l - e c o n o m i c m o d e l s
a n d s o f o u n d t h e m s e l v e s i n m u c h m o r e z e r o - s u m r e l a t i o n s t h a n t h e g r e a t p o w e r s o f t o d a y .
●
I n f o r m a t i o n
: T h e g r e a t p o w e r s h a d l e s s c u l t u r a l a n d i d e o l o g i c a l c o m m o n a l i t y , a n d l e s s i n f o r m a t i o n
a b o u t e a c h o t h e r , t h a n d o t h e g r e a t p o w e r s , a n d t h e i r p u b l i c s , o f t o d a y .
R e a d e r s s h o u l d b e a w a r e o f t w o f u r t h e r m e t h o d o l o g i c a l c a v e a t s :
●
n = 1
. T o s o m e e x t e n t , t h i s h i s t o r i c a l p e r i o d r e p r e s e n t s a s i n g l e o b s e r v a t i o n ( n = 1 ) i n t h a t a s i n g l e l a r g e
s h o c k t o d e c i s i o n p r o c e s s e s c o u l d h a v e l e d t o d i f f e r e n t o u t c o m e s . T h e i m p l i c a t i o n o f t h i s i s t h a t w e
s h o u l d n o t p r i m a r i l y u s e t h e o u t c o m e a s o u r e v i d e n c e , b u t s h o u l d i n s t e a d i n s p e c t a l l t h e i n f o r m a t i v e
h i s t o r i c a l m o m e n t s t h r o u g h o u t t h e e p i s o d e f o r i n s i g h t i n t o h i s t o r i c a l d y n a m i c s a n d m e c h a n i s m s . F o r
e x a m p l e , w e c a n l e a r n f r o m t h e r i c h w a y s i n w h i c h d e c i s i o n m a k e r s r e s p o n d e d t o i n f o r m a t i o n , f o r m e d
b e l i e f s , a n d d e v i s e d s t r a t e g y .
●
S o u r c e s
. O u r a n a l y s i s o f S o v i e t p r o p o s a l s a n d r e s p o n s e s i s r e s t r i c t e d b y t h e r e l a t i v e p a u c i t y o f h i s t o r i c a l
w o r k o n t h e S o v i e t U n i o n . C o n s e q u e n t l y , w e f o c u s l a r g e l y o n t h e U n i t e d S t a t e s ( o n w h i c h t h e r e i s a
5
r i c h e r h i s t o r i c a l l i t e r a t u r e ) , b u t h e r e t o o w e a r e r e s t r i c t e d b y w h a t h i s t o r i a n s h a v e c h o s e n t o e x p l o r e a n d
e l u c i d a t e .
1 0
T h e s e c a v e a t s h a v i n g b e e n s t a t e d , t h i s c a s e s t u d y r e p r e s e n t s a r a r e h i s t o r i c a l m o m e n t w h e n g r e a t p o w e r s
s e r i o u s l y d i s c u s s e d s t r a t e g i e s f o r a v o i d i n g a n a r m s r a c e i n a n e w t e c h n o l o g y ,
a n d w h e r e i n f l u e n t i a l p e o p l e
1 1
w i t h i n t h e s t a t e w i t h t h e t e c h n o l o g i c a l m o n o p o l y s e r i o u s l y c o n s i d e r e d g i v i n g u p t h e i r m o n o p o l y . F u r t h e r m o r e ,
t h i s h i s t o r i c a l e p i s o d e t o o k p l a c e b e t w e e n r e l a t i v e l y m o d e r n g r e a t p o w e r s a n d a t a t i m e w h e n U . S . e l i t e a n d
p u b l i c c u l t u r e w a s n o t e n t i r e l y d i s s i m i l a r t o t o d a y ’ s ; f o r e x a m p l e , t h e m e d i a w a s i m p o r t a n t i n i n f o r m i n g t h e
p u b l i c a n d s h a p i n g i t s o p i n i o n o n m a j o r e v e n t s , p o l i t i c i a n s t o o k p u b l i c o p i n i o n i n t o a c c o u n t w h e n m a k i n g
p o l i c y d e c i s i o n s , i n t e r s e r v i c e r i v a l r y p l a y e d a r o l e i n s o m e p o l i c y d e c i s i o n s , a n d p o l i c y m a k i n g w a s d o n e t h r o u g h
a m i x o f c o m m i t t e e s , e x p e r t s , c a r e e r s t a t e s m e n , a n d t r u s t e d a d v i s o r s .
1 0
T h e r e i s n e w m a t e r i a l a p p e a r i n g i n t o t h e p u b l i c d o m a i n a t r e g u l a r i n t e r v a l s , s e e f o r e x a m p l e D a v i d H o l l o w a y , “ T h e
S o v i e t U n i o n a n d t h e B a r u c h P l a n ” .
h t t p s : / / w w w . w i l s o n c e n t e r . o r g / b l o g - p o s t / s o v i e t - u n i o n - a n d - b a r u c h - p l a n
, a c c e s s e d 1 3
J u n e 2 0 2 0 . W e u s e w h a t w e c o n s i d e r t o b e t h e l a t e s t a n d b e s t h i s t o r i c a l w o r k .
1 1
T h e n e x t b e s t a n a l o g y m a y b e e f f o r t s t o w a r d s t h e i n t e r n a t i o n a l c o n t r o l o f a v i a t i o n . S e e W a q a r Z a i d i , “ ‘ A v i a t i o n W i l l
E i t h e r D e s t r o y o r S a v e O u r C i v i l i z a t i o n ’ : P r o p o s a l s f o r t h e I n t e r n a t i o n a l C o n t r o l o f A v i a t i o n , 1 9 2 0 - 1 9 4 5 ” ,
J o u r n a l o f
C o n t e m p o r a r y H i s t o r y
4 6 , 1 ( 2 0 1 1 ) , p p . 1 5 0 - 7 8 .
6
2 . H i s t o r i c a l O v e r v i e w
S u m m a r y
B y t h e s t a r t o f W o r l d W a r I I , s c i e n t i s t s a r o u n d t h e w o r l d w e r e a w a r e t h a t t h e c o n s t r u c t i o n o f a b o m b b a s e d o n
t h e r e l e a s e o f a t o m i c e n e r g y w a s t h e o r e t i c a l l y p o s s i b l e . B r i t a i n w a s t h e f i r s t t o s t a r t a c o n c e r t e d b o m b p r o g r a m ,
j o i n e d s o o n a f t e r b y t h e U n i t e d S t a t e s , G e r m a n y , J a p a n , a n d t h e S o v i e t U n i o n . T h e G e r m a n a n d J a p a n e s e
p r o g r a m s d i d n o t p r o g r e s s f a r . B r i t a i n , f a c e d w i t h m o r e p r e s s i n g r e s o u r c e r e q u i r e m e n t s , e v e n t u a l l y p a u s e d i t s
p r o g r a m a n d t r a n s f e r r e d i t s e x p e r t i s e i n t o t h e M a n h a t t a n P r o j e c t , j o i n i n g t h e U . S . p r o g r a m a s a j u n i o r p a r t n e r .
T h e M a n h a t t a n P r o j e c t , b e g u n i n O c t o b e r 1 9 4 1 , l e d t o a w o r k i n g b o m b t h a t w a s t e s t e d i n J u l y 1 9 4 5 . A t o m i c
b o m b s w e r e d r o p p e d o n t h e J a p a n e s e c i t i e s o f H i r o s h i m a o n A u g u s t 6 a n d N a g a s a k i o n A u g u s t 9 . J a p a n
a n n o u n c e d i t s s u r r e n d e r o n A u g u s t 1 5 , a n d t h e s i g n i n g o f t h e f o r m a l s u r r e n d e r t r e a t y o n S e p t e m b e r 2 b r o u g h t
t h e S e c o n d W o r l d W a r o f f i c i a l l y t o a c l o s e . T h e u s e o f t h e a t o m i c b o m b l e d t o a n a c c e l e r a t i o n o f t h e S o v i e t
b o m b p r o j e c t a n d a r e s t a r t o f t h e B r i t i s h p r o j e c t .
E v e n b e f o r e t h e b o m b w a s u s e d , s c i e n t i s t s e x p r e s s e d c o n c e r n a b o u t i t s d e s t r u c t i v e n e s s a n d a p o s s i b l e a r m s r a c e
a f t e r t h e w a r . S e n i o r D a n i s h p h y s i c i s t N i e l s B o h r b r o u g h t t h e s e c o n c e r n s t o t h e a t t e n t i o n o f B r i t i s h P r i m e
M i n i s t e r W i n s t o n C h u r c h i l l i n M a y 1 9 4 4 a n d U . S . P r e s i d e n t F r a n k l i n D . R o o s e v e l t i n A u g u s t 1 9 4 4 . B y m i d
1 9 4 5 , s c i e n t i s t s w o r k i n g o n t h e M a n h a t t a n P r o j e c t a l s o b e c a m e c o n c e r n e d a b o u t t h e i m p a c t o f a t o m i c w e a p o n s
a n d i s s u e d a s e r i e s o f w a r n i n g s t o t h e g o v e r n m e n t . M a n y o f t h e s u g g e s t i o n s f o r d e a l i n g w i t h t h e b o m b c a l l e d f o r
t h e “ i n t e r n a t i o n a l c o n t r o l o f a t o m i c e n e r g y ” ( t h a t i s , e f f e c t i v e i n t e r n a t i o n a l r e g u l a t i o n o f a t o m i c w e a p o n s a n d
t h e u n d e r l y i n g s c i e n c e a n d t e c h n o l o g y t h r o u g h m u l t i l a t e r a l a g r e e m e n t s o r a n i n t e r n a t i o n a l o r g a n i z a t i o n ) .
V a r i o u s p r o p o s a l s f o r i n t e r n a t i o n a l c o n t r o l w e r e m a d e f r o m l a t e 1 9 4 4 o n w a r d s . T h e s e b e c a m e w i d e s p r e a d a f t e r
t h e a t o m i c b o m b w a s m a d e p u b l i c i n A u g u s t 1 9 4 5 , a n d b y t h e e n d o f t h e y e a r , s c i e n t i s t s h a d o r g a n i z e d
t h e m s e l v e s i n t o v a r i o u s g r o u p s c a l l i n g f o r i n t e r n a t i o n a l c o n t r o l . S t a t e o f f i c i a l s a l s o , a t t i m e s , c o n s i d e r e d
a d o p t i n g i n t e r n a t i o n a l c o n t r o l a s p o l i c y , a n d t h e r e w a s d i s c u s s i o n o n a t o m i c m a t t e r s w i t h t h e S o v i e t U n i o n . I n
l a t e D e c e m b e r 1 9 4 5 , S t a l i n a g r e e d t o t h e f o r m a t i o n o f a U n i t e d N a t i o n s A t o m i c E n e r g y C o m m i s s i o n t o s t u d y
t h e “ c o n t r o l o f a t o m i c e n e r g y , ” a n d t h e U n i t e d N a t i o n s G e n e r a l A s s e m b l y a u t h o r i z e d i t s f o r m a t i o n i n J a n u a r y
1 9 4 6 .
I n J a n u a r y 1 9 4 6 , S e c r e t a r y o f S t a t e J a m e s F . B y r n e s a u t h o r i z e d t h e f o r m a t i o n o f a c o m m i t t e e ( c h a i r e d b y U n d e r
S e c r e t a r y o f S t a t e D e a n A c h e s o n a n d e x - C h a i r m a n o f t h e T e n n e s s e e V a l l e y A u t h o r i t y D a v i d L i l i e n t h a l ) t o s t u d y
t h e i n t e r n a t i o n a l c o n t r o l o f a t o m i c e n e r g y . T h i s c o m m i t t e e i n t u r n a s k e d a g r o u p o f c o n s u l t a n t s ( l e d b y
p r o m i n e n t p h y s i c i s t J . R o b e r t O p p e n h e i m e r ) t o p r e p a r e a p o l i c y p r o p o s a l f o r i n t e r n a t i o n a l c o n t r o l f o r
g o v e r n m e n t c o n s i d e r a t i o n . T h e c o m m i t t e e c o m p l e t e d i t s d e t a i l e d p l a n f o r i n t e r n a t i o n a l c o n t r o l , d u b b e d t h e
A c h e s o n - L i l i e n t h a l R e p o r t , i n M a r c h 1 9 4 6 . T h e r e p o r t w a s a d o p t e d , w i t h i m p o r t a n t m o d i f i c a t i o n s , b y t h e f i r s t
U . S . r e p r e s e n t a t i v e t o t h e n e w l y f o r m e d U n i t e d N a t i o n s A t o m i c E n e r g y C o m m i s s i o n ( U N A E C ) , B e r n a r d
B a r u c h , w h o p r e s e n t e d h i s s o - c a l l e d B a r u c h P l a n a t t h e U N A E C i n J u n e 1 9 4 6 . T h e S o v i e t U n i o n , i m p l i c i t l y
r e j e c t i n g t h i s p l a n , r e s p o n d e d w i t h i t s o w n p r o p o s a l a f e w d a y s l a t e r ( t h e G r o m y k o P l a n ) .
7
S u b s e q u e n t n e g o t i a t i o n s w i t h t h e S o v i e t U n i o n w e r e c a r r i e d o u t a m i d s t d e t e r i o r a t i n g r e l a t i o n s b e t w e e n t h e
s u p e r p o w e r s a n d f a i l e d b y t h e e n d o f t h e y e a r . I n o u r a s s e s s m e n t o f t h i s c a s e , f a i l u r e w a s o v e r d e t e r m i n e d ( s e e
S e c t i o n
C o u l d I n t e r n a t i o n a l C o n t r o l H a v e S u c c e e d e d
) . T h e r e w a s a g r e a t d i v e r g e n c e i n U . S . - S o v i e t e x p e c t a t i o n s
a n d c o n f l i c t i n g i n t e r e s t s a r o u n d t h e w o r l d . M i s t r u s t h a d a l s o b e e n g r o w i n g s i n c e e a r l y 1 9 4 5 . B y m i d 1 9 4 6 , t h e
U . S . a d m i n i s t r a t i o n h a d g i v e n u p w h a t e v e r h o p e i t h a d i n i n t e r n a t i o n a l c o n t r o l a n d o n l y c a r r i e d o u t
n e g o t i a t i o n s f o r p r o p a g a n d a p u r p o s e s . T h e S o v i e t U n i o n s i m i l a r l y w a s i n t e r e s t e d i n t h e p r o p a g a n d a v a l u e o f
n e g o t i a t i o n s . I t h o p e d t o g e n e r a t e n e g a t i v e p u b l i c i t y f o r t h e U . S . a n d e x t r a c t a s m u c h i n f o r m a t i o n a s p o s s i b l e o n
t h e U . S . p r o g r a m . T h e f i n a l v o t e i n t h e S e c u r i t y C o u n c i l i n D e c e m b e r 1 9 4 6 h a d 1 0 U N A E C v o t e s i n f a v o r , a n d
t w o a b s t e n t i o n s ( t h e S o v i e t U n i o n a n d P o l a n d ) ; t h e s e a b s t e n t i o n s w e r e u n d e r s t o o d a s a n e f f e c t i v e v e t o .
B o t h t h e U . S . ’ s a n d t h e S o v i e t U n i o n ’ s a t o m i c p r o g r a m s c o n t i n u e d u n h i n d e r e d w h i l s t t h e n e g o t i a t i o n s w e r e
c a r r i e d o u t . U . S . a t o m i c b o m b s b e c a m e m o r e a d v a n c e d a n d i n c r e a s e d i n s i z e a n d n u m b e r i n t h e l a t e f o r t i e s . T h e
S o v i e t U n i o n e v e n t u a l l y c a r r i e d o u t i t s f i r s t a t o m i c b o m b t e s t i n A u g u s t 1 9 4 9 , c a t c h i n g m o s t U . S . i n t e l l i g e n c e
a n d m i l i t a r y p l a n n e r s b y s u r p r i s e . S t i m u l a t e d b y t h i s , t h e U n i t e d S t a t e s d e v e l o p e d t h e m u c h m o r e p o w e r f u l
h y d r o g e n b o m b , t e s t i n g i t i n N o v e m b e r 1 9 5 2 ; t h e S o v i e t U n i o n f o l l o w e d s o o n t h e r e a f t e r , i n N o v e m b e r 1 9 5 5 .
T h r o u g h t h e s e y e a r s , t h e q u a n t i t y o f a t o m i c b o m b s , a n d t h e n h y d r o g e n b o m b s , i n c r e a s e d e x p o n e n t i a l l y , f r o m
l e s s t h a n 2 0 i n 1 9 4 7 t o m o r e t h a n 1 0 0 i n 1 9 4 9 a n d t o m o r e t h a n 1 0 , 0 0 0 b y 1 9 5 9 ( s e e
F i g u r e 1
) .
1 2
A l o n g s i d e t h e d e v e l o p m e n t o f n u c l e a r a r s e n a l s c a m e p r o l i f e r a t i o n : B r i t a i n t e s t e d i t s f i r s t d e v i c e i n O c t o b e r 1 9 5 2 ,
F r a n c e i n F e b r u a r y 1 9 6 0 , a n d C h i n a i n O c t o b e r 1 9 6 4 . T h e n u c l e a r a r m s r a c e c o n t i n u e d n o t j u s t i n q u a n t i t y a n d
g e o g r a p h i c a l s p r e a d , b u t a l s o t h r o u g h t h e i n v e n t i o n o f q u a l i t a t i v e l y n e w s y s t e m s , l i k e s u b m a r i n e l a u n c h e d
m i s s i l e s , M I R V s ( m u l t i p l e i n d e p e n d e n t l y t a r g e t a b l e r e e n t r y v e h i c l e s ) , b a t t l e f i e l d n u c l e a r w e a p o n s , a n d m i s s i l e
d e f e n s e s y s t e m s ( a n t i - b a l l i s t i c m i s s i l e s a n d “ S t a r W a r s ” ) .
1 2
M a x R o s e r a n d M o h a m e d N a g d y , “ N u c l e a r W e a p o n s ” ,
O u r W o r l d I n D a t a . o r g
( 2 0 1 3 ) . R e t r i e v e d f r o m :
h t t p s : / / o u r w o r l d i n d a t a . o r g / n u c l e a r - w e a p o n s
. A c c e s s e d 2 6 A u g u s t 2 0 2 0 .
8
9
F i g u r e 1 . N u c l e a r w a r h e a d s i n t h e U . S . a n d U . S . S . R . , a t t h r e e t e m p o r a l z o o m l e v e l s s p a n n i n g 1 9 4 5 – 1 9 6 3 .
N o t e t h a t t h e y - a x i s s c a l e i n c r e a s e s b y o r d e r s o f m a g n i t u d e b e t w e e n t h e f i g u r e s . F r o m : M a x R o s e r a n d
M o h a m e d N a g d y , “ N u c l e a r W e a p o n s ” ,
O u r W o r l d I n D a t a . o r g
( 2 0 1 3 ) . A v a i l a b l e a t :
h t t p s : / / o u r w o r l d i n d a t a . o r g / n u c l e a r - w e a p o n s
. A c c e s s e d 2 6 A u g u s t 2 0 2 0 .
K e y D a t e s
F o r a d e t a i l e d c h r o n o l o g y , s e e
A p p e n d i x A
. F o r a b r i e f b i o g r a p h y o f s o m e o f t h e k e y h i s t o r i c a l f i g u r e s , s e e
A p p e n d i x B
.
1 0
D e c 1 9 4 1
P r e s i d e n t F r a n k l i n D . R o o s e v e l t a u t h o r i z e s M a n h a t t a n E n g i n e e r i n g D i s t r i c t
F e b 1 9 4 5
Y a l t a C o n f e r e n c e
A p r i l 1 9 4 5
F r a n k l i n D . R o o s e v e l t d i e s
M a y 1 9 4 5
G e r m a n y s u r r e n d e r s
J u n e 1 9 4 5
U n i t e d N a t i o n s C h a r t e r s i g n e d i n S a n F r a n c i s c o
J u l y 1 9 4 5
T r i n i t y T e s t ; P o t s d a m C o n f e r e n c e
A u g u s t 1 9 4 5
H i r o s h i m a a n d N a g a s a k i b o m b e d ; J a p a n s u r r e n d e r s
O c t o b e r 1 9 4 5
M i l i t a r y - b a c k e d M a y - J o h n s o n B i l l f o r t h e d o m e s t i c r e g u l a t i o n o f a t o m i c e n e r g y
i n t r o d u c e d i n C o n g r e s s
D e c 1 9 4 5
C o n f e r e n c e o f F o r e i g n M i n i s t e r s i n M o s c o w ; T h e M c M a h o n B i l l f o r t h e d o m e s t i c
r e g u l a t i o n o f a t o m i c e n e r g y i n t r o d u c e d i n C o n g r e s s
J a n 1 9 4 6
F i r s t s e s s i o n o f t h e U . N . G e n e r a l A s s e m b l y
F e b - M a r c h 1 9 4 6
S o v i e t s p y r i n g ; L o n g T e l e g r a m ; U . S . S . R . m i s s e d d e a d l i n e o n I r a n ; C h u r c h i l l ’ s “ I r o n
C u r t a i n ” s p e e c h
M a r c h 1 9 4 6
A c h e s o n - L i l i e n t h a l R e p o r t c o m p l e t e d , a n d l e a k e d
J u n e 1 9 4 6
B a r u c h P l a n p r e s e n t e d t o t h e U N A E C
D e c 1 9 4 6
B a r u c h P l a n v o t e , 1 0 i n f a v o r a n d 2 a b s t e n t i o n s ( U . S . S . R . , P o l a n d )
3 . P r o p o s a l s f o r I n t e r n a t i o n a l C o n t r o l : A B r i e f S u m m a r y
T h e r e w e r e m a n y p r o p o s a l s i n t h e U . S . f o r i n t e r n a t i o n a l c o n t r o l b e t w e e n m i d 1 9 4 5 a n d m i d 1 9 4 6 , a r t i c u l a t e d i n
v a r y i n g d e g r e e s o f d e t a i l . T h e m o s t i n f l u e n t i a l a r e n o t e d b e l o w :
N i e l s B o h r ’ s P r o p o s a l s , J u l y 1 9 4 4
S e n i o r D a n i s h p h y s i c i s t N i e l s B o h r m e t B r i t i s h P r i m e M i n i s t e r W i n s t o n C h u r c h i l l i n M a y 1 9 4 4 a n d U . S .
P r e s i d e n t F r a n k l i n D . R o o s e v e l t i n A u g u s t 1 9 4 4 , s u g g e s t i n g t o b o t h t h a t t h e U . S . i n f o r m t h e S o v i e t U n i o n o f i t s
b o m b p r o j e c t a n d b e g i n n e g o t i a t i o n s o n t h e i n t e r n a t i o n a l c o n t r o l o f a t o m i c e n e r g y . H e l a i d o u t h i s i d e a s i n t h e
m o s t d e t a i l i n a m e m o r a n d u m p r e p a r e d f o r R o o s e v e l t , d a t e d J u l y 1 9 4 4 . I t w a r n e d o f a f u t u r e a r m s r a c e a n d
s u g g e s t e d t h a t i n t e r n a t i o n a l c o n t r o l w a s t h e o n l y s o l u t i o n . B o h r e m p h a s i z e d t h e n e e d f o r c o m p l e t e e x c h a n g e o f
i n f o r m a t i o n , s o m e “ g u a r a n t e e o f c o m m o n s e c u r i t y , ” a n d i n t e r n a t i o n a l c o o p e r a t i o n a m o n g s t s c i e n t i s t s . H e d i d
n o t g i v e a n y s p e c i f i c s a s t o h o w i n t e r n a t i o n a l c o n t r o l w o u l d o p e r a t e , e . g . , o n s t a g e s , v e r i f i c a t i o n , d i s p o s a l o f
c u r r e n t f a c i l i t i e s a n d w e a p o n s , a n d r a w m a t e r i a l s .
1 3
T h e B u s h - C o n a n t M e m o , S e p t e m b e r 1 9 4 4
S e n i o r s c i e n c e p o l i c y m a k e r s V a n n e v a r B u s h ( h e a d o f t h e U . S . O f f i c e o f S c i e n t i f i c R e s e a r c h a n d D e v e l o p m e n t )
a n d J a m e s B . C o n a n t p r e p a r e d a p o l i c y s u g g e s t i o n o n a t o m i c e n e r g y f o r S e c r e t a r y H e n r y L . S t i m s o n i n
S e p t e m b e r 1 9 4 4 . T h e y a d v i s e d t h a t o t h e r c o u n t r i e s c o u l d c a t c h u p w i t h t h e U . S . w i t h i n f o u r y e a r s . T h e y
s u g g e s t e d s h a r i n g s c i e n t i f i c i n f o r m a t i o n w i t h a l l c o u n t r i e s ; o n l y m a n u f a c t u r i n g a n d m i l i t a r y d e t a i l s w e r e t o
r e m a i n s e c r e t . E x c e s s i v e s e c r e c y c o u l d l e a d t o a n a r m s r a c e w i t h t h e S o v i e t U n i o n . I t w a s n o t p o s s i b l e f o r t h e
U . S . t o m o n o p o l i z e r a w m a t e r i a l s g o i n g f o r w a r d . T h e i r o n l y c o n c r e t e s u g g e s t i o n f o r i n t e r n a t i o n a l c o n t r o l w a s
f r e e f l o w o f s c i e n t i f i c i n f o r m a t i o n a n d i n t e r n a t i o n a l i n s p e c t i o n s t h r o u g h a n i n t e r n a t i o n a l o r g a n i z a t i o n .
1 4
T h e B u s h P l a n , N o v e m b e r 1 9 4 5
V a n n e v a r B u s h p r e s e n t e d h i s p r o p o s a l f o r i n t e r n a t i o n a l c o n t r o l i n a m e m o r a n d u m t o S e c r e t a r y o f S t a t e B y r n e s
i n N o v e m b e r 1 9 4 5 . T h e p r o p o s a l w a s s i g n i f i c a n t l y m o r e d e t a i l e d t h a n e a r l i e r o n e s a n d e m p h a s i z e d , f o r t h e f i r s t
t i m e , a s t a g e d p r o c e s s , a t t h e e n d o f w h i c h t h e U . S . w o u l d g i v e u p a t o m i c w e a p o n s . T h e s t a g e s w e r e , f i r s t , t h e
f o r m a t i o n o f a U . N . a g e n c y f o r t h e d i s s e m i n a t i o n o f s c i e n t i f i c i n f o r m a t i o n , i n c l u d i n g f r e e a c c e s s f o r s c i e n t i s t s t o
1 3
N i e l s B o h r ,
N i e l s B o h r ' s M e m o r a n d u m t o P r e s i d e n t R o o s e v e l t
( J u l y 1 9 4 4 ) . A v a i l a b l e a t :
h t t p : / / w w w . a t o m i c a r c h i v e . c o m / D o c s / M a n h a t t a n P r o j e c t / B o h r m e m o . s h t m l
. A c c e s s e d 2 2 S e p t e m b e r 2 0 1 8 . O n B o h r ’ s
a c t i v i s m , s e e M a r t i n J . S h e r w i n , “ N i e l s B o h r a n d t h e F i r s t P r i n c i p l e s o f A r m s C o n t r o l ” , i n H e r m a n F e s h b a c h , T e t s u o
M a t s u i , a n d A l e x a n d r a O l e s o n ( e d s . ) ,
N i e l s B o h r : P h y s i c s a n d t h e W o r l d
( A b i n g d o n : R o u t l e d g e , 1 9 9 8 ) , p p . 3 1 9 - 3 0 ; F i n n
A a s e r u d , “ T h e S c i e n t i s t a n d t h e S t a t e s m e n : N i e l s B o h r ' s P o l i t i c a l C r u s a d e d u r i n g W o r l d W a r I I ” ,
H i s t o r i c a l S t u d i e s i n t h e
P h y s i c a l a n d B i o l o g i c a l S c i e n c e s
3 0 , 1 ( 1 9 9 9 ) , p p . 1 - 4 7 .
1 4
V a n n e v a r B u s h a n d J a m e s B . C o n a n t , “ M e m o r a n d u m ” ( 3 0 S e p t e m b e r 1 9 4 4 ) . A v a i l a b l e a t :
h t t p s : / / n s a r c h i v e 2 . g w u . e d u / / N S A E B B / N S A E B B 1 6 2 / 1 . p d f
. A c c e s s e d 2 2 S e p t e m b e r 2 0 1 8 . A l s o : G . P a s c a l Z a c h a r y ,
E n d l e s s F r o n t i e r : V a n n e v a r B u s h , E n g i n e e r o f t h e A m e r i c a n C e n t u r y
( N e w Y o r k : S i m o n a n d S c h u s t e r , 2 0 1 8 )
,
p . 2 4 3 ; J a m e s
H e r s h b e r g , J a m e s B . C o n a n t :
H a r v a r d t o H i r o s h i m a a n d t h e M a k i n g o f t h e N u c l e a r A g e
( N e w Y o r k : A l f r e d A . K n o p f ,
1 9 9 3 ) , p p . 2 0 4 - 5 .
1 1
b a s i c r e s e a r c h . S e c o n d c a m e t h e e s t a b l i s h m e n t o f a U . N . i n s p e c t i o n s y s t e m , w i t h a g r a d u a l e x c h a n g e o f
i n f o r m a t i o n o n r a w m a t e r i a l s a n d f a c i l i t i e s . T h i s w a s t o c u l m i n a t e w i t h t h e s h a r i n g o f t h e m o s t p r a c t i c a l a n d
m o s t s e c r e t a t o m i c k n o w - h o w . N a t i o n s w e r e t o a g r e e t o u s e s u c h i n f o r m a t i o n f o r c o m m e r c i a l p u r p o s e s o n l y .
F i n a l l y , t h e U . S . w a s t o c o n v e r t i t s b o m b s t o p e a c e f u l u s e s .
T h i s p l a n w a s a d o p t e d b y B y r n e s ( i n a s o m e w h a t
1 5
m o d i f i e d f o r m ) i n D e c e m b e r . B y r n e s h a d f o r m e r l y t a k e n a m o r e h a w k i s h p o s i t i o n o n t h e b o m b a n d t h e S o v i e t
U n i o n b u t a c c e p t e d t h i s c o o p e r a t i v e a p p r o a c h a f t e r h e r e a l i z e d t h a t t h e U . S . ’ s a t o m i c b o m b w a s n o t h e l p i n g i n
g e o p o l i t i c a l n e g o t i a t i o n s w i t h t h e S o v i e t s .
1 6
T h e C o h e n - P a s v o l s k y P l a n , D e c e m b e r 1 9 4 5
T h i s w a s t h e o f f i c i a l S t a t e D e p a r t m e n t p l a n , d r a w n u p b y a c o m m i t t e e h e a d e d b y S t a t e D e p a r t m e n t o f f i c i a l s
B e n j a m i n C o h e n a n d L e o P a s v o l s k y . T h e p l a n w a s a c c e p t e d b y B y r n e s a n d p r e s e n t e d t o t h e S o v i e t s i n t h e
D e c e m b e r 1 9 4 5 M o s c o w c o n f e r e n c e . T h e C o h e n - P a s v o l s k y P l a n w a s b a s e d o n t h e N o v e m b e r 1 9 4 5 B u s h P l a n ,
b u t i n c l u d e d o n e c r u c i a l c h a n g e ; a l t h o u g h i t e m p h a s i z e d s t a g e s , i t a d d e d t h a t t h e i n t e r n a t i o n a l c o n t r o l p r o c e s s
c o u l d m o v e o n t o t h e n e x t s t a g e w i t h o u t h a v i n g c o m p l e t e d t h e p r e v i o u s s t a g e . T h e s t a g e s t h u s d i d n o t h a v e t o
p r o g r e s s i n a s t r i c t s e q u e n c e .
1 7
T h e A c h e s o n - L i l i e n t h a l P l a n , M a r c h 1 9 4 6
T h e A c h e s o n - L i l i e n t h a l R e p o r t s e t o u t a p l a n p r o d u c e d b y a g r o u p o f e x p e r t c o n s u l t a n t s ( i n c l u d i n g t h e l e a d i n g
M a n h a t t a n P r o j e c t p h y s i c i s t J . R o b e r t O p p e n h e i m e r ) f o r t h e S t a t e D e p a r t m e n t i n M a r c h 1 9 4 6 . T h i s w a s t h e
s i n g l e m o s t d e t a i l e d p r o p o s a l f o r i n t e r n a t i o n a l c o n t r o l p r o d u c e d i n t h e U . S . a n d w a s t h e b a s i s ( w i t h c r u c i a l
c h a n g e s ) f o r t h e o f f i c i a l U . S . p r o p o s a l — t h e s o - c a l l e d B a r u c h P l a n — a t t h e U n i t e d N a t i o n s A t o m i c E n e r g y
C o m m i s s i o n ( U N A E C ) a f e w m o n t h s l a t e r .
1 8
T h e A c h e s o n - L i l i e n t h a l P l a n w a s p r e m i s e d o n t h e a s s u m p t i o n t h a t i n s p e c t i o n s w e r e i n s u f f i c i e n t f o r
i n t e r n a t i o n a l c o n t r o l . I n s t e a d , t h e U . N . , t h r o u g h a n A t o m i c D e v e l o p m e n t A u t h o r i t y ( A D A ) , w a s t o c o n t r o l a l l
f i s s i o n a b l e r a w m a t e r i a l s a n d h a v e a m o n o p o l y o n a l l “ d a n g e r o u s ” a c t i v i t i e s ( i . e . , t h o s e w i t h m i l i t a r y
a p p l i c a t i o n s ) . S t a t e s w o u l d s h u t d o w n a l l d a n g e r o u s a c t i v i t i e s , a n d a l l a t o m i c m a t e r i a l w o u l d b e t r a n s f e r r e d t o
U . N . o w n e r s h i p . P e a c e f u l d e v e l o p m e n t ( R & D , p o w e r p l a n t s ) , h o w e v e r , c o u l d c o n t i n u e i n s t a t e s . T h e U n i t e d
S t a t e s w o u l d b e g i n a p h a s e d t r a n s i t i o n o f i t s b o m b s , m a t e r i a l , a n d f a c i l i t i e s t o t h e A D A , o n c e s e t u p . T h e U . S .
1 5
T h e B u s h m e m o c a n b e f o u n d a t : U n i t e d S t a t e s D e p a r t m e n t o f S t a t e ,
F o r e i g n R e l a t i o n s o f t h e U n i t e d S t a t e s : D i p l o m a t i c
P a p e r s , 1 9 4 5 . G e n e r a l : P o l i t i c a l a n d E c o n o m i c M a t t e r s
V o l u m e I I ( W a s h i n g t o n , D . C . , 1 9 4 6 ) , p p . 6 9 - 7 3 . A v a i l a b l e a t :
h t t p : / / d i g i c o l l . l i b r a r y . w i s c . e d u / c g i - b i n / F R U S / F R U S - i d x ? t y p e = a r t i c l e & d i d = F R U S . F R U S 1 9 4 5 v 0 2 . i 0 0 0 5 & i d = F R U S . F R
U S 1 9 4 5 v 0 2 & i s i z e = M
. A c c e s s e d 2 4 A p r i l 2 0 1 9 . F o r c o n t e x t , s e e B a r t o n J . B e r n s t e i n , “ T h e Q u e s t f o r S e c u r i t y : A m e r i c a n
F o r e i g n P o l i c y a n d I n t e r n a t i o n a l C o n t r o l o f A t o m i c E n e r g y , 1 9 4 2 - 1 9 4 6 ” ,
T h e J o u r n a l o f A m e r i c a n H i s t o r y
6 0 , 4 ( M a r c h
1 9 7 4 ) , p p . 1 0 0 3 - 1 0 4 4 ; Z a c h a r y ,
E n d l e s s F r o n t i e r
, p p . 3 0 9 - 3 1 0 .
1 6
H e r k e n ,
p p . 6 1 , 7 1 , 7 2 .
1 7
H e r k e n , p p . 7 1 - 2 . M e m o r a n d u m , “ D r a f t P r o p o s a l s o n A t o m i c E n e r g y f o r S u b m i s s i o n t o S o v i e t G o v e r n m e n t ”
( 1 0
D e c e m b e r 1 9 4 5 ) , i n U n i t e d S t a t e s D e p a r t m e n t o f S t a t e ,
F o r e i g n R e l a t i o n s o f t h e U n i t e d S t a t e s : D i p l o m a t i c P a p e r s , 1 9 4 5 .
G e n e r a l : P o l i t i c a l a n d E c o n o m i c M a t t e r s
V o l u m e I I ( W a s h i n g t o n , D C , 1 9 4 6 ) . A v a i l a b l e a t :
h t t p s : / / h i s t o r y . s t a t e . g o v / h i s t o r i c a l d o c u m e n t s / f r u s 1 9 4 5 v 0 2 / d 3 8
. A c c e s s e d 2 2 A p r i l 2 0 1 9 .
1 8
C h e s t e r I . B a r n a r d , J . R . O p p e n h e i m e r , C h a r l e s A . T h o m a s e t a l . ,
A R e p o r t o n t h e I n t e r n a t i o n a l C o n t r o l o f A t o m i c E n e r g y
( W a s h i n g t o n , D . C . : T h e S t a t e D e p a r t m e n t , 1 9 4 6 ) .
1 2
w o u l d
n o t
c e a s e a t o m i c o p e r a t i o n s p r i o r t o t h e s e t t i n g u p o f t h e A D A . T h e p l a n p l a c e d s i g n i f i c a n t e m p h a s i s o n
t h e c o o p e r a t i o n o f i n t e r n a t i o n a l i s t s c i e n t i s t s w o r k i n g a t t h e A D A .
1 9
T h e A D A w a s t h e c e n t e r p i e c e o f t h e A c h e s o n - L i l i e n t h a l P l a n . I t w a s t o s e t u p l a r g e R & D c e n t e r s a n d c o n d u c t
r e s e a r c h o n p e a c e f u l a n d w a r l i k e u s e s o f a t o m i c e n e r g y . I t w o u l d a l s o h a v e i t s o w n o p e r a t i o n a l r e a c t o r s . T h e s e
r e a c t o r s a n d o t h e r a t o m i c f a c i l i t i e s w e r e t o b e s p r e a d a c r o s s a n u m b e r o f ( u n s p e c i f i e d ) c o u n t r i e s i n a “ s t r a t e g i c
b a l a n c e a m o n g n a t i o n s ” s o t h a t i n t h e e v e n t o f a b r e a k d o w n o f t h e A D A ( o r t h e U . N . i t s e l f ) t h e r e w o u l d b e a
“ b a l a n c e o f f a c i l i t i e s ” a c r o s s s t a t e s . T h i s , i t w a s h o p e d , w o u l d r e d u c e t h e f e a r s o f a n y o n e s t a t e t h a t j o i n i n g
w o u l d u n d e r m i n e i t s s e c u r i t y i n t h e e v e n t o f a d i p l o m a t i c b r e a k d o w n . T h e A D A w o u l d o w n a n d o p e r a t e a l l
m i n i n g , r e f i n i n g , a n d p r o d u c t i o n o f f i s s i o n a b l e r a w m a t e r i a l s . E x i s t i n g m i n e s , p l a n t s , a n d f a c t o r i e s ( e . g . , a t
H a n f o r d a n d O a k R i d g e ) w e r e t o b e t r a n s f e r r e d t o t h e c o n t r o l o f t h e A D A . I t w o u l d d i s p e n s e “ d e n a t u r e d ”
2 0
f i s s i o n a b l e r a w m a t e r i a l s t o i n d i v i d u a l n a t i o n s f o r t h e i r n u c l e a r p o w e r p l a n t s a n d l i c e n s e a n d i n s p e c t t h e i r
( c i v i l i a n ) n u c l e a r f a c i l i t i e s .
2 1
T h e B a r u c h P l a n , J u n e 1 9 4 6
T h e B a r u c h P l a n w a s d e v e l o p e d b y B e r n a r d B a r u c h — a b u s i n e s s m a n a n d f i n a n c i e r w h o w a s t h e U . S .
r e p r e s e n t a t i v e o n t h e U N A E C — b e t w e e n M a r c h a n d J u n e 1 9 4 6 . T h e p l a n w a s a d o p t e d a s o f f i c i a l U . S . p o l i c y
a n d p r e s e n t e d a t t h e U N A E C i n J u n e 1 9 4 6 .
I t w a s b a s e d o n t h e A c h e s o n - L i l i e n t h a l R e p o r t b u t i n c l u d e d s o m e
2 2
c r u c i a l c h a n g e s t h a t m a d e t h e p l a n m o r e h a w k i s h a n d p r o - b u s i n e s s . T h e r e w e r e l i k e l y s e v e r a l r e a s o n s f o r t h e s e
c h a n g e s . O n e m i g h t h a v e b e e n t o h e l p B a r u c h a p p e a r a s t h e a u t h o r o f t h e p r o p o s a l , r a t h e r t h a n j u s t a
“ m e s s e n g e r b o y ” f o r t h e A c h e s o n - L i l i e n t h a l R e p o r t .
A n o t h e r m i g h t h a v e b e e n t o r e d u c e t h e r i s k t o t h e U . S . i f
2 3
t h e p l a n f a i l e d o r t h e S o v i e t U n i o n r e n e g e d . A t h i r d m a y h a v e b e e n t o r e t a i n p r i v a t e s e c t o r a u t o n o m y i n t h e
n u c l e a r i n d u s t r y .
2 4
T h e c e n t r a l e l e m e n t s o f t h e B a r u c h P l a n w e r e t h a t i t a b o l i s h e d t h e v e t o p o w e r o f t h e S e c u r i t y C o u n c i l i n
r e l a t i o n t o a t o m i c m a t t e r s . I t e m p h a s i z e d “ i m m e d i a t e , s w i f t , a n d s u r e p u n i s h m e n t , ” i n c l u d i n g t h e p o s s i b i l i t y o f
a t o m i c a t t a c k , o n v i o l a t o r s o f t h e p l a n . T h e p l a n i n s i s t e d o n a s u r v e y o f S o v i e t r e s o u r c e s a s a f i r s t s t e p . T h i s
w o u l d h a v e p u t t h e S o v i e t s a t a g r e a t d i s a d v a n t a g e , a s t h e y w o u l d r e v e a l s e c r e t i n f o r m a t i o n w i t h o u t t h e U . S .
r e c i p r o c a t i n g a t t h a t p o i n t . T h i s c a n b e t h o u g h t o f a s a “ h i d d e n U . S . v e t o ” b u i l t i n t o t h i s i n t e r n a t i o n a l c o n t r o l
p r o c e s s , b e c a u s e t h e U . S . w o u l d g i v e u p l i t t l e i n t h e i n i t i a l s t a g e s a n d s o c o u l d a b o r t t h e p r o c e s s p a r t w a y
t h r o u g h w i t h m i n i m a l d o w n s i d e . T h e p l a n d e - e m p h a s i z e d t h e r o l e o f t h e A t o m i c D e v e l o p m e n t A u t h o r i t y
1 9
I b i d . , p p . 4 1 - 5 3 .
2 0
B y “ d e n a t u r e d ” t h e p l a n m e a n t f i s s i o n a b l e r a w m a t e r i a l s w h i c h c o u l d n o t b e u s e d f o r “ d a n g e r o u s ” p u r p o s e s . I n r e a l i t y ,
t h i s i s n o t t e c h n i c a l l y p o s s i b l e , b u t i t w a s b e l i e v e d t o b e a t t h e t i m e .
2 1
I b i d . , p p . 4 1 - 5 3 .
2 2
“ T h e A m e r i c a n P r o p o s a l f o r I n t e r n a t i o n a l C o n t r o l P r e s e n t e d b y B e r n a r d B a r u c h ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
1 & 2
( 1 J u l y 1 9 4 6 ) , p p . 3 - 5 , 1 0 . A l s o a t :
h t t p : / / w w w . a t o m i c a r c h i v e . c o m / D o c s / D e t e r r e n c e / B a r u c h P l a n . s h t m l
. A c c e s s e d 2 2
A p r i l 2 0 1 9 .
2 3
H e r k e n , p . 1 6 0 ; C r a i g a n d R a d c h e n k o ,
T h e A t o m i c B o m b a n d t h e O r i g i n s o f t h e C o l d W a r ,
p p . 1 2 2 - 2 3 .
2 4
O n B a r u c h ’ s m o t i v a t i o n s , s e e H e r k e n , p p . 1 6 3 - 1 6 4 . O n t h e h a w k i s h n a t u r e o f t h e p l a n , s e e H e r k e n , p p . 1 6 9 , 1 7 1 . O n e
a l t e r n a t i v e v i e w i s t h a t B a r u c h e x p l i c i t l y d e s i g n e d h i s p l a n t o b e a n “ o b v i o u s p r o p a g a n d a p l o y , ” s e e S h a n e J . M a d d o c k ,
N u c l e a r A p a r t h e i d : T h e Q u e s t f o r A m e r i c a n A t o m i c S u p r e m a c y f r o m W o r l d W a r I I t o t h e P r e s e n t
( C h a p e l H i l l , N C : t h e
U n i v e r s i t y o f N o r t h C a r o l i n a P r e s s , 2 0 1 0 ) , p . 5 7 .
1 3
( A D A ) a n d i n s t e a d s h i f t e d r e s p o n s i b i l i t y f o r m i n i n g a n d r e f i n i n g o f f i s s i o n a b l e m a t e r i a l s t o p r i v a t e i n d u s t r y . I n
t h e A c h e s o n - L i l i e n t h a l P l a n , a l l m i n i n g a n d r e f i n i n g w a s t o b e c a r r i e d o u t b y t h e A D A . I n B a r u c h ’ s p l a n , t h e
A D A w o u l d o n l y o w n / m a n a g e “ a c t i v i t i e s p o t e n t i a l l y d a n g e r o u s t o w o r l d s e c u r i t y ” ; t h e r e s t w o u l d o n l y b e
i n s p e c t e d o r l i c e n s e d b y t h e A D A .
2 5
T h e G r o m y k o P l a n , J u n e 1 9 4 6
T h i s w a s t h e o f f i c i a l S o v i e t c o u n t e r p r o p o s a l t o t h e U . S . B a r u c h P l a n . I t w a s a n n o u n c e d a t t h e U N A E C b y
S o v i e t d e l e g a t e A n d r e i G r o m y k o o n 1 9 J u n e 1 9 4 6 a s a n i m p l i c i t r e j e c t i o n o f t h e B a r u c h P l a n .
T h e p l a n
2 6
f o c u s e d o n d i s a r m a m e n t r a t h e r t h a n c o n t r o l l i n g t h e r a w m a t e r i a l s o r t h e s c i e n t i f i c R & D b e h i n d a t o m i c
w e a p o n s . I t c a l l e d f o r a c o m p l e t e b a n o n a t o m i c w e a p o n s , w h i c h w e r e t o b e d e s t r o y e d w i t h i n t h r e e m o n t h s o f
t h e t r e a t y c o m i n g i n t o f o r c e , a n d c o n t r a c t i n g p a r t i e s w e r e t o a g r e e n o t t o m a k e o r u s e a t o m i c w e a p o n s .
V i o l a t i o n s w o u l d b e r e g a r d e d a s a “ c r i m e a g a i n s t h u m a n i t y , ” a n d
p e n a l t i e s w o u l d b e d e t e r m i n e d b y d o m e s t i c
l e g i s l a t i o n
. T h e p l a n i n s i s t e d t h a t t h e S e c u r i t y C o u n c i l v e t o a p p l y t o i n t e r n a t i o n a l c o n t r o l a n d a l l a t o m i c
m a t t e r s ( c o n t r a t h e B a r u c h P l a n ) . I t s u g g e s t e d t h e f o r m a t i o n o f t w o U n i t e d N a t i o n s c o m m i t t e e s o v e r s e e n b y t h e
S e c u r i t y C o u n c i l : t h e f i r s t w a s t o o r g a n i z e t h e e x c h a n g e o f a t o m i c i n f o r m a t i o n a n d t h e s e c o n d t o e n s u r e t h a t t h e
i n t e r n a t i o n a l a g r e e m e n t i s f o l l o w e d . T h e r e q u i r e m e n t o f e a r l y U . S . d i s a r m a m e n t m a d e t h i s p r o p o s a l c o m p l e t e l y
u n a c c e p t a b l e t o t h e U n i t e d S t a t e s . H o w e v e r , i t w a s t h e v e t o t h a t w a s a f o c u s o f B a r u c h ’ s o p p o s i t i o n , e v e n
t h o u g h i t w a s p r o b a b l y s t r a t e g i c a l l y w o r t h l e s s s i n c e — i n t h e e v e n t o f b r e a k d o w n — t h e o n l y r e a l s a n c t i o n w o u l d
b e a t h r e a t o f w a r .
H i s t o r i a n s b e l i e v e t h a t S o v i e t U n i o n i t s e l f d i d n o t e x p e c t t h i s p r o p o s a l t o b e a c c e p t e d a n d
2 7
o n l y p u t i t f o r w a r d f o r p r o p a g a n d a r e a s o n s , a n d p e r h a p s a l s o t o l e a r n m o r e a b o u t t h e U . S . a t o m i c w e a p o n s
p r o g r a m .
2 8
2 5
“ T h e A m e r i c a n P r o p o s a l f o r I n t e r n a t i o n a l C o n t r o l P r e s e n t e d b y B e r n a r d B a r u c h ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
1 & 2
( 1 J u l y 1 9 4 6 ) p p . 3 - 5 , 1 0 . A l s o a t :
h t t p : / / w w w . a t o m i c a r c h i v e . c o m / D o c s / D e t e r r e n c e / B a r u c h P l a n . s h t m l
. A c c e s s e d 2 5 A p r i l
2 0 1 9 .
2 6
“ T h e R u s s i a n P r o p o s a l f o r I n t e r n a t i o n a l C o n t r o l ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
2 , n o . 1 , 2 ( 1 J u l y 1 9 4 6 ) , p p . 8 - 1 0 . A l s o
a v a i l a b l e a t : U . S . D e p a r t m e n t o f S t a t e ,
D o c u m e n t s o n D i s a r m a m e n t
v o l u m e 1 1 9 4 5 - 1 9 5 6 ( W a s h i n g t o n , D C : D e p a r t m e n t
o f S t a t e , 1 9 6 0 ) , p p . 1 7 - 2 4 .
2 7
H e r k e n , p p . 1 7 4 - 7 5 .
2 8
C r a i g a n d R a d c h e n k o ,
T h e A t o m i c B o m b a n d t h e O r i g i n s o f t h e C o l d W a r ,
p p . 1 3 5 - 1 4 0 .
1 4
4 . L e s s o n s
T h i s s e c t i o n o u t l i n e s s o m e o f t h e l e s s o n s w e b e l i e v e c a n b e g l e a n e d f r o m t h e h i s t o r y o f e a r l y a t t e m p t s a t
i n t e r n a t i o n a l c o n t r o l o f n u c l e a r w e a p o n s . T h e s e l e s s o n s s t r e s s t h e c o m p l e x i t y a n d m e s s i n e s s w h i c h h a s l o n g b e e n
r e c o g n i z e d a s a n i n h e r e n t p a r t o f p o l i c y m a k i n g ; n e v e r t h e l e s s , w e f e e l t h a t p i n p o i n t i n g t h e s e l e s s o n s s p e c i f i c a l l y
i n r e l a t i o n t o p o w e r f u l t e c h n o l o g i e s m a k e s t h e m m o r e s a l i e n t f o r t h o s e t h i n k i n g a b o u t s u c h t e c h n o l o g i e s t o d a y .
T h e l e s s o n s b e g i n w i t h s o m e a s p e c t s o f t h e p r o p o s a l s b e f o r e m o v i n g t o c o n s i d e r s o m e o f t h e c o n s t i t u e n c i e s a n d
p r o c e s s e s t h a t g e n e r a t e d t h e m . W e t h e n c o n s i d e r t h e i r l i k e l i h o o d o f s u c c e s s a n d e n d w i t h s o m e c o n s i d e r a t i o n o f
c o o p e r a t i o n a n d u n i l a t e r a l a c t i o n . T h e s e l e s s o n s d o n o t c o n s t i t u t e a c o m p r e h e n s i v e o r h o l i s t i c o v e r v i e w o f t h e s e
p r o p o s a l s : f o r t h a t w e w o u l d p o i n t r e a d e r s t o t h e v a r i o u s h i s t o r i c a l s t u d i e s o n t h i s t o p i c ( s e e
R e f e r e n c e s
f o r a l i s t
a n d
I n t r o d u c t i o n
f o r a g u i d e t o t h e l i t e r a t u r e ) .
4 . 1 S e r i o u s R a d i c a l P r o p o s a l s
L e s s o n s
R a d i c a l p r o p o s a l s w h i c h w o u l d n o r m a l l y a p p e a r n a i v e o r e x t r e m e m a y , i n t h e r i g h t c i r c u m s t a n c e s , b e s e r i o u s l y
p r o p o s e d , d i s c u s s e d , a n d e v e n a d o p t e d a s o f f i c i a l p o l i c y . T w o c o n d i t i o n s a r e c o n d u c i v e t o t h i s . F i r s t , i f t h e
e m e r g e n t t e c h n o l o g y i s s p e c t a c u l a r l y d i s r u p t i v e , i t c a n e x p a n d t h e r e a l m o f p o l i t i c a l l y f e a s i b l e p o l i c i e s . S e c o n d , a
s e n s e o f r u p t u r e o r c r i s i s i n i n t e r n a t i o n a l p o l i t i c a l a f f a i r s c a n m a k e o t h e r w i s e u n r e a l i s t i c p r o p o s a l s m o r e
a c c e p t a b l e a n d p o s s i b l e .
H i s t o r i c a l C a s e
P r o p o s a l s f o r i n t e r n a t i o n a l c o n t r o l o f a t o m i c e n e r g y w e r e r a d i c a l f o r t h e i r t i m e . T h e y p r o p o s e d t h a t s t a t e s b e
b o u n d b y p o w e r f u l a n d w i d e - r a n g i n g m u l t i l a t e r a l a g r e e m e n t s a n d t h a t p o w e r f u l i n t e r n a t i o n a l o r g a n i z a t i o n s b e
c r e a t e d w i t h t h e p o w e r t o p o l i c e s u c h a g r e e m e n t s . I n b o t h t h e s e s e n s e s , t h e s e p r o p o s a l s w e r e m u c h m o r e r a d i c a l
t h a n o t h e r s e r i o u s ( t h a t i s , t a k e n u p a t t h e d i p l o m a t i c l e v e l ) d i s c u s s i o n s o n i n t e r n a t i o n a l g o v e r n a n c e a t t h e t i m e .
T h e U n i t e d N a t i o n s c h a r t e r , f o r e x a m p l e , d i d n o t c r e a t e o b l i g a t i o n s o n s t a t e s a s b i n d i n g , o r i n t r u s i v e o n
n a t i o n a l s o i l , a s s o m e e n v i s a g e d i n t h e A c h e s o n - L i l i e n t h a l R e p o r t .
S i m i l a r l y , n o o t h e r a r m a m e n t s w e r e s u b j e c t
2 9
t o s u c h p r o p o s a l s i n t h e 1 9 4 0 s .
3 0
T h e A c h e s o n - L i l i e n t h a l P l a n , f o r e x a m p l e , p r o p o s e d t h a t a p o w e r f u l n e w U . N . A t o m i c D e v e l o p m e n t
A u t h o r i t y ( A D A ) w o u l d s e t u p l a r g e R & D c e n t e r s a n d c o n d u c t r e s e a r c h o n p e a c e f u l a n d w a r l i k e u s e s o f a t o m i c
e n e r g y . I t w o u l d o w n a n d o p e r a t e a l l m i n i n g , r e f i n i n g , a n d p r o d u c t i o n o f f i s s i o n a b l e r a w m a t e r i a l s — i n c l u d i n g
h a v i n g i t s o w n o p e r a t i o n a l r e a c t o r s — a n d d i s p e n s e f i s s i o n a b l e r a w m a t e r i a l s t o n a t i o n s f o r t h e i r n u c l e a r p o w e r
p l a n t s . I t w o u l d a l s o l i c e n s e a n d i n s p e c t o p e r a t i n g c i v i l i a n n u c l e a r f a c i l i t i e s i n n a t i o n s t a t e s .
3 1
2 9
O n i n t e r n a t i o n a l g o v e r n a n c e i n t h e 3 0 s a n d 4 0 s ( i n c l u d i n g t h e U . N . c h a r t e r ) , s e e M a r k M a z o w e r ,
G o v e r n i n g t h e W o r l d :
T h e H i s t o r y o f a n I d e a
( P e n g u i n : N e w Y o r k , 2 0 1 2 ) , c h a p t e r s 5 a n d 7 .
3 0
O n t h e u n i q u e n e s s o f i n t e r n a t i o n a l c o n t r o l , s e e P a t r i c k M . M o r g a n , “ E l e m e n t s o f a G e n e r a l T h e o r y o f A r m s C o n t r o l ” ,
R o b e r t E . W i l l i a m s a n d P a u l R . V i o t t i ( e d s . ) ,
A r m s C o n t r o l : H i s t o r y , T h e o r y , a n d P o l i c y
( S a n t a B a r b a r a , C A : P r a e g e r , 2 0 1 2 ) ,
p p . 1 5 - 4 0 .
3 1
C h e s t e r I . B a r n a r d , J . R . O p p e n h e i m e r , C h a r l e s A . T h o m a s e t a l . ,
A R e p o r t o n t h e I n t e r n a t i o n a l C o n t r o l o f A t o m i c E n e r g y
( W a s h i n g t o n , D C : T h e S t a t e D e p a r t m e n t , 1 9 4 6 ) .
1 5
T h e s e p r o p o s a l s w e r e t a k e n s e r i o u s l y b y m a n y o f t h e i r p r o p o n e n t s a n d m u c h o f t h e p u b l i c . T h e r e i s e v e r y
i n d i c a t i o n , f o r e x a m p l e , t h a t t h e f r a m e r s o f t h e A c h e s o n - L i l i e n t h a l P l a n g e n u i n e l y b e l i e v e d t h a t t h e i r p l a n c o u l d
w o r k a n d t h a t i t w a s t h e m o s t e f f e c t i v e a n d r e a l i s t i c w a y o f d e a l i n g w i t h t h e p r o b l e m s o f a t o m i c e n e r g y .
3 2
T h e r e a r e s e v e r a l c o n d i t i o n s t h a t a l l o w e d t h e A c h e s o n - L i l i e n t h a l R e p o r t t o g a i n t r a c t i o n . ( 1 ) T h e r e p o r t w a s
a c c e p t e d b e c a u s e o f a g r o w i n g p u b l i c a n d e l i t e p e r c e p t i o n o f t h e t h r e a t t h a t a t o m i c w e a p o n s p o s e d . I t w a s
w i d e l y b e l i e v e d i n l a t e 1 9 4 5 a n d 1 9 4 6 t h a t a t o m i c w e a p o n s w o u l d b e u s e d i n a n y f u t u r e m a j o r w a r , w i p i n g o u t
c i t i e s a n d k i l l i n g m i l l i o n s . T
h e B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
, f o r e x a m p l e , w a s s e t u p t o w a r n o f t h e
d e s t r u c t i v e e f f e c t s o f a t o m i c w e a p o n s . A t t h e s a m e t i m e , m a n y b e l i e v e d t h a t a t o m i c e n e r g y o f f e r e d c h e a p
e l e c t r i c i t y a n d n e w t y p e s o f m e d i c i n e s a n d a g r i c u l t u r a l p r o d u c t s .
P u b l i c s a n d e l i t e s w e r e a m a z e d t h a t a n a t o m i c
3 3
b o m b h a d b e e n d e v e l o p e d s o f a s t a n d o f t e n a s s u m e d t h a t t h i s r a p i d p a c e w o u l d c o n t i n u e i n t o t h e n e a r f u t u r e .
N o v e l s a n d f u t u r i s t i c m a g a z i n e a n d n e w s p a p e r a r t i c l e s c o n t r i b u t e d t o t h e s e b e l i e f s .
( 2 ) T h e r e c e n t e x p e r i e n c e
3 4
w i t h a n a w f u l w a r i n c r e a s e d p u b l i c a n d e l i t e r e c e p t i v e n e s s t o r a d i c a l p o l i t i c a l p r o p o s a l s . T h e r e w a s a s e n s e t h a t
r e g u l a r p o l i t i c s h a d f a i l e d , a n d m o r e r a d i c a l m e a s u r e s w e r e r e q u i r e d . ( 3 ) T h e n e e d f o r p o s t w a r r e c o n s t r u c t i o n
a n d t h e g r o w t h o f U . S . i n f l u e n c e g l o b a l l y m a d e p o s s i b l e n e w i n i t i a t i v e s i n i n t e r n a t i o n a l r e l a t i o n s w h i c h w e r e
n o t p o s s i b l e b e f o r e . T h e f o r m a t i o n o f t h e U n i t e d N a t i o n s a n d o t h e r i n t e r n a t i o n a l o r g a n i z a t i o n s ( e . g . , t h e
B r e t t o n W o o d s s y s t e m t o m a n a g e t h e g l o b a l e c o n o m y ) w e r e w i d e l y w e l c o m e d .
( 4 ) S c i e n t i s t s w h o w o r k e d o n
3 5
a t o m i c m a t t e r s g e n e r a l l y t h r e w t h e i r w e i g h t b e h i n d t h e A c h e s o n - L i l i e n t h a l R e p o r t a n d f o r m e d p o w e r f u l
o r g a n i z a t i o n s t h a t a d v o c a t e d f o r i n t e r n a t i o n a l c o n t r o l . T h e i r s t a t u s a n d n e w f o u n d p r o m i n e n c e g a v e t h e i r
m e s s a g e t r a c t i o n i n t h e m e d i a a n d i n g o v e r n m e n t .
3 6
3 2
B e r n s t e i n , “ T h e Q u e s t f o r S e c u r i t y ” ; H e r k e n , p p . 1 5 5 - 5 8 .
3 3
O n e a r l y f e a r s a n d h o p e s a b o u t a t o m i c e n e r g y i n t h e U . S . , s e e P a u l B o y e r ,
B y t h e B o m b ’ s E a r l y L i g h t : A m e r i c a n T h o u g h t
a n d C u l t u r e a t t h e D a w n o f t h e A t o m i c A g e
( N e w Y o r k : P a n t h e o n B o o k s , 1 9 8 5 ) , P a r t s 4 a n d 5 ; P a u l B o y e r , “ A H i s t o r i c a l
V i e w o f S c a r e T a c t i c s ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
4 2 , 1 ( J a n u a r y 1 9 8 6 ) , p p . 1 7 - 1 9 .
3 4
B o y e r ,
B y t h e B o m b ’ s E a r l y L i g h t
, P a r t s 4 a n d 5 .
3 5
O n U . S . i n t e r n a t i o n a l i s m d u r i n g 1 9 3 9 - 1 9 4 5 , s e e R o b e r t A . D i v i n e ,
S e c o n d C h a n c e : T h e T r i u m p h o f I n t e r n a t i o n a l i s m i n
A m e r i c a D u r i n g W o r l d W a r I I
( N e w Y o r k : A t h e n e u m , 1 9 7 1 ) .
3 6
O n t h e i r a c t i v i s m , s e e A l i c e K i m b a l l - S m i t h ,
A P e r i l a n d a H o p e : T h e S c i e n t i s t s ’ M o v e m e n t i n A m e r i c a : 1 9 4 5 - 4 7
( C h i c a g o :
T h e U n i v e r s i t y o f C h i c a g o P r e s s , 1 9 6 5 ) , c h a p t e r s 5 t o 1 2 . S e e a l s o l a t e r l e s s o n s i n t h i s r e p o r t .
1 6
4 . 2 D i f f e r e n c e s a n d C h a n g e s i n V i e w s
L e s s o n s
G r o u p s o r c o a l i t i o n s s u p p o r t i n g ( o r o p p o s i n g ) i n t e r n a t i o n a l c o n t r o l w i l l c o n t a i n i n d i v i d u a l s w h o h a v e d i f f e r e n t
r e a s o n s a n d r a t i o n a l e f o r t h e i r s u p p o r t ( o r o p p o s i t i o n ) . T h e i r m o t i v a t i o n s w i l l b e o n a s p e c t r u m f r o m t h e
c y n i c a l t o t h e i d e a l i s t i c . T h e i r v i e w s a n d s t r e n g t h o f o p p o s i t i o n ( o r s u p p o r t ) m a y v a r y s i g n i f i c a n t l y o v e r t i m e .
I n d i v i d u a l s m a y t a k e p o s i t i o n s f o r s h o r t - t e r m g a i n . T h o s e t h i n k i n g a b o u t g o v e r n a n c e o f p o w e r f u l t e c h n o l o g i e s
n e e d t o b e a w a r e t h a t s u p p o r t a n d o p p o s i t i o n m a y s h i f t a n d b e p r e p a r e d f o r i t . T h e y n e e d t o b u i l d c o a l i t i o n s
t h a t e n c o m p a s s a b r o a d r a n g e o f a g e n d a s a n d a p p r o a c h e s , a n d w o u l d b e n e f i t f r o m a “ p o l i t i c a l e n t r e p r e n e u r ” t o
h o l d d i v e r g e n t v i e w s t o g e t h e r , b u i l d c o n s e n s u s , a n d f o r g e a w a y f o r w a r d .
3 7
H i s t o r i c a l C a s e
I n d i v i d u a l s H a d D i f f e r e n t U n d e r s t a n d i n g s a b o u t t h e I m p a c t o f t h e A t o m i c B o m b
O n t h e o n e h a n d , s o m e t h o u g h t t h a t t h e a t o m i c b o m b w a s a u s e f u l w e a p o n o f w a r . T h e U . S . a r m y t h o u g h t t h a t
n u c l e a r w e a p o n s c o u l d p l a y a s i g n i f i c a n t , b u t n o t t r a n s f o r m a t i v e , t a c t i c a l r o l e i n s l o w i n g S o v i e t a r m i e s . T h e
n a v y t h o u g h t t h a t t h e a t o m i c b o m b w o u l d b e c e n t r a l t o s t o p p i n g a S o v i e t a t t a c k , a n d t h e a i r f o r c e e v e n t u a l l y
c a m e t o s e e i t a s a s t r a t e g i c w e a p o n t h a t c o u l d l a n d a k n o c k o u t b l o w ( t h e “ a i r - a t o m i c s t r a t e g y ” ) .
E x - P r i m e
3 8
M i n i s t e r W i n s t o n C h u r c h i l l t h o u g h t t h a t t h e a t o m i c b o m b c o u l d b e u s e d t o k e e p t h e S o v i e t U n i o n i n c h e c k
a n d a s s u r e U . S . a n d B r i t i s h d o m i n a t i o n o f w o r l d a f f a i r s i n t o t h e n e a r f u t u r e .
Y e t o t h e r s t h o u g h t t h a t i t h a d
3 9
t r a n s f o r m e d i n t e r n a t i o n a l r e l a t i o n s a n d w a r f a r e i n a n e g a t i v e w a y . T h e s e p e o p l e b e l i e v e d i t m a d e w a r s m o r e
d e s t r u c t i v e a n d s u i c i d a l , a n d i n c r e a s e d t h e l i k e l i h o o d t h a t s m a l l e r a g g r e s s i v e n a t i o n s ( n o w a r m e d w i t h a t o m i c
w e a p o n s ) w o u l d l a u n c h w a r s .
4 0
V i e w s
C h a n g e d
o v e r
T i m e
K e y p o l i c y m a k e r s ’ o p i n i o n s o n i n t e r n a t i o n a l c o n t r o l s h i f t e d o v e r t i m e . I n s o m e w a y s , t h e s e s h i f t s r e f l e c t e d
b r o a d e r s h i f t s i n o p i n i o n a b o u t t h e S o v i e t U n i o n o r t h e d e s t r u c t i v e n e s s o f a t o m i c b o m b s . O p i n i o n w a s
p a r t i c u l a r l y f l u i d d u r i n g t h e w a r . S e c r e t a r y o f W a r H e n r y L . S t i m s o n , f o r e x a m p l e , a d v i s e d P r e s i d e n t R o o s e v e l t
i n D e c e m b e r 1 9 4 4 n o t t o g i v e a t o m i c i n f o r m a t i o n t o t h e S o v i e t U n i o n w i t h o u t a “ r e a l q u i d p r o q u o , ” s u c h a s
l i b e r a l i z a t i o n o f d o m e s t i c S o v i e t r u l e . I n J u n e 1 9 4 5 , h e a d d e d t h a t t h e q u i d p r o q u o c o u l d i n c l u d e i n t e r n a t i o n a l
c o n t r o l , o r a n e g o t i a t e d s e t t l e m e n t o v e r t h e f a t e o f E a s t e r n E u r o p e . B y t h e t i m e o f t h e f i r s t a t o m i c b o m b t e s t i n
J u l y 1 9 4 5 , S t i m s o n h a d a b a n d o n e d h o p e o f s i m p l e c o o p e r a t i o n w i t h t h e S o v i e t U n i o n a n d i n s t e a d a r g u e d t h a t
t h e U . S . s h o u l d f o r c e S o v i e t l i b e r a l i z a t i o n a s a p r e c o n d i t i o n f o r c o o p e r a t i o n o n t h e a t o m i c b o m b . B y
3 7
F o r a r e v i e w o f t h e l i t e r a t u r e o n p o l i t i c a l e n t r e p r e n e u r s , s e e J . H o g a n a n d S . F e e n e y ,
T h e R o l e o f t h e P o l i t i c a l E n t r e p r e n e u r
i n t h e C o n t e x t o f P o l i c y C h a n g e a n d C r i s i s ,
M i d w e s t P o l i t i c a l S c i e n c e A s s o c i a t i o n A n n u a l C o n f e r e n c e , C h i c a g o , A p r i l 1 4 t h
2 0 1 3 . A v a i l a b l e a t
h t t p s : / / a r r o w . d i t . i e / c g i / v i e w c o n t e n t . c g i ? a r t i c l e = 1 0 1 5 & c o n t e x t = b u s c h g r a c o n
. A c c e s s e d 2 2 A p r i l 2 0 1 9 .
3 8
H e r k e n , p p . 2 0 2 - 0 4 , 2 1 2 . G e o r g e W . B a e r ,
O n e H u n d r e d Y e a r s o f S e a P o w e r : T h e U . S . N a v y , 1 8 9 0 - 1 9 9 0
( S t a n f o r d :
S t a n f o r d U n i v e r s i t y P r e s s , 1 9 9 6 ) , p p . 2 8 7 - 8 8 ; E d w a r d K a p l a n ,
T o K i l l N a t i o n s : A m e r i c a n S t r a t e g y i n t h e A i r - A t o m i c A g e
a n d t h e R i s e o f M u t u a l l y A s s u r e d D e s t r u c t i o n
( I t h a c a , N Y : C o r n e l l U n i v e r s i t y P r e s s , 2 0 1 5 ) , p p . 2 2 - 2 8 .
3 9
B a r t o n J . B e r n s t e i n , “ T h e U n e a s y A l l i a n c e : R o o s e v e l t , C h u r c h i l l , a n d t h e A t o m i c B o m b , 1 9 4 0 - 1 9 4 5 ” ,
T h e W e s t e r n
P o l i t i c a l Q u a r t e r l y
2 9 , 2 ( J u n e 1 9 7 6 ) , p p . 2 0 2 - 3 0 ; G r a h a m F a r m e l o ,
C h u r c h i l l ’ s B o m b : H o w t h e U n i t e d S t a t e s O v e r t o o k
B r i t a i n i n t h e F i r s t N u c l e a r A r m s R a c e
( N e w Y o r k : B a s i c B o o k s , 2 0 1 3 ) , p . 3 3 1 .
4 0
S e e , f o r e x a m p l e , A t o m i c S c i e n t i s t s o f C h i c a g o ,
T h e A t o m i c B o m b : F a c t s a n d I m p l i c a t i o n s
( C h i c a g o : T h e A t o m i c
S c i e n t i s t s o f C h i c a g o , 1 9 4 6 ) .
1 7
S e p t e m b e r , h e h a d g i v e n u p o n a q u i d p r o q u o , r e c o g n i z i n g h o w u n l i k e l y t h e S o v i e t U n i o n w a s t o m a k e t h o s e
c o n c e s s i o n s a n d h a v i n g g r e a t e r c o n c e r n a b o u t t h e d a n g e r s o f t h e a t o m i c b o m b . H e r e e m p h a s i z e d t h e i m p e r a t i v e
o f i n t e r n a t i o n a l c o n t r o l a n d a p r o p o s e d U . S . - U . K . - S o v i e t “ c o v e n a n t ” : t h a t t h e S o v i e t s r e f r a i n f r o m a t o m i c
d e v e l o p m e n t , a n d i n r e t u r n , t h e W e s t w o u l d s h a r e t h e p e a c e f u l a p p l i c a t i o n s o f a t o m i c e n e r g y a n d a g r e e n o t t o
e m p l o y t h e a t o m i c b o m b . O n e o f t h e r e a s o n s f o r S t i m s o n ’ s c o n v e r s i o n t o i n t e r n a t i o n a l c o n t r o l m a y h a v e b e e n
h i s i m p e n d i n g r e t i r e m e n t ; h e p r e s e n t e d h i s i n f l u e n t i a l m e m o r a n d u m c a l l i n g f o r i n t e r n a t i o n a l c o n t r o l a t a
C a b i n e t m e e t i n g i n S e p t e m b e r 1 9 4 5 j u s t p r i o r t o h i s r e t i r e m e n t . S t i m s o n m a y h a v e s e e n i t a s a n i d e a l p o l i c y t o
p u r s u e , p o t e n t i a l l y c r e a t i n g a l e g a c y b u t w i t h l i t t l e p e r s o n a l r i s k t o h i m s e l f .
O t h e r p o l i c y m a k e r s m a y h a v e
4 1
p u s h e d f o r i n t e r n a t i o n a l c o n t r o l f o r s i m i l a r r e a s o n s . S e c r e t a r y o f S t a t e B y r n e s w a s d e m o t e d i n D e c e m b e r 1 9 4 5 ,
a n d t h a t m a y h a v e b e e n a f a c t o r i n h i m s e t t i n g u p a S t a t e D e p a r t m e n t C o m m i t t e e t o l o o k i n t o i n t e r n a t i o n a l
c o n t r o l a s o n e o f h i s f i n a l a c t s .
T h i s s u g g e s t s a p o t e n t i a l l e s s o n : l o o k f o r o f f i c i a l s n e a r r e t i r e m e n t o r l e a v i n g
4 2
o f f i c e a s c a n d i d a t e s f o r e n a c t i n g m o r e i d e a l i s t i c p o l i c i e s , w i t h p a t h - d e p e n d e n t i m p a c t s .
P e r c e p t i o n s o f S o v i e t t h r e a t c o n t i n u e d t o h a v e a l a r g e i m p a c t o n i n d i v i d u a l s ’ o p i n i o n s i n t o 1 9 4 6 . A s t h e S o v i e t
U n i o n a p p e a r e d t o g r o w e c o n o m i c a l l y i n t h e f i r s t h a l f o f 1 9 4 6 a n d b e c o m e m o r e a s s e r t i v e i n i n t e r n a t i o n a l a f f a i r s
( e s p e c i a l l y i n r e l a t i o n t o T u r k e y ) , U . S . p o l i c y m a k e r s ( i n c l u d i n g B y r n e s a n d T r u m a n ) c a m e t o i n c r e a s i n g l y s e e
c o o p e r a t i o n a s i m p o s s i b l e .
T h e f a m o u s “ L o n g T e l e g r a m , ” s e n t t o t h e S t a t e D e p a r t m e n t i n F e b r u a r y 1 9 4 6 b y
4 3
G e o r g e F . K e n n a n , t h e c h a r g é d ' a f f a i r e s a t t h e U n i t e d S t a t e s E m b a s s y i n M o s c o w , t y p i f i e d t h i s r e p o s i t i o n i n g .
4 4
T h i s s h i f t r e d u c e d w i l l i n g n e s s t o n e g o t i a t e o n i n t e r n a t i o n a l c o n t r o l a n d l e d s e n i o r p o l i c y m a k e r s t o i n c r e a s i n g l y
s e e i n t e r n a t i o n a l c o n t r o l n e g o t i a t i o n s a s o n l y u s e f u l f o r p r o p a g a n d a p u r p o s e s . P o l i c y m a k e r s a l s o r e s p o n d e d t o
g r o w i n g p u b l i c c o n c e r n s o v e r S o v i e t s p y i n g . T r u m a n , f o r e x a m p l e , r e f u s e d t o s u p p o r t c i v i l i a n c o n t r o l o f a t o m i c
e n e r g y f o l l o w i n g a s p y s c a n d a l i n F e b r u a r y / M a r c h 1 9 4 6 . H e i n s t e a d b e n t t o p u b l i c o p i n i o n , w h i c h i n c r e a s i n g l y
p r e f e r r e d a s t r o n g m i l i t a r y r o l e i n c o n t r o l l i n g a t o m i c e n e r g y d o m e s t i c a l l y .
B y t h e m i d d l e o f 1 9 4 7 , e v e n
4 5
p r e v i o u s l y s t r o n g s u p p o r t e r s o f i n t e r n a t i o n a l c o n t r o l g a v e u p o n i t , h a v i n g d e c i d e d t h a t t h e S o v i e t U n i o n c o u l d
n o t b e t r u s t e d o r n e g o t i a t e d w i t h . O p p e n h e i m e r , f o r e x a m p l e , m e t B a r u c h ’ s r e p l a c e m e n t a s t h e U . S . d e l e g a t e o n
t h e U N A E C , F r e d e r i c k O s b o r n , s p e c i f i c a l l y t o r e q u e s t t h a t t h e U . S . w i t h d r a w f r o m a t o m i c c o n t r o l
n e g o t i a t i o n s .
4 6
B u i l d i n g a C o a l i t i o n o f S u p p o r t f r o m D i f f e r i n g C o n s t i t u e n c i e s w a s I m p o r t a n t
4 1
H e r k e n , p p . 1 4 , 1 6 , 1 9 , 2 4 - 7 , 2 9 - 3 1 . S e a n L . M a l l o y ,
A t o m i c T r a g e d y : H e n r y L . S t i m s o n a n d t h e D e c i s i o n t o U s e t h e B o m b
A g a i n s t J a p a n
( I t h a c a , N Y : C o r n e l l U n i v e r s i t y P r e s s , 2 0 0 8 ) , p p . 1 1 0 - 1 2 , 1 4 5 - 5 3 .
4 2
H e r k e n , p p . 9 2 , 9 7 , 9 8 ; L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a ,
p p . 2 3 4 - 5 .
4 3
H e r k e n , p p . 1 3 9 - 4 0 ; J o h n L e w i s G a d d i s ,
T h e C o l d W a r
, p p . 2 8 - 9 ; F e r n a n d e S c h e i d R a i n e , “ T h e I r a n i a n C r i s i s o f 1 9 4 6
a n d t h e O r i g i n s o f t h e C o l d W a r ” i n M e l v y n P . L e f f l e r a n d D a v i d S . P a i n t e r ( e d s . ) ,
O r i g i n s o f t h e C o l d W a r : A n
I n t e r n a t i o n a l H i s t o r y
2 n d e d .
( N e w Y o r k : R o u t l e d g e , 2 0 0 5 ) , p p . 9 3 - 1 1 1 ; a n d E d u a r d M a r k , " T h e T u r k i s h W a r S c a r e o f
1 9 4 6 , " i n i b i d . , p p . 1 1 2 - 3 3 .
4 4
J o h n L e w i s G a d d i s ,
S t r a t e g i e s o f C o n t a i n m e n t : A C r i t i c a l A p p r a i s a l o f A m e r i c a n N a t i o n a l S e c u r i t y P o l i c y d u r i n g t h e
C o l d W a r
2 n d e d i t i o n ( N e w Y o r k : O x f o r d U n i v e r s i t y P r e s s , 2 0 0 5 ) , p p . 5 3 - 4 . T h e “ L o n g T e l e g r a m ” w a s a n i n f l u e n t i a l n o t e
w r i t t e n b y K e n n a n a r g u i n g t h a t S o v i e t p o l i c y w a s i n h e r e n t l y e x p a n s i o n i s t a n d b a s e d o n a n e u r o t i c v i e w o f w o r l d a f f a i r s . H e
a r g u e d a g a i n s t c o o p e r a t i o n w i t h t h e S o v i e t U n i o n .
4 5
H e r k e n , p p . 1 3 5 - 3 6 .
4 6
J o s e p h I . L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a : T h e S t r u g g l e t o C o n t r o l A t o m i c W e a p o n s 1 9 4 5 - 1 9 4 9
( B o s t o n :
H o u g h t o n M i f f l i n C o m p a n y , 1 9 7 0 ) , p p . 3 9 0 - 9 1 .
1 8
I n d i v i d u a l s a n d i n i t i a t i v e s t h a t b u i l t s u p p o r t f r o m a r a n g e o f c o n s t i t u e n c i e s a n d v i e w p o i n t s w e r e m o r e l i k e l y t o
s u c c e e d . T h e A c h e s o n - L i l i e n t h a l R e p o r t g a i n e d t r a c t i o n n o t o n l y b e c a u s e i t w a s s p o n s o r e d b y t h e S t a t e
D e p a r t m e n t , b u t b e c a u s e i t h a d t h e p o w e r a n d p r e s t i g e o f i t s c o m m i t t e e m e m b e r s b e h i n d i t , w h o r e p r e s e n t e d a
r a n g e o f d i f f e r e n t i n t e r e s t s . T h e s e i n c l u d e d t h e l i b e r a l N e w D e a l e r D a v i d L i l i e n t h a l ( f o r m e r h e a d o f t h e
T V A T e n n e s s e e V a l l e y A u t h o r i t y ) , U n d e r S e c r e t a r y o f S t a t e D e a n A c h e s o n ( w h o h a d t h e s u p p o r t o f m u c h o f
t h e S t a t e D e p a r t m e n t ) , a n d J . R o b e r t O p p e n h e i m e r , w h o c a r r i e d w i t h h i m t h e s u p p o r t o f t h e A t o m i c S c i e n t i s t s ’
M o v e m e n t . E a r l y a t t e m p t s a t i n t e r n a t i o n a l c o n t r o l b y S e c r e t a r y o f S t a t e B y r n e s ( i n l a t e 1 9 4 5 ) , o n t h e o t h e r
h a n d , h a d f a i l e d b e c a u s e h e h a d e x c l u d e d c e r t a i n p o w e r f u l , b u t s k e p t i c a l , c o n g r e s s m e n . T h e y f e l t s l i g h t e d t h a t
h e h a d n o t s h a r e d p o l i c y m a k i n g w i t h t h e m , a n d s o u n d e r m i n e d h i s p o l i c y i n i t i a t i v e s .
4 7
4 7
L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a
, p p . 2 7 0 - 7 2 ; K i m b a l l - S m i t h ,
A P e r i l a n d a H o p e
, p p . 4 6 1 - 6 2 ; H e r k e n ,
p p . 7 4 - 8 7 .
1 9
4 . 3 C a u t i o u s o r C y n i c a l C o o p e r a t o r s
L e s s o n s
S c h e m e s f o r i n t e r n a t i o n a l g o v e r n a n c e c a n g a r n e r s u p p o r t f r o m “ r e a l i s t s , ” u n d e r s t o o d h e r e a s p o l i c y m a k e r s w h o
b e l i e v e i n t h e p r i m a c y a n d i n e v i t a b i l i t y o f p o w e r p o l i t i c s a n d w h o f o c u s f i r s t a n d f o r e m o s t o n n a t i o n a l i n t e r e s t .
4 8
T h e i r s u p p o r t , t h o u g h o f t e n c a u t i o u s o r c y n i c a l , c a n b e c r u c i a l , b u t i t c a n a l s o b e f i c k l e .
H i s t o r i c a l C a s e
A l t h o u g h r e a l i s t s w e r e d i s t u r b e d b y t h e i m p l i c a t i o n s o f a t o m i c w e a p o n s , t h e y d i d n o t s e e t h e m a s f u n d a m e n t a l l y
c h a n g i n g t h e w o r l d o r i n t e r n a t i o n a l r e l a t i o n s . R e a l i s t s t e n d e d t o b e l i e v e t h a t c o o p e r a t i n g o n t h e c o n t r o l o f
n u c l e a r w e a p o n s w a s f u t i l e a n d , i n a n y c a s e , b e s t d o n e f r o m a p o s i t i o n o f s t r e n g t h . T h e y t e n d e d t o n o t p e r c e i v e
a n u c l e a r a r m s r a c e a s p a r t i c u l a r l y d a n g e r o u s , e s p e c i a l l y c o m p a r e d t o t h e d a n g e r s f r o m m a k i n g o n e s e l f
v u l n e r a b l e t h r o u g h s t e p s t o w a r d s i n t e r n a t i o n a l c o n t r o l . S u c h s t e p s c o u l d b e d a n g e r o u s f o r t h e U . S . b e c a u s e t h e y
c o u l d l e a d t o t h e d i f f u s i o n o f c a p a b i l i t i e s a n d c o u l d u n d e r m i n e t h e c o u n t r y ’ s r e s o l v e t o r e s i s t C o m m u n i s m .
T h e y t h u s c o n c l u d e d t h a t t h e U . S . s h o u l d c o n t i n u e w i t h a t o m i c w e a p o n s d e v e l o p m e n t .
I n a ( r e a l i s t ) a n a l y s i s
4 9
o f a t o m i c w e a p o n s p r e s e n t e d t o C o n g r e s s i n J a n u a r y 1 9 4 6 , t h e f o r m e r h e a d o f t h e M a n h a t t a n P r o j e c t L e s l i e
G r o v e s p r e s e n t e d o n l y t w o a l t e r n a t i v e s f o r t h e U . S . : “ E i t h e r w e m u s t h a v e a h a r d - b o i l e d , r e a l i s t i c e n f o r c e a b l e
w o r l d a g r e e m e n t e n s u r i n g t h e o u t l a w i n g o f a t o m i c w e a p o n s o r w e a n d o u r d e p e n d a b l e a l l i e s m u s t h a v e a n
e x c l u s i v e s u p r e m a c y i n t h e f i e l d . ” I t w a s c l e a r t o h i m t h a t i f “ t h e r e a r e t o b e a t o m i c w e a p o n s i n t h e w o r l d , w e
m u s t h a v e t h e b e s t , t h e b i g g e s t a n d t h e m o s t . ”
5 0
T h e r e a l i s t c o m m i t m e n t t o t h e U . S . a t o m i c a r s e n a l w a s b o l s t e r e d b y t h e i r b e l i e f t h a t t h e S o v i e t U n i o n w o u l d
n o t b e a b l e t o b u i l d a n a t o m i c b o m b f o r m a n y y e a r s . G r o v e s , f o r e x a m p l e , g a v e v a r i o u s e s t i m a t e s b u t , f r o m
N o v e m b e r 1 9 4 5 o n w a r d s , u s u a l l y s a i d t w e n t y y e a r s .
T h e h i g h f i n a n c i a l c o s t s o f t h e M a n h a t t a n P r o j e c t a n d t h e
5 1
s c i e n t i f i c , t e c h n o l o g i c a l , i n d u s t r i a l , a n d o r g a n i z a t i o n a l h u r d l e s t h e U n i t e d S t a t e s h a d t o o v e r c o m e i n o r d e r t o
b u i l d t h e b o m b l e d m a n y t o b e l i e v e t h a t , e v e n i n t h e b e s t o f c i r c u m s t a n c e s , i t w o u l d t a k e t h e S o v i e t U n i o n
m u c h l o n g e r t o a c h i e v e t h i s f e a t .
E v e n i f t h e S o v i e t U n i o n m a n a g e d t o d e v e l o p a t o m i c b o m b s , m a n y b e l i e v e d
5 2
t h a t t h e U . S . c o u l d r e m a i n a h e a d i n a t o m i c w e a p o n s R & D , p r o d u c t i o n , d e p l o y m e n t , a n d d e l i v e r y o v e r t h e
c o m i n g d e c a d e s . M o r e o v e r , t h e S o v i e t e c o n o m y a p p e a r e d t o b e i n n o s h a p e t o t a k e t h e b u r d e n o f a n e x p e n s i v e
a t o m i c p r o g r a m . I t w a s s u f f e r i n g f r o m w a r t i m e d e v a s t a t i o n , o v e r b u r d e n e d w i t h t h e c o s t s o f t h e o c c u p a t i o n o f
E a s t e r n E u r o p e , a n d h a d a n e c o n o m y s t i l l t u n e d t o t h e p r o d u c t i o n o f ( c o n v e n t i o n a l ) m i l i t a r y f o r c e s .
I n
5 3
a d d i t i o n , G r o v e s b e l i e v e d t h a t t h e U . S . a n d i t s a l l i e s c o u l d m o n o p o l i z e n u c l e a r f u e l , t h u s d e l a y i n g o r s t o p p i n g
t h e S o v i e t a t o m i c p r o g r a m a l t o g e t h e r , a n d h e w o r k e d h a r d t o a c h i e v e t h i s . T h i s b e l i e f a p p e a r e d t o b e i m p o r t a n t
f o r h i s c o n f i d e n t a s s e r t i o n s t h a t t h e S o v i e t s w o u l d n o t s o o n g e t a t o m i c w e a p o n s , b u t h e d i d n o t s h a r e i t b e c a u s e
h e r e g a r d e d t h e U . S . ’ s m o n o p o l i z a t i o n e f f o r t s a s a s t a t e s e c r e t .
5 4
4 8
F o r d e f i n i t i o n s o f r e a l i s m a n d r e a l i s t d i s p o s i t i o n s , s e e D o n n e l l y ,
R e a l i s m a n d I n t e r n a t i o n a l R e l a t i o n s
, p p . 6 - 1 3 .
4 9
Z a c h a r y S . D a v i s , “ T h e R e a l i s t N u c l e a r R e g i m e ” ,
S e c u r i t y S t u d i e s
2 , 3 - 4 ( 1 9 9 3 ) , p p . 7 9 - 9 9 .
5 0
H e r k e n , p . 1 1 2 .
5 1
G o r d i n ,
R e d C l o u d a t D a w n
, p . 7 0 .
5 2
H e r k e n , p . 2 3 1 ; G o r d i n ,
R e d C l o u d a t D a w n
, p p . 7 0 - 1 .
5 3
H e r k e n , p . 1 3 8 .
5 4
I n a S e p t e m b e r 1 9 4 4 r e p o r t , f o r e x a m p l e , G r o v e s p r e d i c t e d t h a t t h e U . S . , t h r o u g h t h e C o m b i n e d D e v e l o p m e n t T r u s t ,
c o u l d c o n t r o l 9 0 % o f t h e w o r l d ’ s h i g h - g r a d e u r a n i u m o r e b y t h e w a r ’ s e n d . H e r k e n , p p . 1 0 1 - 0 2 . I n D e c e m b e r 1 9 4 5 , G r o v e s
c l a i m e d t h a t t h e T r u s t c o n t r o l l e d 9 7 % o f t h e w o r l d ’ s u r a n i u m o u t p u t a n d 6 5 % o f t h e w o r l d ’ s s u p p l y o f t h o r i u m , s e e
H o l l o w a y ,
S t a l i n a n d t h e B o m b
, p . 1 7 4 . A l s o : C h a r l e s A . Z i e g l e r , “ I n t e l l i g e n c e A s s e s s m e n t s o f S o v i e t A t o m i c C a p a b i l i t y ,
2 0
L a s t l y , t h e S o v i e t U n i o n e c o n o m y a p p e a r e d t o U . S . p o l i c y m a k e r s t o b e i n n o s h a p e t o p o s e a c h a l l e n g e t o U . S .
i n t e r e s t s . T h e c o u n t r y , t h e y r e a s o n e d , w a s s u f f e r i n g f r o m w a r t i m e d e v a s t a t i o n , o v e r b u r d e n e d w i t h t h e c o s t s o f
t h e o c c u p a t i o n o f E a s t e r n E u r o p e , f o c u s e d o n p o s t w a r r e c o n s t r u c t i o n , a n d s t i l l t u n e d t o t h e p r o d u c t i o n o f
( c o n v e n t i o n a l ) m i l i t a r y f o r c e s .
5 5
W h i l e r e a l i s t s r e p r e s e n t e d o n l y a m i n o r i t y o f t h e e l i t e s i n f a v o r o f i n t e r n a t i o n a l c o n t r o l , t h e i r s u p p o r t w a s
d i s p r o p o r t i o n a t e l y i m p o r t a n t , a s t h e y w e r e o f t e n p o w e r f u l i n d i v i d u a l s , e m b e d d e d h i g h i n t h e s t a t e , w i t h s t r o n g
i n f l u e n c e o n p o l i c y . H o w e v e r , w h i l e t h e y s u p p o r t e d i n t e r n a t i o n a l c o n t r o l , t h e y w e r e n o t
s t r o n g l y
c o m m i t t e d t o
i t . T h e y h a d , a t b e s t , a w e a k p r e f e r e n c e f o r i t a n d q u i c k l y a b a n d o n e d i t a s c i r c u m s t a n c e s c h a n g e d . E x a m p l e s
i n c l u d e B e r n a r d B a r u c h a n d p e r h a p s e v e n P r e s i d e n t T r u m a n h i m s e l f . B a r u c h w a s p i c k e d b y T r u m a n t o s h a p e
t h e U . S . p r o p o s a l o n i n t e r n a t i o n a l c o n t r o l a n d p r e s e n t i t a t t h e U n i t e d N a t i o n s A t o m i c E n e r g y C o m m i s s i o n
( U N A E C ) . B a r u c h q u i c k l y a b a n d o n e d h o p e f o r i n t e r n a t i o n a l c o n t r o l o n c e n e g o t i a t i o n s s t a l l e d . H i s t o r i a n s n o w
b e l i e v e t h a t h e w a s o n l y w e a k l y c o m m i t t e d t o i t , i f a t a l l .
T h e e x t e n t o f T r u m a n ’ s c o m m i t m e n t t o i n t e r n a t i o n a l
5 6
c o n t r o l i n e a r l y 1 9 4 6 i s u n c l e a r , b u t h e w a s c e r t a i n l y a t t r a c t e d t o t h e i d e a a n d i n t e r e s t e d e n o u g h t o a l l o w f o r a
U . S . p r o p o s a l t o b e f o r m u l a t e d a n d p l a c e d b e f o r e t h e U N A E C . W h a t i s c l e a r , h o w e v e r , i s t h a t h e q u i c k l y l o s t
i n t e r e s t b y m i d 1 9 4 6 o n c e h e b e c a m e c o n v i n c e d t h a t a g g r e s s i v e S o v i e t f o r e i g n p o l i c y c o u l d n o t b e m e t w i t h
c o n c e s s i o n s .
5 7
T h e r e w e r e a n u m b e r o f r e a s o n s f o r w h y r e a l i s t s t h o u g h t t h a t i n t e r n a t i o n a l c o n t r o l w o u l d s u p p o r t t h e i r
o b j e c t i v e t o m a x i m i z e U . S . p o w e r i n i n t e r n a t i o n a l a f f a i r s :
( 1 )
B y s u p p o r t i n g a p o l i c y t h a t w a s p o p u l a r a m o n g s t t h e p u b l i c , t h e y h o p e d t o b o o s t t h e i r o w n p o p u l a r i t y .
T h e y w a n t e d t o b e s e e n a d d r e s s i n g p u b l i c c o n c e r n s a b o u t a t o m i c w e a p o n s . C a r r y i n g o u t n e g o t i a t i o n s ,
t h e y c a l c u l a t e d , w o u l d b e e n o u g h t o m e e t t h i s c o n c e r n . I f n e g o t i a t i o n s f a i l e d , t h e y h o p e d t o p l a c e t h e
b l a m e o n t h e S o v i e t U n i o n , t h u s h i g h l i g h t i n g t h e S o v i e t U n i o n a s , a t b e s t , a n u n r e l i a b l e p a r t n e r i n
i n t e r n a t i o n a l a f f a i r s , a n d a t w o r s t , a t h r e a t t o U . S . s e c u r i t y .
5 8
( 2 )
S o m e s u p p o r t e d i n t e r n a t i o n a l c o n t r o l f o r a d m i n i s t r a t i v e p o l i t i c a l r e a s o n s . O n e o f t h e r e a s o n s w h y
S e c r e t a r y o f S t a t e B y r n e s f o r m e d a c o m m i t t e e o f c o n s u l t a n t s t o l o o k i n t o i n t e r n a t i o n a l c o n t r o l i n
1 9 4 5 - 1 9 4 9 : M y t h s , M o n o p o l i e s a n d M a s k i r o v k a ” ,
I n t e l l i g e n c e a n d N a t i o n a l S e c u r i t y
1 2 , 4 ( 1 9 9 7 ) , p p . 1 - 2 4 ; G o r d i n ,
R e d
C l o u d a t D a w n
, p p . 7 2 - 7 4 ; J o n a t h a n E . H e l m r e i c h ,
G a t h e r i n g R a r e O r e s : T h e D i p l o m a c y o f U r a n i u m A c q u i s i t i o n ,
1 9 4 3 - 1 9 5 4
( P r i n c e t o n , N J : P r i n c e t o n U n i v e r s i t y P r e s s , 1 9 8 6 ) , p p . 2 4 8 - 9 .
5 5
H e r k e n , p . 1 3 8 . C a r l S h u l m a n h a s p o i n t e d o u t t h a t t h i s c o n c l u s i o n i s i n s o m e w a y s s u r p r i s i n g g i v e n t h a t t h e G D P o f t h e
S o v i e t U n i o n a t t h a t t i m e w a s r o u g h l y 1 4 t i m e s t h e c o s t o f t h e M a n h a t t a n P r o j e c t . T h i s i s c a l c u l a t e d b y t a k i n g t h e
e s t i m a t e d G D P o f t h e S o v i e t U n i o n i n 1 9 4 6 ( t h i s i s a r o u g h e s t i m a t e , a n d n o r e l i a b l e f i g u r e f o r 1 9 4 5 i s a v a i l a b l e ) a s U S D
6 6 4 . 6 4 6 b i l l i o n ( i n 2 0 1 1 U S D f r o m t h e M a d d i s o n P r o j e c t D a t a b a s e 2 0 1 8 a t
h t t p s : / / w w w . r u g . n l / g g d c / h i s t o r i c a l d e v e l o p m e n t / m a d d i s o n /
, a c c e s s e d 2 8 M a y 2 0 2 0 ) a n d t h e c o s t o f t h e M a n h a t t a n
P r o j e c t a s U S D 1 . 8 8 9 b i l l i o n ( i n 1 9 4 2 – 1 9 4 5 d o l l a r s , f r o m S t e p h e n I . S c h w a r t z ,
A t o m i c A u d i t : T h e C o s t s a n d C o n s e q u e n c e s
o f U . S . N u c l e a r W e a p o n s S i n c e 1 9 4 0
( W a s h i n g t o n , D C : B r o o k i n g s I n s t i t u t i o n P r e s s , 1 9 9 8 ) , p . 6 0 ) . C o n v e r t i n g t h e l a t t e r
f i g u r e t o 2 0 1 1 d o l l a r s u s i n g
h t t p s : / / w w w . m e a s u r i n g w o r t h . c o m / c a l c u l a t o r s / u s c o m p a r e /
g i v e s a M a n h a t t a n P r o j e c t c o s t
f i g u r e ( u s i n g a p r o d u c t i o n w o r k e r c o m p e n s a t i o n i n f l a t o r ) a s U S D 4 8 . 4 b i l l i o n .
5 6
M a d d o c k ,
N u c l e a r A p a r t h e i d
, c h a p t e r 3 .
5 7
H e r k e n , p p . 1 3 9 - 4 0 , 1 7 5 . S . D a v i d B r o s c i o u s , “ L o n g i n g f o r I n t e r n a t i o n a l C o n t r o l , B a n k i n g o n A m e r i c a n S u p e r i o r i t y :
H a r r y S T r u m a n ’ s A p p r o a c h t o N u c l e a r W e a p o n s ” , i n J o h n L e w i s G a d d i s e t a l ( e d s . ) ,
C o l d W a r S t a t e s m e n C o n f r o n t t h e
B o m b : N u c l e a r D i p l o m a c y S i n c e 1 9 4 5
( O x f o r d : O x f o r d U n i v e r s i t y P r e s s , 1 9 9 9 ) , p p . 1 5 - 3 8 . O n T r u m a n ’ s g r o w i n g
c o n f r o n t a t i o n a l s t a n c e t o w a r d s t h e S o v i e t U n i o n , s e e G a d d i s ,
T h e C o l d W a r
, p p . 2 8 - 9 ; R a i n e , “ T h e I r a n i a n C r i s i s o f 1 9 4 6
a n d t h e O r i g i n s o f t h e C o l d W a r ” ; M a r k , " T h e T u r k i s h W a r S c a r e o f 1 9 4 6 ” .
5 8
C r a i g a n d R a d c h e n k o ,
T h e A t o m i c B o m b a n d t h e O r i g i n s o f t h e C o l d W a r
, p . 1 3 0 . S e e a l s o :
P u b l i c S p h e r e l e s s o n
.
2 1
J a n u a r y 1 9 4 6 w a s t h a t h e w a n t e d t o r e t a i n a t o m i c f o r e i g n p o l i c y m a k i n g a n d e x p e r t i s e w i t h i n t h e S t a t e
D e p a r t m e n t , r a t h e r t h a n t o l o s e i t t o C o n g r e s s o r s o m e o t h e r s t a t e e n t i t y .
5 9
( 3 )
S o m e m a y h a v e b e l i e v e d t h a t t h e o f f i c i a l U . S . p r o p o s a l , t h e B a r u c h P l a n , w a s d e s i g n e d t o s a f e g u a r d U . S .
n a t i o n a l s e c u r i t y i n t e r e s t s i n t h a t i t i n c l u d e d t h e m a x i m u m p o s s i b l e c o n c e s s i o n s t h a t t h e U . S . c o u l d
m a k e w i t h o u t j e o p a r d i z i n g i t s s e c u r i t y . T h e y c o n s e q u e n t l y s u p p o r t e d i t b e c a u s e i t w a s n o t o n l y t h e b e s t
p o s s i b l e i n t e r n a t i o n a l c o n t r o l p l a n b u t a l s o b e c a u s e i t m i t i g a t e d e n o u g h r i s k f o r t h e U . S . t o b e
a c c e p t a b l e . B e r n a r d B a r u c h a n d h i s n e g o t i a t o r s , h i s t o r i a n s l a r g e l y c o n c l u d e , d i d n o t t r y h a r d e r t o r e a c h
a n a g r e e m e n t w i t h t h e S o v i e t s f o r t h i s v e r y r e a s o n . B a r u c h a n d h i s a s s o c i a t e s b e l i e v e d t h a t t h e p l a n w a s
t h e b e s t t h a t t h e U . S . c o u l d o f f e r , a n d p r e f e r r e d a n o - d e a l s c e n a r i o t h a t w o u l d s t i l l l e a v e t h e U . S . i n a
d o m i n a n t p o s i t i o n i n a t o m i c w e a p o n s f o r d e c a d e s t o c o m e .
6 0
5 9
H e r k e n , p . 9 7 ; L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a ,
p p . 2 3 4 - 5 .
6 0
M a d d o c k ,
N u c l e a r A p a r t h e i d
, c h a p t e r 3 .
2 2
4 . 4 S e c r e c y a n d S e c u r i t y
L e s s o n s
S e c r e c y a n d s e c u r i t y w i l l p l a y a c e n t r a l r o l e i n a n y d i s c u s s i o n o n t h e g o v e r n a n c e o f p o w e r f u l t e c h n o l o g i e s . T h e y
c a n h a v e a s i g n i f i c a n t e f f e c t o n t h e p o s s i b i l i t y o f i n t e r n a t i o n a l c o o p e r a t i o n a s w e l l a s o n i n t r a s t a t e p o w e r
s t r u g g l e s . T h e y c a n g i v e t r e m e n d o u s p o w e r t o i n d i v i d u a l s a n d s t a t e i n s t i t u t i o n s ( s u c h a s t h e m i l i t a r y ) w h i c h
c o n t r o l t h e f l o w o f i n f o r m a t i o n a n d , i n p a r t i c u l a r , c a n b e u s e d t o u n d e r m i n e o p p o n e n t s . S e c r e c y c a n b e t e r r i b l e
f o r e p i s t e m i c s , u n d e r m i n i n g c o m p e t e n t o r g a n i z a t i o n a l d e l i b e r a t i o n . S e c r e c y i s o f t e n a n t i t h e t i c a l t o c o o p e r a t i o n
a n d t r u s t , i n p a r t b e c a u s e t h e p u b l i c a n d a c t o r s w h o a r e o u t s i d e s e c r e t a c c e s s a r e o f t e n m o r e i n f a v o r o f
c o o p e r a t i o n a n d t r u s t .
S e c r e c y c a n b e u s e d t o e m p o w e r n a r r a t i v e s o f f e a r a n d b e l l i g e r e n c e .
6 1
P o l i c y m a k e r s n e e d t o c a r e f u l l y w e i g h d e c i s i o n s t o e x p a n d s e c r e t d o m a i n s . T h e y n e e d t o m a k e s u r e s u c h
d e c i s i o n s a r e c o u n t e r b a l a n c e d w i t h i n i n s t i t u t i o n s a n d t h a t a w i d e r a n g e o f p e r s p e c t i v e s a r e u s e d t o i n f o r m
p o l i c y m a k i n g . T h i s w i l l r e d u c e t h e r i s k s o f c o r r u p t i o n a n d a b u s e a s w e l l a s d e c i s i o n - m a k i n g f r o m a n o v e r l y
n a r r o w p e r s p e c t i v e . P o l i c y m a k e r s a l s o n e e d t o b e w a r y o f a r g u m e n t s t h a t w a r n o f i m m i n e n t s e c u r i t y t h r e a t s .
S u c h n a r r a t i v e s c a n e a s i l y l e a d t o i n c r e a s e d s e c r e c y . P o l i c y m a k e r s s h o u l d o b t a i n a w i d e r a n g e o f v i e w s s o a s t o g e t
b e t t e r q u a l i t y i n f o r m a t i o n a n d m a k e i n f o r m e d d e c i s i o n s . T h e y s h o u l d a l s o e n s u r e t h a t t e c h n i c a l e x p e r t s a r e
i n v o l v e d i n k e y s t r a t e g i c d e c i s i o n s .
H i s t o r i c a l C a s e
A p p e a l s t o t h e s e c u r i t y o f t h e U n i t e d S t a t e s a n d w a r n i n g s o f i m m i n e n t s e c u r i t y t h r e a t s w e r e a n i n t r i n s i c p a r t o f
d e b a t e s o v e r i n t e r n a t i o n a l c o n t r o l o f a t o m i c e n e r g y . T h e s e w e r e n o t c o n s t a n t b u t w a x e d a n d w a n e d i n i n t e n s i t y
a n d i n f l u e n c e . O n e m a j o r d r o p i n p u b l i c a n d e l i t e s u p p o r t f o r i n t e r n a t i o n a l c o n t r o l ( a n d c i v i l i a n c o n t r o l o f
d o m e s t i c a t o m i c e n e r g y ) o c c u r r e d d u e t o t h e s e n s a t i o n a l r e v e l a t i o n o f S o v i e t “ a t o m i c s p i e s ” i n F e b r u a r y 1 9 4 6 .
T h e S o v i e t U n i o n i n c r e a s i n g l y d i d n o t a p p e a r t o b e r e l i a b l e a l l i e s a n d , i n s t e a d , w e r e s e e n a s h a v i n g m a l i g n i n t e n t
a n d a g r o w i n g a t o m i c b o m b p r o g r a m .
6 2
M o r e s p e c i f i c a l l y , s e c r e c y g a v e t r e m e n d o u s p o w e r t o t h o s e w h o c o n t r o l l e d t h e f l o w o f i n f o r m a t i o n , n o t a b l y
L e s l i e R . G r o v e s ( t h e h e a d o f t h e M a n h a t t a n P r o j e c t ) .
G r o v e s a c c r u e d p o w e r a n d r e s p o n s i b i l i t y f o r h i m s e l f b y
6 3
t a k i n g a d v a n t a g e o f h i s p r i v i l e g e d a c c e s s t o i n f o r m a t i o n a n d h i s a b i l i t y t o d e m a r c a t e t h e b o u n d a r i e s a n d
h i e r a r c h i e s o f s e c r e c y . H e b e g a n w i t h r e s p o n s i b i l i t y f o r b u i l d i n g t h e p l a n t s a n d f a c t o r i e s t o m a k e a t o m i c f u e l .
T h i s e v o l v e d i n t o r e s p o n s i b i l i t y f o r s c i e n t i f i c r e s e a r c h , w e a p o n d e s i g n , a n d a t o m i c s e c u r i t y , i n t e l l i g e n c e , a n d
c o u n t e r i n t e l l i g e n c e . H e w a s a l s o e v e n t u a l l y i n v o l v e d i n h i g h - l e v e l p o l i c y m a k i n g o n b o t h d o m e s t i c a n d
6 1
N o t e t h a t , i n s o m e c a s e s , s e c r e c y c a n p e r m i t c o o p e r a t i v e g e s t u r e s t h a t a h a w k i s h / n a t i o n a l i s t i c p u b l i c w o u l d p u n i s h
( e s p e c i a l l y i f u n s u c c e s s f u l ) a n d c a n a l l o w f o r c o l l u s i o n t o p r e v e n t e s c a l a t i o n . C o u n t e r e x a m p l e s i n c l u d e o u t r e a c h i n t h e s t y l e
o f N i x o n t o C h i n a , t h e O s l o A c c o r d s , a n d p r e v e n t i n g e s c a l a t i o n f r o m S o v i e t - U S c o n f l i c t i n K o r e a . A . C a r s o n “ F a c i n g O f f
a n d S a v i n g F a c e : C o v e r t I n t e r v e n t i o n a n d E s c a l a t i o n M a n a g e m e n t i n t h e K o r e a n W a r ” ,
I n t e r n a t i o n a l O r g a n i z a t i o n
7 0 , 1
( 2 0 1 6 ) , p p . 1 0 3 - 3 1 . A l s o : A l l i s o n C a r n e g i e a n d A u s t i n C a r s o n , “ T h e D i s c l o s u r e D i l e m m a : N u c l e a r I n t e l l i g e n c e a n d
I n t e r n a t i o n a l O r g a n i z a t i o n s ” ,
T h e A m e r i c a n J o u r n a l o f P o l i t i c a l S c i e n c e
6 3 , 2 ( 2 0 1 9 ) , p p . 2 6 9 - 2 8 5 .
6 2
H e r k e n , p p . 1 2 7 , 1 3 2 , 1 3 6 ; G r e g g H e r k e n , “ ‘ A M o s t D e a d l y I l l u s i o n ’ : T h e A t o m i c S e c r e t a n d A m e r i c a n N u c l e a r
W e a p o n s P o l i c y , 1 9 4 5 - 1 9 5 0 ” ,
P a c i f i c H i s t o r i c a l R e v i e w
4 9 , 1 ( F e b r u a r y 1 9 8 0 ) , p p . 5 1 - 7 6 ; K i m b a l l - S m i t h ,
A P e r i l a n d a
H o p e
, p p . 3 7 3 - 5 , 3 8 7 - 8 ; H o g a n ,
A C r o s s o f I r o n
, p . 2 3 8 .
6 3
O n G r o v e s ’ “ c o m p a r t m e n t a l i z a t i o n ’ s t r a t e g y s e e : S h e r w i n ,
A W o r l d D e s t r o y e d
, p p . 5 8 - 6 2 .
2 3
i n t e r n a t i o n a l i s s u e s , a n d i n t h e p l a n n i n g a n d e x e c u t i o n o f t h e a t o m i c b o m b i n g m i s s i o n s .
U s i n g h i s p r i v i l e g e d
6 4
a c c e s s t o s e c r e t i n f o r m a t i o n , G r o v e s w a s a b l e t o w i n a r g u m e n t s w i t h a d v e r s a r i e s b y p o i n t i n g t o t h e i r i g n o r a n c e ,
e . g . , a b o u t t h e d u r a t i o n o f t h e U . S . m o n o p o l y .
H e w a s a b l e t o a v o i d h i s p o l i c i e s b e i n g q u e s t i o n e d , e . g . , b y
6 5
C o n g r e s s . E v e n h i s c o l l a b o r a t o r s f o u n d t h e m s e l v e s p u l l e d a l o n g b y G r o v e s ’ f a i t s a c c o m p l i s .
H e d i s c r e d i t e d
6 6
o t h e r s a s b e i n g u n c a r e f u l w i t h s e c r e t i n f o r m a t i o n o r e v e n b e i n g t r e a s o n o u s . H e d i d t h i s w i t h N i e l s B o h r , J .
R o b e r t O p p e n h e i m e r , L e o S z i l a r d , a n d D a v i d L i l i e n t h a l .
G r o v e s u s e d h i s a c c e s s t o s e c r e t i n f o r m a t i o n t o s h a p e
6 7
p o l i c y m a k i n g b y r e c o m m e n d i n g h i s “ t e c h n i c a l a d v i s o r s ” f o r k e y d e c i s i o n s . F o r e x a m p l e , h e a l m o s t s c u t t l e d
L i l i e n t h a l ’ s c o n s u l t a n t g r o u p o n i n t e r n a t i o n a l c o n t r o l b y r e c o m m e n d i n g h i s t e c h n i c a l a d v i s o r s i n s t e a d .
G r o v e s
6 8
a l s o u s e d h i s p r i v i l e g e d a c c e s s t o d i r e c t l y i n f l u e n c e k e y p o l i c y m a k e r s s u c h a s T r u m a n . T r u m a n ’ s b e l i e f i n e a r l y
1 9 4 6 t h a t t h e U . S . w o u l d h a v e a l o n g - l i v e d a t o m i c m o n o p o l y w a s d u e d i r e c t l y t o G r o v e s ’ a r g u m e n t s a n d
i n f l u e n c e .
G r o v e s w a s a l s o a b l e t o s h a p e p u b l i c o p i n i o n a n d w h i p u p p u b l i c a n d p o l i c y m a k e r c o n c e r n o v e r
6 9
s e c u r i t y a n d s p y i n g .
7 0
S e c r e c y l e d t o p o o r i n f o r m a t i o n f l o w a n d s o t o b a d d e c i s i o n - m a k i n g . T w o e x a m p l e s o f t h i s a r e e s p e c i a l l y
p r o m i n e n t . F i r s t , G r o v e s i m p o s e d h i s v i e w o f t h e d u r a t i o n o f t h e U . S . a t o m i c m o n o p o l y o n t h e U . S .
g o v e r n m e n t b y p r e v e n t i n g r e a s o n e d d e b a t e . I n l a t e 1 9 4 4 a n d e a r l y 1 9 4 5 , V a n n e v a r B u s h , J a m e s B . C o n a n t , a n d
L e o S z i l a r d — w h o d i s a g r e e d w i t h G r o v e s a n d b e l i e v e d t h a t t h e S o v i e t U n i o n c o u l d a c c e s s h i g h - g r a d e u r a n i u m
o r e a n d s o b u i l d a n a t o m i c b o m b r e l a t i v e l y q u i c k l y — w e r e s t i f l e d t h r o u g h s e c r e c y r e g u l a t i o n s ( s p e c i f i c a l l y t h e
s i l o i n f o r m a t i o n s t r u c t u r e w i t h i n t h e M a n h a t t a n P r o j e c t ) .
G r o v e s ’ s v i e w s a l s o c a r r i e d s i g n i f i c a n t w e i g h t w i t h
7 1
t h e p u b l i c . O n e h i s t o r i a n h a s n o t e d t h a t “ G r o v e s ’ s p r e d i c t i o n s c a r r i e d m o r e w e i g h t w i t h t h e p u b l i c t h a n a n y o n e
e l s e ’ s , p r e c i s e l y b e c a u s e h e w a s t h e i n d i v i d u a l e x p e c t e d t o h a v e t h e l a r g e s t a m o u n t o f s e c r e t i n f o r m a t i o n u p o n
w h i c h t o b a s e a n e s t i m a t e . F e w c h a l l e n g e d h i m d i r e c t l y … . ”
. S e c o n d , i n f o r m a t i o n o n a t o m i c b o m b p r o d u c t i o n
7 2
a n d s t o c k p i l e s w a s s u c h a c l o s e l y g u a r d e d s e c r e t w i t h i n t h e m i l i t a r y t h a t p a r t s o f t h e m i l i t a r y c o n c e r n e d w i t h
a t o m i c w e a p o n u s e w e r e d e n i e d c r u c i a l i n f o r m a t i o n . W a r p l a n s w e r e c r e a t e d i n 1 9 4 5 , 1 9 4 6 , a n d 1 9 4 7 , f o r
e x a m p l e , w i t h l i t t l e u n d e r s t a n d i n g o f t h e s i z e o f t h e U . S . a t o m i c a r s e n a l a n d i t s d e l i v e r a b i l i t y , o r b a s e d o n
i n a c c u r a t e e s t i m a t e s o f i t s s i z e . S t a f f o f f i c e r s d r a w i n g u p p l a n s d i d n o t e v e n k n o w f o r c e r t a i n h o w t h e y w e r e
a l l o w e d t o u s e a t o m i c w e a p o n s . O n e J a n u a r y 1 9 4 6 p l a n f r o m t h e J o i n t W a r P l a n s C o m m i t t e e d e s i g n a t e d 1 7
S o v i e t c i t i e s a s t a r g e t s f o r w h i c h i t a s s u m e d t h e A i r C o r p s w o u l d r e q u i r e 9 8 a t o m i c b o m b s p l u s 9 8 i n r e s e r v e ,
6 4
A c c o r d i n g t o G r o v e s ’ b i o g r a p h e r , s e c u r i t y t h r o u g h c o m p a r t m e n t a l i z a t i o n w a s t h e “ s e c r e t o f G r o v e s ’ s p o w e r . ” R o b e r t S .
N o r r i s ,
R a c i n g f o r t h e B o m b : T h e T r u e S t o r y o f G e n e r a l L e s l i e R . G r o v e s , t h e M a n b e h i n d t h e B i r t h o f t h e A t o m i c A g e
( H a n o v e r , N H . : S t e e r f o r t h P r e s s , 2 0 0 2 ) , p p . x i i i , 1 1 , 1 8 5 . A l s o H e r k e n , p p . 1 1 0 - 1 1 .
6 5
H e r k e n , p p . 1 0 9 - 1 0 ; M i c h a e l G o r d i n ,
R e d C l o u d a t D a w n : T r u m a n , S t a l i n , a n d t h e E n d o f t h e A t o m i c M o n o p o l y
( N e w
Y o r k : F a r r a r , S t r a u s a n d G i r o u x , 2 0 0 9 ) ,
p p . 7 3 - 5 ; C h a r l e s A . Z i e g l e r a n d D a v i d J a c o b s o n ,
S p y i n g W i t h o u t S p i e s : O r i g i n s o f
A m e r i c a ' s S e c r e t N u c l e a r S u r v e i l l a n c e S y s t e m
( W e s t p o r t , C T : P r a e g e r , 1 9 9 5 ) , p p . 2 4 - 5 .
6 6
N o r r i s ,
R a c i n g f o r t h e B o m b
, p . 1 4 .
6 7
H e r k e n , p . 2 4 2 ; W i l l i a m L a n o u e t t e a n d B e l a S i l a r d ,
G e n i u s i n t h e S h a d o w s : A B i o g r a p h y o f L e o S z i l a r d , t h e M a n B e h i n d
t h e B o m b
( N e w Y o r k : C . S c r i b n e r ’ s S o n s , 1 9 9 2 ) , p . 3 1 0 ; M i c h a e l J . H o g a n ,
A C r o s s o f I r o n : H a r r y S . T r u m a n a n d t h e
O r i g i n s o f t h e N a t i o n a l S e c u r i t y S t a t e
( C a m b r i d g e : C a m b r i d g e U n i v e r s i t y P r e s s , 2 0 0 0 ) , p . 2 4 4 , 2 5 8 .
6 8
H e r k e n , p . 1 5 4 .
6 9
H e r k e n , p . 9 9 ; Z i e g l e r , “ I n t e l l i g e n c e A s s e s s m e n t s o f S o v i e t A t o m i c C a p a b i l i t y , 1 9 4 5 - 1 9 4 9 ” ; Z i e g l e r a n d J a c o b s o n ,
S p y i n g
W i t h o u t S p i e s
, p p . 2 4 - 5 .
7 0
H e r k e n , p . 2 7 3 . P a r t i c u l a r l y d u r i n g t h e S o v i e t s p y s c a r e o f F e b r u a r y 1 9 4 6 ; s e e C r a i g a n d R a d c h e n k o ,
T h e A t o m i c B o m b
a n d t h e O r i g i n s o f t h e C o l d W a r
, p p . 1 2 1 - 2 2 ; R e g i n a l d W h i t a k e r ,
C o l d W a r C a n a d a : T h e M a k i n g o f a N a t i o n a l I n s e c u r i t y
S t a t e , 1 9 4 5 - 1 9 5 7
( T o r o n t o : U n i v e r s i t y o f T o r o n t o P r e s s , 1 9 9 4 ) , p . 9 3 ; H o g a n ,
A C r o s s o f I r o n
, p . 2 5 8 .
7 1
H e r k e n , p . 1 1 0 - 1 1 ; G o r d i n ,
R e d C l o u d a t D a w n ,
p p . 7 5 - 8 .
7 2
G o r d i n ,
R e d C l o u d a t D a w n
, p . 7 5 .
2 4
g i v i n g a t o t a l o f r e q u i r e m e n t o f 1 9 6 . A t t h e e n d o f 1 9 4 6 , t h e a c t u a l n u m b e r o f b o m b s i n U . S . p o s s e s s i o n w a s
a r o u n d 7 ! M o r e o v e r , n o n e o f t h e b a s e s f r o m w h i c h t h e a t o m i c s t r i k e s w e r e t o b e l a u n c h e d w e r e e q u i p p e d w i t h
a t o m i c w e a p o n s l o a d i n g p i t s o r a t o m i c s t o r a g e f a c i l i t i e s . T h e J o i n t C h i e f s o f S t a f f e v e n o f f i c i a l l y c o m p l a i n e d
t h a t t h e y w e r e d e n i e d a c c e s s t o i n t e l l i g e n c e c o l l e c t e d a n d h e l d b y G r o v e s . T h i s l e d t o s i g n i f i c a n t d i s a g r e e m e n t
a n d i n c o n s i s t e n c y b e t w e e n v a r i o u s m i l i t a r y p l a n s o n t h e u s e o f a t o m i c w e a p o n s .
7 3
S e c r e c y l e d t o a p o o r p u b l i c u n d e r s t a n d i n g o f d a m a g e t o c i v i l i a n p o p u l a t i o n s f r o m a t o m i c w a r f a r e . I n o n e f a l l
1 9 4 5 p u b l i c h e a r i n g , G r o v e s ' c a s u a l c o m m e n t t h a t a n a t o m i c w a r w o u l d “ o n l y ” l e a d t o 4 0 m i l l i o n U . S . c a s u a l t i e s
s h o c k e d t h e a u d i e n c e .
S o m e p o l i c y m a k e r s s a w b e n e f i t i n h i d i n g t h e d e s t r u c t i v e n e s s o f t h e a t o m i c b o m b f r o m
7 4
t h e p u b l i c . O n e l a t e r J o i n t C h i e f s r e p o r t n o t e d t h a t “ A s i t u a t i o n d a n g e r o u s t o o u r s e c u r i t y c o u l d r e s u l t f r o m
i m p r e s s i n g o n o u r o w n d e m o c r a t i c p e o p l e s t h e h o r r o r s o f f u t u r e w a r s o f m a s s d e s t r u c t i o n w h i l e t h e p o p u l a t i o n s
o f t h e ' p o l i c e ' s t a t e s r e m a i n u n a w a r e o f t h e t e r r i b l e i m p l i c a t i o n s . ”
7 5
C o n c e r n s o v e r t h e p o s s i b l e l o s s o f t h e “ a t o m i c s e c r e t , ” d r i v e n b y r e p o r t s o f S o v i e t s p i e s i n F e b r u a r y 1 9 4 6 ,
e m p o w e r e d n a r r a t i v e s o f f e a r a n d h a w k i s h n e s s . W i t h o u t f u l l e r i n f o r m a t i o n , i t w a s e a s i e r f o r t h e p u b l i c t o p a n i c
a b o u t s e c u r i t y r i s k s . I n s u c h a n a t m o s p h e r e , t h o s e p o s s e s s i n g t o p s e c r e t c l e a r a n c e ( e s p e c i a l l y G r o v e s ) w e r e
e m p o w e r e d . I n a n a t m o s p h e r e o f i n c r e a s e d s e c u r i t y c o n s c i o u s n e s s , i t w a s e a s i e r f o r t h e m i l i t a r y a n d i t s
s u p p o r t e r s t o g a i n b a c k i n g f o r t h e i r v i e w s o n a t o m i c t e c h n o l o g y .
I t a l s o r e d u c e d s u p p o r t f o r i n t e r n a t i o n a l
7 6
c o o p e r a t i o n o n t h e a t o m i c b o m b a m o n g s t t h e p u b l i c , a n d e v e n s c i e n t i s t s r e d u c e d t h e i r s u p p o r t f o r i n t e r n a t i o n a l
c o n t r o l a s r u m o r s c i r c u l a t e d t h a t t h e H o u s e U n - A m e r i c a n A c t i v i t i e s C o m m i t t e e w a s t o i n v e s t i g a t e O a k R i d g e
s c i e n t i s t s f o r s e c u r i t y t h r e a t s .
S e c r e c y m a y h a v e c o n t r i b u t e d t o t h e l a c k o f t r u s t b e t w e e n t h e U . S . a n d t h e S o v i e t
7 7
U n i o n , a n d p o s s i b l y a l s o b e t w e e n t h e U . S . a n d B r i t a i n . I t a l s o l e d t o t e n s i o n s b e t w e e n G r o v e s a n d t h e c i v i l i a n s
a n d s c i e n t i s t s w o r k i n g o n a t o m i c p o l i c y , w h o r e s e n t e d h i s c o n t r o l o v e r a t o m i c i n f o r m a t i o n a n d h i s u s e o f
s e c r e c y t o w i n a r g u m e n t s .
7 8
7 3
H e r k e n , p p . 1 9 9 - 2 0 0 , 2 1 9 - 2 0 , 2 2 9 ; S t e v e n T . R o s s ,
A m e r i c a n W a r P l a n s 1 9 4 5 - 1 9 5 0
( N e w Y o r k : R o u t l e d g e , 2 0 1 3 ) , p p .
1 2 - 1 7 ; Z i e g l e r , “ I n t e l l i g e n c e A s s e s s m e n t s o f S o v i e t A t o m i c C a p a b i l i t y ” . T h e J a n u a r y 1 9 4 6 p l a n i s n o t e d i n R o s s ,
A m e r i c a n
W a r P l a n s
, p p . 1 6 - 1 7 . O n t h e n u m b e r o f a t o m i c b o m b s , s e e J o h n M . C u r a t o l a ,
B i g g e r B o m b s f o r a B r i g h t e r T o m o r r o w :
S t r a t e g i c A i r C o m m a n d a n d A m e r i c a n W a r P l a n s a t t h e D a w n o f t h e A t o m i c A g e , 1 9 4 5 - 1 9 5 0
( J e f f e r s o n , N C : M a c F a r l a n d ,
2 0 1 6 ) , p p . 5 3 .
7 4
H e r k e n , p . 2 2 2 .
7 5
H e r k e n , p . 2 2 1 .
7 6
H e r k e n , p p . 1 2 7 , 1 3 2 , 1 3 6 ; G r e g g H e r k e n , “ ‘ A M o s t D e a d l y I l l u s i o n ’ : T h e A t o m i c S e c r e t a n d A m e r i c a n N u c l e a r
W e a p o n s P o l i c y , 1 9 4 5 - 1 9 5 0 ” ,
P a c i f i c H i s t o r i c a l R e v i e w
4 9 , 1 ( F e b r u a r y 1 9 8 0 ) , p p . 5 1 - 7 6 ; H o g a n ,
A C r o s s o f I r o n
, p . 2 3 8 .
T h e r e a r e e c h o e s o f t h i s i n c u r r e n t d a y c o n c e r n s o v e r C h i n e s e c y b e r e s p i o n a g e i n t h e U . S . S e e , f o r e x a m p l e , S a m m S a c k s ,
“ T h e R i p p l e E f f e c t s o f t h e C h i n a C h i p H a c k i n g S t o r y ” ,
W a r o n t h e R o c k s
( 1 7 O c t o b e r 2 0 1 8 ) . A t :
h t t p s : / / w a r o n t h e r o c k s . c o m / 2 0 1 8 / 1 0 / t h e - r i p p l e - e f f e c t s - o f - t h e - c h i n a - c h i p - h a c k i n g - s t o r y
. A c c e s s e d 2 4 A p r i l 2 0 1 9 .
7 7
K i m b a l l - S m i t h ,
A P e r i l a n d a H o p e
, p p . 3 7 3 - 5 , 3 8 7 - 8 .
7 8
O n t h e r e l a t i o n s h i p w i t h B r i t a i n s e e H e r k e n , p p . 3 6 , 5 8 , 6 1 - 3 , 1 0 3 - 0 5 , 1 4 6 - 4 7 . O n s c i e n t i s t s ’ c o n c e r n s w i t h t h e m i l i t a r y ’ s
s e c r e c y o n a t o m i c m a t t e r s , s e e K i m b a l l - S m i t h ,
A P e r i l a n d a H o p e
, p p . 2 4 2 - 3 6 1 .
2 5
4 . 5 P u b l i c S p h e r e
L e s s o n s
T h e p u b l i c s p h e r e w i l l h a v e a p o w e r f u l i m p a c t o n d e b a t e s o n i n t e r n a t i o n a l c o n t r o l . P o l i t i c a l a n d p o l i c y m a k i n g
e l i t e s w i l l b e s e n s i t i v e t o l a r g e s h i f t s i n p u b l i c o p i n i o n , w h i c h w i l l ( f o r e x a m p l e ) i n f l u e n c e t h e i r p o l i t i c a l
p r o s p e c t s . T h e y w i l l s e e k t o m o b i l i z e p u b l i c o p i n i o n i n s u p p o r t o f t h e i r p r e f e r r e d p o l i c i e s . P u b l i c o p i n i o n o n
t h i s i s s u e i s m a l l e a b l e a n d c a n b e s h a p e d i n s u p p o r t o f o r a g a i n s t p a r t i c u l a r p o l i c i e s , p e o p l e , i n s t i t u t i o n s , o r
c o u n t r i e s .
A l t h o u g h e l i t e o p i n i o n r e m a i n s p a r a m o u n t , p a r t i c i p a n t s i n d e b a t e s o n i n t e r n a t i o n a l c o n t r o l b e n e f i t f r o m
h a r n e s s i n g t h e p o w e r o f t h e p u b l i c s p h e r e a n d s h a p i n g p u b l i c o p i n i o n . T h e y s h o u l d r u n p u b l i c i t y c a m p a i g n s ,
m a k e a r g u m e n t s t h a t w o u l d a p p e a l t o t h e p u b l i c , a n d g a r n e r s u p p o r t f r o m i n d i v i d u a l s a n d g r o u p s w i t h h i g h
p u b l i c p r o f i l e s .
H i s t o r i c a l C a s e
I n r e l a t i o n t o t h e i n t e r n a t i o n a l c o n t r o l o f a t o m i c e n e r g y , a l t h o u g h e l i t e o p i n i o n w a s p a r a m o u n t , p o l i c y m a k e r s
n e v e r t h e l e s s w e r e i n f l u e n c e d b y a n d r e a c t e d t o p u b l i c o p i n i o n .
D o m e s t i c p r e s s u r e s , f o r e x a m p l e , f o r c e d
7 9
T r u m a n t o c o m m i t t h e U . S . t o t h e p r i n c i p l e o f i n t e r n a t i o n a l c o n t r o l b e f o r e S e c r e t a r y o f S t a t e B y r n e s h a d e v e n
a t t e m p t e d t o e x t r a c t a q u i d p r o q u o f r o m M o s c o w i n l a t e 1 9 4 5 .
P u b l i c c o n c e r n s o v e r t h e s a f e t y o f t h e “ a t o m i c
8 0
s e c r e t ” a n d S o v i e t s p i e s l e d T r u m a n t o m a i n t a i n h i g h e r s e c r e c y a r o u n d a t o m i c o p e r a t i o n s t h a n m i g h t o t h e r w i s e
h a v e b e e n i d e a l .
A l t e r n a t i v e s t o i n t e r n a t i o n a l c o n t r o l , s u c h a s t h e c o n c e p t o f a n “ a t o m i c l e a g u e , ” w e r e n o t f u l l y
8 1
e x p l o r e d a s t h e y e m p h a s i z e d p r e v e n t i v e w a r , w h i c h w a s o u t o f f a v o r w i t h t h e p u b l i c .
B a r u c h b e c a m e o b s e s s e d
8 2
w i t h e v e n t h e s m a l l e s t s h i f t i n p u b l i c s u p p o r t f o r h i s p r o p o s a l s d u r i n g U N A E C n e g o t i a t i o n s i n S e p t e m b e r
1 9 4 6 .
P o l i c i e s w e r e s o m e t i m e s v a g u e l y s t a t e d i n o r d e r t o s a t i s f y t h e p u b l i c : N S C - 3 0 , t h e “ P o l i c y o n A t o m i c
8 3
W a r f a r e ” r e l e a s e d a t t h e e n d o f 1 9 4 8 , f o r e x a m p l e , a v o i d e d t h e i s s u e o f f i r s t u s e o f a t o m i c w e a p o n s i n o r d e r t o
n o t u p s e t p u b l i c o p i n i o n .
B a r u c h a n d h i s a d v i s o r y t e a m , w h e n t h e y h e a r d o f S o v i e t d e l e g a t e A n d r e i
8 4
G r o m y k o ’ s r e j e c t i o n o f t h e B a r u c h P l a n i n J u n e 1 9 4 6 , d i d n o t w a n t t o o p e n l y r e j e c t t h e S o v i e t c o u n t e r p r o p o s a l
( t h e s o - c a l l e d “ G r o m y k o P l a n ” ) s o e a r l y i n t h e n e g o t i a t i o n s b u t n e v e r t h e l e s s w a n t e d t o s i g n a l t h e i r r e j e c t i o n .
T h e y t h u s l e a k e d a s e r i e s o f s t o r i e s t o t h e p r e s s f r o m “ a n o n y m o u s b u t r e l i a b l e s o u r c e s ” t h a t t h e U . S . d e l e g a t i o n
c o u l d n o t a c c e p t t h e G r o m y k o P l a n .
8 5
E l i t e s s h a p e d p u b l i c o p i n i o n i n t h e i r f a v o r , u s i n g p r e s s r e l e a s e s , b r i e f i n g s , a n d p u b l i c a t i o n s . G r o v e s , f o r e x a m p l e ,
m a y h a v e b e e n t h e c o n f i d e n t i a l s o u r c e c i t e d b y n e w s r e p o r t s t h a t b r o k e s e n s a t i o n a l i s t n e w s o f S o v i e t s p i e s
7 9
A n o t h e r e x a m p l e o f t h e i n f l u e n c e o f p u b l i c o p i n i o n o n n u c l e a r a r m s c o n t r o l p o l i c y m a k i n g i s J a m e s C a m e r o n ,
T h e
D o u b l e G a m e T h e D e m i s e o f A m e r i c a ' s F i r s t M i s s i l e D e f e n s e S y s t e m a n d t h e R i s e o f S t r a t e g i c A r m s L i m i t a t i o n
( N e w Y o r k :
O x f o r d U n i v e r s i t y P r e s s , 2 0 1 7 ) .
8 0
H e r k e n , p . 5 2 .
8 1
H e r k e n , p . 1 3 6 ; H o g a n ,
A C r o s s o f I r o n
, p . 2 3 8 .
8 2
H e r k e n , p . 2 6 5 .
8 3
H e r k e n , p . 1 7 9 - 8 0 .
8 4
H e r k e n , p . 2 6 8 ; S t e v e n M i l l e r ,
S t r a t e g y a n d N u c l e a r D e t e r r e n c e : A n I n t e r n a t i o n a l S e c u r i t y R e a d e r
( P r i n c e t o n , N J :
P r i n c e t o n U n i v e r s i t y P r e s s , 1 9 8 4 ) ,
p . 1 2 3 .
8 5
L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a
, p . 3 1 1 .
2 6
o p e r a t i n g i n C a n a d a a n d t h e U n i t e d S t a t e s i n F e b r u a r y 1 9 4 6 . T h e s e n e w s s t o r i e s i n c r e a s e d s u p p o r t f o r t h e
M a y - J o h n s o n B i l l , w h i c h a d v o c a t e d m i l i t a r y c o n t r o l o f a t o m i c e n e r g y p o l i c y .
8 6
P u b l i c o p i n i o n w a s n o t u n i f o r m o r e n t i r e l y c o h e r e n t . I t s o m e t i m e s c o n t a i n e d v i e w s t h a t w e r e i n t e n s i o n w i t h
o n e a n o t h e r . F o r e x a m p l e , s u r v e y s i n S e p t e m b e r 1 9 4 5 r e v e a l e d t h a t ~ 7 0 % o f c i t i z e n s d i d n o t w a n t t o s h a r e t h e
s e c r e t o f t h e a t o m i c b o m b w i t h o t h e r c o u n t r i e s . A t t h e s a m e t i m e , h o w e v e r , 9 0 % t h o u g h t t h a t t h e U . S . w o u l d
n o t b e a b l e t o k e e p t h e s e c r e t f o r l o n g a n y w a y a n d o t h e r c o u n t r i e s w o u l d s o o n b u i l d t h e b o m b .
P o l l s i n
8 7
O c t o b e r 1 9 4 5 s h o w e d 1 7 % s u p p o r t f o r i n t e r n a t i o n a l c o n t r o l t h r o u g h t h e U . N . S e c u r i t y C o u n c i l , b u t 6 7 %
s u p p o r t f o r “ E n g l a n d , R u s s i a , F r a n c e , [ t h e ] U n i t e d S t a t e s , C h i n a , a n d o t h e r c o u n t r i e s ” t o “ g e t t o g e t h e r t o a g r e e
t h a t a t o m i c b o m b s s h o u l d n e v e r b e u s e d a s a w a r w e a p o n . ”
A 1 9 4 7 p o l l s h o w e d t h a t a m a j o r i t y b e l i e v e d t h a t
8 8
a t o m i c b o m b s m a d e w a r l e s s l i k e l y b u t w e r e a l s o w i l l i n g t o i n i t i a t e a n a t o m i c w a r .
T h e p u b l i c h a d e r r o n e o u s
8 9
t e c h n i c a l b e l i e f s , s u c h a s t h a t , b y O c t o b e r 1 9 4 7 , ~ 6 0 % o f t h e p u b l i c s u r v e y e d “ t h o u g h t t h a t R u s s i a w a s
m a n u f a c t u r i n g a t o m i c b o m b s i n q u a n t i t y " ; i r o n i c a l l y , w h i l e t h e p u b l i c w a s v e r y m i s t a k e n h e r e , t h e y w e r e
c o m p a r a b l y m i s t a k e n b u t i n t h e o p p o s i t e d i r e c t i o n a s t h e m o s t i n f o r m e d e x p e r t , G e n e r a l G r o v e s .
9 0
8 6
G r o v e s w o u l d h i m s e l f l a t e r d i s c o u n t t h e e f f e c t i v e n e s s a n d i m p o r t a n c e o f t h i s s p y i n g . H e r k e n , p . 1 3 0 - 3 3 . T h e r e i s a
p o s s i b i l i t y t h a t t h e s o u r c e w a s i n t h e F B I ( p e r h a p s t h e D i r e c t o r h i m s e l f , J . E d g a r H o o v e r ) o r f r o m w i t h i n t h e J u s t i c e
D e p a r t m e n t ; s e e : E l l e n S c h r e c k e r ,
M a n y A r e t h e C r i m e s : M c C a r t h y i s m i n A m e r i c a
( P r i n c e t o n , N J : P r i n c e t o n U n i v e r s i t y
P r e s s , 1 9 9 9 ) , p . 1 7 0 ; C r a i g a n d R a d c h e n k o ,
T h e A t o m i c B o m b a n d t h e O r i g i n s o f t h e C o l d W a r
, p p . 1 2 1 - 2 2 .
8 7
H e r k e n , p . 3 2 .
8 8
H a d l e y C a n t r i l ,
P u b l i c O p i n i o n 1 9 3 5 - 1 9 4 6
( P r i n c e t o n , N J : P r i n c e t o n U n i v e r s i t y P r e s s , 1 9 5 1 ) , p . 2 2 .
8 9
H e r k e n , p . 3 1 1 .
9 0
H e r k e n , p . 2 3 2 .
2 7
4 . 6 T e c h n i c a l E x p e r t s
L e s s o n s
T e c h n i c a l e x p e r t s a n d s p e c i a l i s t s ( s c i e n t i s t s , e n g i n e e r s , t e c h n i c i a n s , a c a d e m i c s ) h a v e s i g n i f i c a n t p o w e r t o s h a p e
p r o p o s a l s a n d p o l i c y t o b e m o r e e f f e c t i v e a n d m o r e c o o p e r a t i v e , t h o u g h t h e i r i n f l u e n c e d e p e n d s o n t h e i r
p o l i t i c a l s o p h i s t i c a t i o n .
T e c h n i c a l e x p e r t s s h o u l d b e g i v e n a c e n t r a l r o l e i n d r a f t i n g p r o p o s a l s , p o l i c y m a k i n g , a n d p u b l i c e n g a g e m e n t .
E x p e r t s s h o u l d i n v e s t t i m e i n u n d e r s t a n d i n g t h e p o l i t i c a l l a n d s c a p e a n d i d e n t i f y i n g p o l i t i c a l a l l i e s ; t h e i r m o s t
i m p o r t a n t c o n t r i b u t i o n t o t h e p r o b l e m s o f p o w e r f u l t e c h n o l o g i e s m a y b e f r o m t h e i r s h a p i n g o f p o l i t i c a l
d i s c u s s i o n s , r a t h e r t h a n m a r g i n a l l y m o r e s c i e n t i f i c o r a c a d e m i c w o r k .
H i s t o r i c a l C a s e
F r o m 1 9 4 4 t o 1 9 4 6 , a t o m i c s c i e n t i s t s p l a y e d a c e n t r a l r o l e i n p r o p o s a l s f o r i n t e r n a t i o n a l c o n t r o l . T h e y w e r e
l e a d i n g a d v o c a t e s , p a s s i o n a t e a n d c o m m i t t e d . T h e y w e r e s o m e o f t h e f i r s t t o w a r n o f t h e d a n g e r s o f a t o m i c
e n e r g y , t h e y f o r m e d a d v o c a c y g r o u p s a n d r a i s e d f u n d s f o r t h e i r a c t i v i s m , a n d t h e y c a r r i e d o u t p u b l i c
e n g a g e m e n t t o e x p l a i n a t o m i c e n e r g y a n d w a r n o f i t s d a n g e r s . T h e y p l a y e d a n i m p o r t a n t r o l e i n p r o v i d i n g
i n f o r m a t i o n a n d a d v i c e t o e l i t e s a n d t o t h e p u b l i c , w i t h w h o m t h e y g a r n e r e d e n o r m o u s r e s p e c t , a u t h o r i t y , a n d
c r e d i b i l i t y .
F o r e m o s t a m o n g s t t h e m w a s J . R o b e r t O p p e n h e i m e r ( h e a d o f t h e L o s A l a m o s l a b o r a t o r y d u r i n g
9 1
t h e w a r ) , w h o w a s c o n s i d e r e d a n a t i o n a l h e r o , t h e “ f a t h e r o f t h e a t o m i c b o m b , ” a n d m a d e t h e c o v e r o f
T i m e
m a g a z i n e i n N o v e m b e r 1 9 4 8 .
S c i e n t i s t s w e r e a l s o c e n t r a l i n s h a p i n g p r o p o s a l s . T h e A c h e s o n - L i l i e n t h a l R e p o r t
9 2
w a s d r a f t e d l a r g e l y b y O p p e n h e i m e r , w h o b r o u g h t i n t o i t m a n y o f t h e i d e a s o f t h e A t o m i c S c i e n t i s t s ’
M o v e m e n t .
9 3
S c i e n t i s t s a l s o p r o v i d e d t e c h n i c a l a n d s t r a t e g i c i n s i g h t s f o r e l i t e s . L e o S z i l a r d , f o r e x a m p l e , l o b b i e d h a r d t o
c o n v i n c e t h e U . S . g o v e r n m e n t t h a t t h e y n e e d e d t o s t a r t a n a t o m i c b o m b p r o g r a m i n 1 9 3 9 .
J a m e s B . C o n a n t ,
9 4
O p p e n h e i m e r , a n d V a n n e v a r B u s h h a d s i g n i f i c a n t i n p u t o n a t o m i c p o l i c y m a k i n g d u r i n g t h e w a r . T h e w a r t i m e
“ S c i e n t i f i c A d v i s o r y P a n e l ” ( s t a f f e d b y s c i e n t i s t s O p p e n h e i m e r , E r n e s t L a w r e n c e , E n r i c o F e r m i , a n d A r t h u r H .
C o m p t o n ) w a s i n s t r u m e n t a l i n d e t e r m i n i n g h o w t h e a t o m i c b o m b w o u l d b e u s e d .
I n 1 9 4 4 a n d 1 9 4 5 , o t h e r
9 5
s c i e n t i s t s i n f o r m e d p o l i c y m a k e r s o f t h e s t r a t e g i c i m p l i c a t i o n s o f a t o m i c w e a p o n s i n r e p o r t s a n d m e m o s . M a n y
9 1
O n t h e i r a c t i v i s m , s e e K i m b a l l - S m i t h ,
A P e r i l a n d a H o p e
, c h a p t e r s 5 t o 1 2 . H i s t o r i a n J e s s i c a W a n g h a s s u m m a r i z e d t h e i r
i m p a c t t h u s : “ D e s p i t e t h e i r i n d i v i d u a l a n o n y m i t y , t h e a t o m i c s c i e n t i s t s s o o n e s t a b l i s h e d a p o w e r f u l p r e s e n c e i n A m e r i c a n
p o l i t i c a l l i f e . T h e y a p p e a l e d d i r e c t l y t o t h e p u b l i c t h r o u g h a l o n g s e r i e s o f m e d i a i n t e r v i e w s , a r t i c l e s , r a d i o a d d r e s s e s , a n d
p u b l i c s p e a k i n g e n g a g e m e n t s i n w h i c h t h e y d i s c u s s e d b o t h t h e s p e c i f i c l e g i s l a t i o n a t h a n d a n d t h e g e n e r a l p o l i t i c a l a n d
s o c i a l i m p l i c a t i o n s o f a t o m i c e n e r g y . . . B e t w e e n O c t o b e r a n d D e c e m b e r 1 9 4 5 , s o m e t h i r t y - o d d s c i e n t i s t s w e n t t o
W a s h i n g t o n , w h e r e , i n a w h i r l w i n d o f s o c i a l a n d p o l i t i c a l a c t i v i t y , t h e y b u i l t i n f l u e n c e i n e x c e s s o f t h e i r n u m b e r s . ” J e s s i c a
W a n g ,
A m e r i c a n S c i e n c e i n a n A g e o f A n x i e t y : S c i e n t i s t s , A n t i c o m m u n i s m , a n d t h e C o l d W a r
( C h a p e l H i l l , N C : U n i v e r s i t y
o f N o r t h C a r o l i n a P r e s s , 1 9 9 9 ) , p . 1 6 .
9 2
K a i B i r d a n d M a r t i n J . S h e r w i n ,
A m e r i c a n P r o m e t h e u s : T h e T r i u m p h a n d T r a g e d y o f J . R o b e r t O p p e n h e i m e r
( N e w Y o r k :
V i n t a g e B o o k s , 2 0 0 6 ) , x i .
9 3
H e r k e n , p p . 1 5 5 - 5 8 .
9 4
B a r t o n J . B e r n s t e i n , “ S c i e n t i s t s a n d N u c l e a r W e a p o n s i n W o r l d W a r I I ” , i n T h o m a s W . Z e i l e r a n d D a n i e l M . D u B o i s
( e d s . ) ,
A C o m p a n i o n t o W o r l d W a r I I
v o l u m e 1 ( O x f o r d : W i l e y - B l a c k w e l l , 2 0 1 3 ) , p p . 5 1 6 - 4 8 .
9 5
I b i d .
2 8
o f t h e s e w a r n e d t h a t t h e r e c o u l d b e a n a r m s r a c e , t h a t t h e r e w a s n o s u c h t h i n g a s a n “ a t o m i c s e c r e t ” t h a t c o u l d
b e k e p t , a n d t h a t t h e S o v i e t s w o u l d a c q u i r e t h e a t o m i c b o m b w i t h i n a f e w y e a r s . T h e s e w a r n i n g s t u r n e d o u t t o
b e m o r e a c c u r a t e t h a n G r o v e s ’ p r e d i c t i o n s o n w h e n t h e S o v i e t U n i o n w o u l d g e t t h e a t o m i c b o m b .
S c i e n t i s t s
9 6
a l s o p l a y e d a k e y r o l e i n t h e d e v e l o p m e n t o f m o n i t o r i n g t e c h n o l o g i e s a n d s t r a t e g i e s i n t h e l a t e 4 0 s .
9 7
T h e m o s t d e t a i l e d p r o p o s a l f o r i n t e r n a t i o n a l c o n t r o l , t h e A c h e s o n - L i l i e n t h a l P l a n , c o u l d n o t h a v e b e e n d r a w n
u p w i t h o u t t h e p a r t i c i p a t i o n o f a s c i e n t i s t . O t h e r s w o r k i n g o n t h e p l a n k n e w v e r y l i t t l e , i f a n y t h i n g a t a l l , a b o u t
t h e w o r k i n g s o f a t o m i c e n e r g y o r t h e a t o m i c b o m b . B e c a u s e o f h i s t e c h n i c a l e x p e r t i s e , O p p e n h e i m e r t o o k t h e
l e a d i n e d u c a t i n g o t h e r c o m m i t t e e m e m b e r s a b o u t a t o m i c e n e r g y a n d t h e n i n d r a f t i n g t h e p l a n .
H e w a s
9 8
u n i q u e l y a b l e t o s u g g e s t p o t e n t i a l t e c h n i c a l d e v e l o p m e n t s ( s u c h a s s e p a r a b i l i t y a n d d e n a t u r i n g ) t o e a s e t h e
p r o b l e m s i n h e r e n t t o i n t e r n a t i o n a l c o n t r o l .
9 9
O n t h e o t h e r h a n d , s c i e n t i s t s w e r e s o m e t i m e s p e r c e i v e d a s n a i v e , e s p e c i a l l y b y p o l i t i c i a n s a n d d i p l o m a t s .
S c i e n t i s t s ( e s p e c i a l l y t h o s e i n t h e A t o m i c S c i e n t i s t s ’ M o v e m e n t o r s u p p o r t e r s o f w o r l d g o v e r n m e n t ) w e r e o f t e n
d e r i d e d f o r t h e i r i d e a l i s t i c v i e w s a n d l a c k o f u n d e r s t a n d i n g o f t h e p o l i t i c a l l y p o s s i b l e . P r o m i n e n t S t a t e
D e p a r t m e n t o f f i c i a l G e o r g e K e n n a n , f o r e x a m p l e , r e p o r t e d b a c k t o t h e S t a t e D e p a r t m e n t i n A u g u s t 1 9 4 6 ,
f o l l o w i n g m e e t i n g s w i t h m e m b e r s o f t h e A t o m i c S c i e n t i s t s ’ M o v e m e n t , t h a t “ [ p ] o l i t i c a l l y , t h e s e p e o p l e a r e a s
i n n o c e n t a s s i x - y e a r - o l d m a i d e n s . I n t r y i n g t o e x p l a i n t h i n g s t o t h e m I f e l t l i k e o n e w h o s h a t t e r s t h e p u r e i d e a l s
o f t e n d e r y o u t h . ”
E i n s t e i n ’ s c a l l s f o r w o r l d g o v e r n m e n t i n 1 9 4 7 / 4 8 ( w h i c h i n c l u d e d i n t e r n a t i o n a l c o n t r o l o f
1 0 0
n u c l e a r w e a p o n s ) w e r e s i m i l a r l y d e r i d e d b y t h e S t a t e D e p a r t m e n t a s “ n a i v e … T h e m a n w h o p o p u l a r i z e d t h e
c o n c e p t o f t h e f o u r t h d i m e n s i o n c o u l d t h i n k i n o n l y t w o o f t h e m i n c o n s i d e r a t i o n o f W o r l d G o v e r n m e n t . ”
1 0 1
A f t e r a m e e t i n g i n O c t o b e r 1 9 4 5 , O p p e n h e i m e r w a s d e s c r i b e d b y T r u m a n a s a “ c r y - b a b y s c i e n t i s t ” w h o h a d
c o m e t o h i s o f f i c e a n d “ s p e n t m o s t o f h i s t i m e w r i n g i n g h i s h a n d s a n d t e l l i n g m e t h e y h a d b l o o d o n t h e m
b e c a u s e o f h i s d i s c o v e r y o f a t o m i c e n e r g y . ”
1 0 2
S c i e n t i s t s w e r e a l s o v u l n e r a b l e t o t h e c r i t i c i s m t h a t t h e y w e r e a s e c u r i t y t h r e a t . T h i s w a s d u e l a r g e l y t o t h e i r
o p e n n e s s , t h e i r i n t e r n a t i o n a l c o n n e c t i o n s a n d c o m m u n i c a t i o n s ( f o r e x a m p l e w i t h E a s t e r n b l o c s c i e n t i s t s ) , a n d
t h e i r ( g e n e r a l l y ) p r o g r e s s i v e p o l i t i c s . O n e o f t h e e a r l i e s t c a l l s f o r i n t e r n a t i o n a l c o n t r o l , b y D a n i s h p h y s i c i s t N i e l s
B o h r t o W i n s t o n C h u r c h i l l a n d F r a n k l i n D . R o o s e v e l t i n 1 9 4 4 , w a s m e t w i t h t h e r e s p o n s e f r o m C h u r c h i l l t h a t
9 6
F o r e x a m p l e , t h r o u g h t h e J u n e 1 9 4 5 F r a n c k R e p o r t a n d t h e N o v e m b e r 1 9 4 4 J e f f r i e s R e p o r t : K i m b a l l S m i t h ,
A P e r i l a n d
a H o p e ,
p p . 1 9 - 2 4 , 4 1 - 8 .
9 7
G o r d i n ,
R e d C l o u d a t D a w n ,
p p . 1 8 9 - 2 1 3 .
9 8
B i r d a n d S h e r w i n ,
A m e r i c a n P r o m e t h e u s ,
p . 3 4 1 .
9 9
O p p e n h e i m e r s u g g e s t e d t h a t a n e w t e c h n i q u e c a l l e d i s o t o p i c d e n a t u r i n g c o u l d b e u s e d t o p r e v e n t m i s u s e o f u r a n i u m
r e a c t o r f u e l f o r n u c l e a r w e a p o n s . I s o t o p i c d e n a t u r i n g m e a n t t h e a d d i t i o n o f a d i f f e r e n t i s o t o p e o f n u c l e a r f u e l w h i c h w o u l d
r e n d e r t h e f u e l u s e l e s s a s a n e x p l o s i v e . T h e f u e l , h o w e v e r , c o u l d s t i l l b e u s e d f o r p o w e r r e a c t o r s . S u c h a n i s o t o p e , i t w a s
t h o u g h t , c o u l d n o t b e c h e m i c a l l y s e p a r a t e d . T h i s w a s t h e f o u n d a t i o n f o r h i s i d e a t h a t m i l i t a r y u s e s o f a t o m i c e n e r g y c o u l d
b e p r a c t i c a l l y s e p a r a t e d f r o m c i v i l i a n u s e s . T h e m i l i t a r y u s e s c o u l d t h e n b e t i g h t l y c o n t r o l l e d b y a n i n t e r n a t i o n a l a u t h o r i t y .
C i v i l i a n u s e s c o u l d b e s a f e l y l e f t t o s o v e r e i g n s t a t e s . S e e : B a r n a r d e t a l ,
A R e p o r t o n t h e I n t e r n a t i o n a l C o n t r o l o f A t o m i c
E n e r g y
, p p . 2 6 - 7 ; a n d H e r k e n , p p . 1 5 5 - 5 9 . T h e r e l i a n c e o n d e n a t u r i n g t u r n e d o u t t o b e m i s g u i d e d ; s e e
l a t e r s e c t i o n s
.
1 0 0
L a w r e n c e S . W i t t n e r ,
T h e S t r u g g l e A g a i n s t t h e B o m b
v o l u m e 1 : O n e W o r l d o r N o n e : A H i s t o r y o f t h e W o r l d N u c l e a r
D i s a r m a m e n t M o v e m e n t T h r o u g h 1 9 5 3 ( S t a n f o r d : S t a n f o r d U n i v e r s i t y P r e s s , 1 9 9 3 ) , p . 2 6 4 .
1 0 1
H e r k e n , p . 2 6 4 .
1 0 2
R a y M o n k ,
R o b e r t O p p e n h e i m e r : A L i f e I n s i d e t h e C e n t e r
( N e w Y o r k : D o u b l e d a y , 2 0 1 2 ) , p . 4 9 4 .
2 9
“ [ B o h r ] i s v e r y n e a r t h e e d g e o f m o r t a l c r i m e s ” f o r d i s c u s s i n g a t o m i c m a t t e r s w i t h S o v i e t c i t i z e n s .
T h e r e w e r e
1 0 3
a f e w p u b l i c d e n o u n c e m e n t s o f s c i e n t i s t s i n t h e f o r t i e s ; o n e o f t h e m o s t s e r i o u s w a s i n M a r c h 1 9 4 8 w h e n t h e
H o u s e U n - A m e r i c a n A c t i v i t i e s C o m m i t t e e d e n o u n c e d t h e o r e t i c a l p h y s i c i s t E d w a r d C o n d o n , t h e n d i r e c t o r o f
t h e N a t i o n a l B u r e a u o f S t a n d a r d s , a s ‘ ‘ o n e o f t h e w e a k e s t l i n k s i n o u r a t o m i c s e c u r i t y . ”
P r i v a t e l y , t h e s e c u r i t y
1 0 4
e s t a b l i s h m e n t w a s i n t e n s e l y s u s p i c i o u s o f m a n y i n t e r n a t i o n a l i s t s c i e n t i s t s . I n e a r l y 1 9 4 8 , t h e J u s t i c e D e p a r t m e n t
c o n s i d e r e d p r o s e c u t i n g n u c l e a r p h y s i c i s t L e o S z i l a r d u n d e r t h e L o g a n A c t ( w h i c h c r i m i n a l i z e s n e g o t i a t i o n b y
u n a u t h o r i z e d p e r s o n s w i t h f o r e i g n g o v e r n m e n t s h a v i n g a d i s p u t e w i t h t h e U n i t e d S t a t e s ) f o l l o w i n g t h e
p u b l i c a t i o n o f h i s “ L e t t e r t o S t a l i n ” i n t h e
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
. I n t h i s o p e n l e t t e r ( w h i c h w a s n o t
a c t u a l l y s e p a r a t e l y d e l i v e r e d t o t h e S o v i e t U n i o n ) , S z i l a r d c a l l e d f o r i n t e r n a t i o n a l c o n t r o l t h r o u g h t h e U n i t e d
N a t i o n s a n d f o r S t a l i n t o s p e a k d i r e c t l y t o t h e U . S . p e o p l e i n t h i s r e g a r d . I n 1 9 5 0 , t h e F B I e v e n l a u n c h e d a n
i n v e s t i g a t i o n i n t o A l b e r t E i n s t e i n , e v e n t u a l l y a c c u m u l a t i n g o v e r 1 , 5 0 0 p a g e s o f e v i d e n c e o n h i m .
1 0 5
S c i e n t i s t s w e r e o f t e n r e a c t i v e r a t h e r t h a n p r o a c t i v e a n d s o m e t i m e s w e r e o v e r t a k e n b y e v e n t s o r f a i l e d t o r e s p o n d
e f f e c t i v e l y t o t h e m . O t h e r e l i t e s ( s u c h a s G r o v e s , f o r e x a m p l e , o r T r u m a n ) o f t e n d r o v e p o l i c y m a k i n g o r n u d g e d
p u b l i c o p i n i o n i n v a r i o u s d i r e c t i o n s , a n d s c i e n t i s t s c o u l d o n l y r e a c t , o f t e n n o t e f f e c t i v e l y . T w o p r o m i n e n t
e x a m p l e s o f t h i s a r e t h e s p y r e v e l a t i o n s ( a n d s u b s e q u e n t c o n t r o v e r s y ) i n F e b r u a r y / M a r c h 1 9 4 6 a n d t h e
a p p o i n t m e n t o f B e r n a r d B a r u c h a s t h e U . S . r e p r e s e n t a t i v e o f t h e U N A E C w i t h a u t h o r i t y t o s h a p e i n t e r n a t i o n a l
c o n t r o l p r o p o s a l s t o h i s o w n l i k i n g . I n b o t h c a s e s , s c i e n t i s t s w e r e u n h a p p y w i t h t h e t u r n o f e v e n t s , w e r e c a u g h t
u n p r e p a r e d , a n d w e r e u n a b l e t o d e v e l o p e f f e c t i v e r e s p o n s e s .
S c i e n t i s t s ’ i n a b i l i t y t o r e a c t e f f e c t i v e l y w a s d u e t o
1 0 6
t h e i r d i s t a n c e f r o m t h e p o w e r s o f d e c i s i o n - m a k i n g . I n b o t h c a s e s , T r u m a n a n d t h o s e d i r e c t l y a r o u n d h i m
d e t e r m i n e d p o l i c y ; s c i e n t i s t s w e r e n o t p a r t o f t h e i n n e r c i r c l e .
1 0 3
A a s e r u d , “ T h e S c i e n t i s t a n d t h e S t a t e s m e n ” .
1 0 4
D a v i d K a i s e r , “ T h e A t o m i c S e c r e t i n R e d H a n d s ? A m e r i c a n S u s p i c i o n s o f T h e o r e t i c a l P h y s i c i s t s D u r i n g t h e E a r l y C o l d
W a r ” ,
R e p r e s e n t a t i o n s
9 0 , 1 ( S p r i n g 2 0 0 5 ) , p p 2 8 - 6 0 .
1 0 5
W i t t n e r ,
T h e S t r u g g l e A g a i n s t t h e B o m b
v o l u m e 1 , p p . 2 6 7 - 8 ; H e l e n S . H a w k i n s e t a l ( e d s . ) ,
T o w a r d a L i v a b l e W o r l d :
L e o S z i l a r d a n d t h e C r u s a d e f o r N u c l e a r A r m s C o n t r o l
( C a m b r i d g e , M A : T h e M I T P r e s s , 1 9 8 7 ) , p . X l ; L e o S z i l a r d ,
“ C a l l i n g f o r a C r u s a d e ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
( M a y 1 9 4 7 ) , p p . 1 0 2 - 6 , 1 2 5 . E i n s t e i n w a s p a r t i c u l a r l y f o r w a r d i n
w r i t i n g t o S o v i e t s c i e n t i s t s o n a r m s c o n t r o l , s e e C r a i g a n d R a d c h e n k o ,
T h e A t o m i c B o m b a n d t h e O r i g i n s o f t h e C o l d W a r
,
p p . 1 4 7 - 8 .
1 0 6
H e r k e n , p p . 1 2 9 - 3 6 , 1 5 9 - 6 3 . K i m b a l l - S m i t h ,
A P e r i l a n d a H o p e ,
p p . 3 8 6 - 8 , 4 6 4 .
3 0
4 . 7 M u d d l e d P o l i c y m a k i n g
L e s s o n s
P o l i c y m a k i n g i n v o l v e s s i g n i f i c a n t m u d d l i n g t h r o u g h r a t h e r t h a n g r a n d s t r a t e g y .
I t i s a l s o d e e p l y a f f e c t e d b y
1 0 7
d o m e s t i c p o l i t i c s a n d p u b l i c o p i n i o n , a n d o f t e n d e v e l o p e d o n t h e b a s i s o f s h o r t - t e r m o b j e c t i v e s a n d p o o r
q u a l i t y i n f o r m a t i o n . P r o p o s a l s a r e s o m e t i m e s s l o w t o b e d e v e l o p e d , p o o r l y t h o u g h t - o u t o r e x p r e s s e d , g a m b l e d
o n t e c h n i c a l s o l u t i o n s , a n d l a c k i n g c r u c i a l d e t a i l s . T h e r e c a n b e a l a c k o f c l a r i t y o n r e s p o n s i b i l i t y f o r
p o l i c y m a k i n g , w h i c h i s o f t e n d e p e n d e n t o n p e r s o n a l i t i e s .
H i s t o r i c a l C a s e
I n r e l a t i o n t o t h e i n t e r n a t i o n a l c o n t r o l o f a t o m i c e n e r g y , d e b a t e s w e r e o f t e n u n c l e a r a b o u t w h a t w a s b e i n g
d i s c u s s e d a n d p r o p o s e d a n d w h a t w a s a t s t a k e . D e b a t e s w e r e h u g e l y s h a p e d b y d o m e s t i c p o l i t i c s , l a c k o f
i n f o r m a t i o n , v a g u e u n d e r s t a n d i n g s , r e t i r e m e n t s , p r i v a t e i n i t i a t i v e s , c o m m i t t e e r o o m d e a l i n g s , e g o s , a n d
o r g a n i z a t i o n a l i n t e r e s t s . D e b a t e s a b o u t i n t e r n a t i o n a l c o n t r o l w e r e a l s o s o m e t i m e s i n t e r t w i n e d w i t h d e b a t e s o v e r
d o m e s t i c p o l i c y .
S e c r e c y a n d L a c k o f I n f o r m a t i o n
A s n o t e d i n
l e s s o n 4 . 4
, s e c r e c y , l a c k o f i n f o r m a t i o n , a n d m i s c o n c e p t i o n s a b o u t t h e d e v e l o p m e n t o f a t o m i c
w e a p o n s s h a p e d p o l i c y a n d n e g o t i a t i o n s . N o o n e i n g o v e r n m e n t , n o t e v e n T r u m a n , h a d a c l e a r i d e a o f t h e
n u m b e r o f a t o m i c b o m b s i n U . S . p o s s e s s i o n i n 1 9 4 6 ( T r u m a n w o u l d b e v i s i b l y s h o c k e d w h e n L i l i e n t h a l
r e v e a l e d i n a 1 9 4 7 i n v e n t o r y e x e r c i s e h o w f e w t h e U . S . p o s s e s s e d ) .
E v e n t h e U . S . m i l i t a r y m a d e w a r p l a n s
1 0 8
w i t h o u t u n d e r s t a n d i n g t h e n u m b e r a n d d e l i v e r a b i l i t y o f U . S . a t o m i c w e a p o n s .
C u r t i s L e M a y ( t h e n d e p u t y
1 0 9
c h i e f o f A i r S t a f f f o r R e s e a r c h & D e v e l o p m e n t , a n d t h e f u t u r e c o m m a n d e r o f t h e A i r F o r c e ’ s S t r a t e g i c A i r
C o m m a n d ) c o m p l a i n e d i n 1 9 4 6 t h a t t h e A i r F o r c e s t r u g g l e d t o p l a n f o r a t o m i c b o m b d e l i v e r y b e c a u s e o f
s e c r e c y s u r r o u n d i n g t h e n u m b e r a n d n a t u r e o f t h e U . S . b o m b s t o c k .
I n d e e d , o n e J a n u a r y 1 9 4 6 p l a n b y t h e
1 1 0
J o i n t W a r P l a n s C o m m i t t e e w o r k e d o f f t h e a s s u m p t i o n t h a t t h e A i r C o r p s h a d a c c e s s t o 1 9 6 b o m b s , w h i l e t h e
a c t u a l n u m b e r f o r t h e U . S . a s a w h o l e w a s a r o u n d 7 .
I n t e r n a t i o n a l c o n t r o l p o l i c y a n d n e g o t i a t i o n s w e r e
1 1 1
l a r g e l y d e v e l o p e d u n d e r t h e g e n e r a l i m p r e s s i o n t h a t t h e S o v i e t s w e r e m a n y y e a r s a w a y f r o m c o m p l e t i n g t h e i r
f i r s t a t o m i c b o m b .
A n i m p o r t a n t w a r t i m e a g r e e m e n t b e t w e e n R o o s e v e l t a n d C h u r c h i l l t o c o o p e r a t e o n
1 1 2
1 0 7
C h a r l e s E . L i n d b l o m , “ T h e S c i e n c e o f ‘ M u d d l i n g T h r o u g h ’ ” ,
P u b l i c A d m i n i s t r a t i o n R e v i e w
1 9 , 2 ( S p r i n g , 1 9 5 9 ) , p p .
7 9 - 8 8 .
1 0 8
C u r a t o l a ,
B i g g e r B o m b s f o r a B r i g h t e r T o m o r r o w
, p p . 5 2 - 3 ; D a v i d A l a n R o s e n b e r g , “ U . S . N u c l e a r S t o c k p i l e , 1 9 4 5 t o
1 9 5 0 ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
3 8 , 5 ( M a y 1 9 8 2 ) , p p . 2 5 - 3 0 ; D a v i d A l a n R o s e n b e r g , “ T h e O r i g i n s o f O v e r k i l l :
N u c l e a r W e a p o n s a n d A m e r i c a n S t r a t e g y , 1 9 4 5 - 1 9 6 0 ” ,
I n t e r n a t i o n a l S e c u r i t y
7 , 4 ( S p r i n g 1 9 8 3 ) , p p . 3 - 7 1 ; H e r k e n , p . 1 9 7 .
1 0 9
H e r k e n , p p . 2 1 9 - 2 0 , 2 2 9 ; R o s e n b e r g , “ T h e O r i g i n s o f O v e r k i l l ” ; R o s s ,
A m e r i c a n W a r P l a n s 1 9 4 5 - 1 9 5 0
, p p . 1 2 - 1 7 .
1 1 0
H e r k e n , p . 1 9 8 .
1 1 1
H e r k e n , p p . 1 9 9 - 2 0 0 , 2 1 9 - 2 0 , 2 2 9 ; R o s s ,
A m e r i c a n W a r P l a n s 1 9 4 5 - 1 9 5 0
, p p . 1 2 - 1 7 ; Z i e g l e r , “ I n t e l l i g e n c e A s s e s s m e n t s
o f S o v i e t A t o m i c C a p a b i l i t y ” . T h e J a n u a r y 1 9 4 6 p l a n i s i n R o s s , A m e r i c a n W a r P l a n s , p p . 1 6 - 1 7 . O n t h e n u m b e r o f a t o m i c
b o m b s s e e J o h n M . C u r a t o l a ,
B i g g e r B o m b s f o r a B r i g h t e r T o m o r r o w : S t r a t e g i c A i r C o m m a n d a n d A m e r i c a n W a r P l a n s a t
t h e D a w n o f t h e A t o m i c A g e , 1 9 4 5 - 1 9 5 0
( J e f f e r s o n , N C : M a c F a r l a n d , 2 0 1 6 ) , p p . 5 3 .
1 1 2
H e r k e n , p p . 1 1 3 ; G o r d i n ,
R e d C l o u d a t D a w n ,
p p . 7 5 - 8 . O n t h e C I A ’ s m i s t a k e n a s s e s s m e n t s s e e D o n a l d P . S t e u r y ,
“ H o w t h e C I A M i s s e d S t a l i n ' s B o m b : D i s s e c t i n g S o v i e t A n a l y s i s , 1 9 4 6 - 5 0 ” . A v a i l a b l e a t :
h t t p s : / / w w w . c i a . g o v / l i b r a r y / c e n t e r - f o r - t h e - s t u d y - o f - i n t e l l i g e n c e / c s i - p u b l i c a t i o n s / c s i - s t u d i e s / s t u d i e s / v o l 4 9 n o 1 / h t m l _ f i l e s
/ s t a l i n s _ b o m b _ 3 . h t m l
, a c c e s s e d 1 J u n e 2 0 1 9 .
3 1
a t o m i c d e v e l o p m e n t w a s k e p t s e c r e t f r o m m o s t p o l i c y m a k e r s , i n c l u d i n g T r u m a n , w h o l a t e r i g n o r e d i t w h e n h e
f o u n d o u t .
1 1 3
P r o p o s a l s w e r e S l o w t o b e D e v e l o p e d , P o o r l y T h o u g h t - o u t o r E x p r e s s e d , G a m b l e d o n T e c h n i c a l
S o l u t i o n s , a n d L a c k e d C r u c i a l D e t a i l s
S e r i o u s t h i n k i n g a b o u t p r o p o s a l s f o r i n t e r n a t i o n a l c o n t r o l w e r e n o t f o r m u l a t e d u n t i l r e l a t i v e l y l a t e . T h i s i s d u e
i n l a r g e p a r t t o t h e e x i s t e n c e o f n u c l e a r w e a p o n s b e i n g a c l o s e l y h e l d s e c r e t u n t i l t h e d e t o n a t i o n o f t h e f i r s t
a t o m i c b o m b a t H i r o s h i m a i n A u g u s t 1 9 4 5 . E f f o r t s t o m a k e s e n s e o f n u c l e a r w e a p o n s o n l y d r a m a t i c a l l y
a c c e l e r a t e d a f t e r t h e b o m b ’ s e x i s t e n c e w a s m a d e p u b l i c .
1 1 4
S e c r e t a r y o f W a r H e n r y L . S t i m s o n c r i s p l y s t a t e d t h e h i g h - l e v e l p r o b l e m o f p o s t w a r g o v e r n a n c e o f n u c l e a r
w e a p o n s i n A p r i l 1 9 4 5 , b u t l i t t l e s e r i o u s t h o u g h t a b o u t i n t e r n a t i o n a l c o n t r o l t o o k p l a c e a f t e r t h e n .
T h i s
1 1 5
n e g l e c t c a n b e i n p a r t u n d e r s t o o d , b e c a u s e p o l i c y m a k e r s w e r e s w a m p e d w i t h m a s s i v e g e o p o l i t i c a l p r o b l e m s :
e n d i n g t h e w a r w i t h J a p a n , n e g o t i a t i n g a p o s t w a r o r d e r w i t h R u s s i a , s e t t i n g u p t h e U n i t e d N a t i o n s , a n d
r e b u i l d i n g E u r o p e .
S e c r e t a r y o f W a r S t i m s o n ’ s i d e a s i n l a t e 1 9 4 4 a n d e a r l y 1 9 4 5 o n i n t e r n a t i o n a l c o n t r o l a n d p o s s i b l e q u i d p r o
q u o s w e r e c u r s o r y . M o s t p r o p o s a l s f o r i n t e r n a t i o n a l c o n t r o l l a c k e d d e t a i l a n d f a i l e d t o t h i n k t h r o u g h t h e S o v i e t
r e s p o n s e . T h i s i s p a r t i c u l a r l y t r u e o f t h e e a r l i e s t t h i n k i n g i n 1 9 4 4 a n d 1 9 4 5 . S t i m s o n t h o u g h t t h a t i n t e r n a t i o n a l
c o n t r o l c o u l d b e a c h i e v e d t h r o u g h “ f r e e d o m b o t h o f s c i e n c e a n d a c c e s s ” t o a t o m i c i n f o r m a t i o n . B y t h i s h e
m e a n t s o m e f o r m o f s h a r i n g o f a t o m i c s c i e n t i f i c i n f o r m a t i o n , b u t h e d i d n o t g o i n t o f u r t h e r d e t a i l . S t i m s o n a l s o
t h o u g h t t h a t t h e U . S . c o u l d d e m a n d l i b e r a l i z a t i o n o f i n t e r n a l r u l e i n t h e S o v i e t U n i o n i n r e t u r n f o r i n f o r m a t i o n
o n t h e a t o m i c b o m b , w h i c h r e f l e c t e d a c o m p l e t e m i s u n d e r s t a n d i n g o f S o v i e t p r e f e r e n c e s . S i m i l a r l y h i s l a t e r i d e a
o f a s i m p l e “ c o v e n a n t ” b e t w e e n t h e g r e a t p o w e r s t o n o t u s e a t o m i c e n e r g y f o r m i l i t a r y p u r p o s e s , w i t h o u t a n y
t h o u g h t o f s a f e g u a r d s o r p u n i s h m e n t s , w a s i l l t h o u g h t - o u t .
S t i m s o n ’ s m e m o r a n d u m o n i n t e r n a t i o n a l c o n t r o l ,
1 1 6
p r e s e n t e d a t a c r u c i a l S e p t e m b e r 1 9 4 5 m e e t i n g o n a t o m i c e n e r g y , g e n e r a t e d s i g n i f i c a n t c o n f u s i o n o v e r w h a t h e
w a s p r o p o s i n g . S o m e p a r t i c i p a n t s t h o u g h t h e w a n t e d t o g i v e t h e a t o m i c b o m b t o t h e S o v i e t s . T h i s w a s d u e t o a n
u n f a m i l i a r i t y o n t h e i r p a r t w i t h t h e d e v e l o p m e n t o f t h e b o m b a n d i t s i n t e r n a t i o n a l p o l i t i c s , a s w e l l a s a l a c k o f
c l a r i t y o n S t i m s o n ’ s p a r t .
1 1 7
1 1 3
B a r t o n J . B e r n s t e i n , “ R o o s e v e l t , T r u m a n , a n d t h e A t o m i c B o m b , 1 9 4 1 - 1 9 4 5 : A R e i n t e r p r e t a t i o n ” ,
P o l i t i c a l S c i e n c e
Q u a r t e r l y
9 0 , 1 ( S p r i n g 1 9 7 5 ) , p p . 2 3 - 6 9 ; H e r k e n , p . 6 2 .
1 1 4
A s , f o r e x a m p l e , i s i l l u s t r a t e d b y B e r n a r d B r o d i e r e c o u n t i n g t h e m o m e n t w h e n h i s c a r e e r p i v o t e d t o t h e s t u d y o f n u c l e a r
w e a p o n s . F r e d K a p l a n ,
T h e W i z a r d s o f A r m a g e d d o n
2 n d e d i t i o n ( S t a n f o r d : S t a n f o r d U n i v e r s i t y P r e s s , 1 9 9 1 ) , p . 1 0 .
1 1 5
H e n r y S t i m s o n ,
M e m o r a n d u m D i s c u s s e d w i t h t h e P r e s i d e n t
( 2 5 A p r i l 1 9 4 5 ) . A v a i l a b l e a t :
h t t p : / / w w w . n u c l e a r f i l e s . o r g / m e n u / l i b r a r y / c o r r e s p o n d e n c e / s t i m s o n - h e n r y / c o r r _ s t i m s o n _ 1 9 4 5 - 0 4 - 2 5 . h t m
, a c c e s s e d 1 5
O c t o b e r 2 0 1 9 . T h i s m e m o r a n d u m w a s d i s c u s s e d i n t h e M a y 3 1 m e e t i n g o f t h e I n t e r i m C o m m i t t e e , s e e
N o t e s o f t h e I n t e r i m
C o m m i t t e e M e e t i n g , T h u r s d a y 3 1 M a y 1 9 4 5
( 3 1 M a y 1 9 4 5 ) . A v a i l a b l e a t :
h t t p : / / w w w . n u c l e a r f i l e s . o r g / m e n u / k e y - i s s u e s / n u c l e a r - w e a p o n s / h i s t o r y / p r e - c o l d - w a r / i n t e r i m - c o m m i t t e e / i n t e r i m - c o m m i
t t e e - i n f o r m a l - n o t e s _ 1 9 4 5 - 0 5 - 3 1 . h t m
, a c c e s s e d 1 5 O c t o b e r 2 0 1 9 .
h / t L u k e M u e h l h a u s e r , w h o r e f l e c t s o n t h i s :
h t t p s : / / t w i t t e r . c o m / l u k e p r o g / s t a t u s / 1 1 8 1 7 7 4 8 7 0 0 9 6 4 0 0 3 8 4
1 1 6
H e r k e n , p p . 1 4 , 2 5 ; M a l l o y ,
A t o m i c T r a g e d y
, p p . 1 7 3 - 4 .
1 1 7
H e n r y S t i m s o n ,
M e m o r a n d u m o n t h e E f f e c t s o f t h e A t o m i c B o m b
( 1 1 S e p t e m b e r 1 9 4 5 ) . A v a i l a b l e a t :
h t t p : / / w w w . n u c l e a r f i l e s . o r g / m e n u / l i b r a r y / c o r r e s p o n d e n c e / s t i m s o n - h e n r y / c o r r _ s t i m s o n _ 1 9 4 5 - 0 9 - 1 1 . h t m
, a c c e s s e d 1 6
O c t o b e r 2 0 1 9 . H e r k e n , p . 3 0 ; L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a
, p p . 1 4 5 - 6 .
3 2
T h e i n t e r n a t i o n a l c o n t r o l o f t e c h n o l o g y i s m a d e e a s i e r w h e n t h e d a n g e r o u s u s e s a n d p r o d u c t i v e u s e s c a n e a c h b e
c l e a n l y , v e r i f i a b l y , a n d r o b u s t l y s e p a r a t e d f r o m e a c h o t h e r . I n t h i s c a s e , t h e r e w a s o n e t e c h n i c a l p o s s i b i l i t y t h a t
t h e m o s t t h o u g h t - t h r o u g h p r o p o s a l , t h e A c h e s o n - L i l i e n t h a l P l a n , i n v e s t e d a l o t o f h o p e i n , b u t w h i c h i n
r e t r o s p e c t w a s m i s g u i d e d . ( T h i s w a s p l a u s i b l y f o r e s e e a b l e a t t h e t i m e ) . T h e p l a n n e r s a l l o w e d t h e m s e l v e s t o
d e p e n d o n a s p e c u l a t i v e t e c h n i c a l f i x , c a l l e d i s o t o p i c d e n a t u r i n g . I t w a s h o p e d t h i s c o u l d b e u s e d t o p r e v e n t ( t h e
e a s y ) m i s u s e o f u r a n i u m o r p l u t o n i u m r e a c t o r f u e l f o r n u c l e a r w e a p o n s . E f f e c t i v e d e n a t u r i n g w a s , h o w e v e r , a
h i g h l y s p e c u l a t i v e c o n c e p t , a n d h i s t o r y h a s n o t b o r n e i t o u t .
T h e r e p o r t w a s c r i t i c i z e d b y o t h e r s c i e n t i s t s ( a n d
1 1 8
b y L e s l i e G r o v e s ) f o r r e l y i n g o n s u c h a n i m p r a c t i c a l i d e a s o o n a f t e r t h e p l a n w a s m a d e p u b l i c .
1 1 9
T h e p l a n a l s o d i d n o t e x p l a i n w h a t w a s t o b e d o n e w i t h t h e e x i s t i n g a t o m i c b o m b s o n c e i n t e r n a t i o n a l c o n t r o l
w a s i n s t i t u t e d : w e r e t h e y t o b e d e s t r o y e d , o r p a s s e d o n t o t h e U . N . ? I n t h i s c a s e , t h e v a g u e n e s s w a s p r o b a b l y
d e l i b e r a t e a n d r e f l e c t e d t h e p o l i t i c a l r a m i f i c a t i o n s o f c h o o s i n g b e t w e e n t h e s e o p t i o n s : t h e r e w e r e m a n y c a l l i n g
f o r t h e c o m p l e t e d e s t r u c t i o n o f a t o m i c b o m b s a n d y e t o t h e r s f o r t h e m t o b e k e p t a n d u s e d b y t h e U . N . f o r
c o l l e c t i v e s e c u r i t y o r p o l i c i n g p u r p o s e s .
1 2 0
L a c k o f C l a r i t y o n R e s p o n s i b i l i t y f o r P o l i c y m a k i n g
R e s p o n s i b i l i t y f o r p o l i c y m a k i n g o n a t o m i c m a t t e r s w a s n o t c l e a r l y d e m a r c a t e d , e s p e c i a l l y i n l a t e 1 9 4 5 ( i t w o u l d
i m p r o v e l a t e r o n w i t h t h e A c h e s o n - L i l i e n t h a l c o m m i t t e e , t h o u g h t h e c o m m i t t e e i t s e l f w a s a g a m b i t f o r S t a t e
D e p a r t m e n t i n f l u e n c e ) . C o n g r e s s i o n a l c o m m i t t e e s , c o n g r e s s m e n , S t a t e D e p a r t m e n t o f f i c i a l s , m i l i t a r y m e n ,
s c i e n t i f i c a d v i s o r s , e x t e r n a l c o n s u l t a n t s , a n d o t h e r a d v i s o r s a l l h a d i n f l u e n c e a t o n e p o i n t o r a n o t h e r . S c i e n t i f i c
a d m i n i s t r a t o r V a n n e v a r B u s h w o u l d w r i t e t o S t i m s o n i n N o v e m b e r 1 9 4 5 e x p r e s s i n g f r u s t r a t i o n a t t h e l a c k o f
1 1 8
B e l o w i s a s e l e c t i o n f r o m t h e A c h e s o n - L i l i e n t h a l R e p o r t , e m p h a s i s o u r s . N o t e t h a t t h e a c t u a l t e x t a d m i t s t h a t t h i s
t e c h n i q u e i s n o t f o o l p r o o f .
“ U 2 3 5 a n d p l u t o n i u m c a n b e d e n a t u r e d ; s u c h d e n a t u r e d m a t e r i a l s d o n o t
r e a d i l y
l e n d t h e m s e l v e s t o t h e m a k i n g o f
a t o m i c e x p l o s i v e s , b u t t h e y c a n s t i l l b e u s e d w i t h n o e s s e n t i a l l o s s o f e f f e c t i v e n e s s f o r t h e p e a c e f u l a p p l i c a t i o n s o f a t o m i c
e n e r g y . … I t i s i m p o r t a n t t o u n d e r s t a n d t h e s e n s e i n w h i c h d e n a t u r i n g r e n d e r s m a t e r i a l s a f e r . I n t h e f i r s t p l a c e , i t w i l l m a k e
t h e m a t e r i a l u n u s e a b l e b y a n y m e t h o d s
w e n o w k n o w
f o r e f f e c t i v e a t o m i c e x p l o s i v e s
u n l e s s s t e p s a r e t a k e n t o r e m o v e
t h e d e n a t u r a n t s .
I n t h e s e c o n d p l a c e , t h e d e v e l o p m e n t o f m o r e i n g e n i o u s m e t h o d s i n t h e f i e l d o f a t o m i c e x p l o s i v e s
w h i c h m i g h t m a k e t h i s m a t e r i a l e f f e c t i v e l y u s e a b l e i s n o t o n l y d u b i o u s , b u t i s c e r t a i n l y n o t p o s s i b l e w i t h o u t a v e r y m a j o r
s c i e n t i f i c a n d t e c h n i c a l e f f o r t .
I t i s p o s s i b l e , b o t h f o r U 2 3 5 a n d f o r p l u t o n i u m , t o r e m o v e t h e d e n a t u r a n t , b u t d o i n g s o c a l l s f o r r a t h e r c o m p l e x
i n s t a l l a t i o n s w h i c h , t h o u g h n o t o f t h e s c a l e o f t h o s e a t O a k R i d g e o r H a n f o r d , n e v e r t h e l e s s w i l l r e q u i r e a l a r g e e f f o r t a n d ,
a b o v e a l l , s c i e n t i f i c a n d e n g i n e e r i n g s k i l l o f a n a p p r e c i a b l e o r d e r f o r t h e i r d e v e l o p m e n t . I t i s n o t w i t h o u t i m p o r t a n c e t o b e a r
i n m i n d t h a t , a l t h o u g h a s t h e a r t n o w s t a n d s d e n a t u r e d m a t e r i a l s a r e u n s u i t a b l e f o r b o m b m a n u f a c t u r e , d e v e l o p m e n t s
w h i c h d o n o t a p p e a r t o b e i n p r i n c i p l e i m p o s s i b l e m i g h t a l t e r t h e s i t u a t i o n . T h i s i s a g o o d e x a m p l e o f t h e n e e d f o r c o n s t a n t
r e c o n s i d e r a t i o n o f t h e d i v i d i n g l i n e b e t w e e n w h a t i s s a f e a n d w h a t i s d a n g e r o u s . ” B a r n a r d e t a l ,
A R e p o r t o n t h e
I n t e r n a t i o n a l C o n t r o l o f A t o m i c E n e r g y
.
1 1 9
I s o t o p i c d e n a t u r i n g m e a n t t h e a d d i t i o n o f a d i f f e r e n t i s o t o p e o f n u c l e a r f u e l w h i c h w o u l d r e n d e r t h e f u e l u s e l e s s a s a n
e x p l o s i v e . T h e f u e l h o w e v e r c o u l d s t i l l b e u s e d f o r p o w e r r e a c t o r s . S u c h a n i s o t o p e , i t w a s t h o u g h t , c o u l d n o t b e c h e m i c a l l y
s e p a r a t e d . B a r n a r d e t a l ,
A R e p o r t o n t h e I n t e r n a t i o n a l C o n t r o l o f A t o m i c E n e r g y
, p p . 2 6 - 7 . A n t h o n y L e v i e r o , “ D e n a t u r i n g
G u a r d o n A t o m N o t S u r e ” ,
N e w Y o r k T i m e s
( 1 0 A p r i l 1 9 4 6 ) , p . 1 6 . K i m b a l l - S m i t h ,
A P e r i l a n d a H o p e
, p p . 4 6 2 - 3 . F o r
m o r e r e c e n t t h i n k i n g o n d e n a t u r i n g s e e : A . D e V o l p i , “ D e n a t u r i n g F i s s i l e M a t e r i a l s ” ,
P r o g r e s s i n N u c l e a r E n e r g y
1 0 , 2
( 1 9 8 2 ) , p p . 1 6 1 - 2 2 0 . C . E . T i l l , “ D e n a t u r e d F u e l C y c l e s ” , i n J o s e p h L . F o w l e r , C l e l a n d H . J o h n s o n , C h a r l e s D . B o w m a n
( e d s . ) ,
N u c l e a r C r o s s S e c t i o n s f o r T e c h n o l o g y : P r o c e e d i n g s o f t h e I n t e r n a t i o n a l C o n f e r e n c e
v o l 1 3 ( O a k R i d g e , T N : O a k R i d g e
N a t i o n a l L a b o r a t o r y , 1 9 8 0 ) , p . 1 1 5 - 1 8 .
1 2 0
B a r n a r d e t a l ,
A R e p o r t o n t h e I n t e r n a t i o n a l C o n t r o l o f A t o m i c E n e r g y
.
3 3
c l e a r l y d e m a r c a t e d r e s p o n s i b i l i t i e s f o r a t o m i c p o l i c y m a k i n g a n d a d v i c e : “ I h a v e n e v e r p a r t i c i p a t e d i n a n y t h i n g
t h a t w a s s o c o m p l e t e l y u n o r g a n i z e d o r s o i r r e g u l a r . . . I t i s s o m e w h a t a p p a l l i n g . . . t o t h i n k o f t h i s c o u n t r y h a n d l i n g
m a n y m a t t e r s i n s u c h a n a t m o s p h e r e . ”
M a n y , s u c h a s G r o v e s i n l a t e 1 9 4 5 , b e c a m e p o w e r f u l , b u t s t i l l
1 2 1
u n o f f i c i a l , a d v i s o r s .
1 2 2
B e c a u s e o f t h i s l a c k o f c l a r i t y , i n d i v i d u a l s o f t e n c o m p e t e d a g a i n s t e a c h o t h e r t o i n f l u e n c e p o l i c y m a k i n g .
V a n n e v a r B u s h , f o r e x a m p l e , s u g g e s t e d a n e w c o m m i t t e e , i n c l u d i n g t h e S t a t e D e p a r t m e n t , s c i e n t i s t s , a n d
r e p r e s e n t a t i v e s o f C o n g r e s s , t o m a k e a t o m i c p o l i c y i n N o v e m b e r 1 9 4 5 b u t w a s r e b u f f e d .
S e c r e t a r y o f S t a t e
1 2 3
B y r n e s f o u g h t w i t h S e n a t o r A r t h u r H . V a n d e n b e r g a n d o t h e r s t o r e t a i n s o l e c o n t r o l o v e r a t o m i c d i p l o m a c y i n
l a t e 1 9 4 5 .
O n e o f t h e r e a s o n s B y r n e s s e t u p a c o m m i t t e e t o m a k e a p r o p o s a l f o r i n t e r n a t i o n a l c o n t r o l i n
1 2 4
D e c e m b e r 1 9 4 5 / J a n u a r y 1 9 4 6 w a s t o k e e p a t o m i c d i p l o m a c y w i t h i n t h e S t a t e D e p a r t m e n t , i n s t e a d o f l o s i n g i t
t o C o n g r e s s i o n a l c o m m i t t e e s .
T h e A c h e s o n - L i l i e n t h a l R e p o r t w a s p r o d u c e d l a r g e l y b y i t s t e c h n i c a l b o a r d o f
1 2 5
c o n s u l t a n t s . L e s l i e G r o v e s h a d o p p o s e d t h e a p p o i n t m e n t o f t h i s b o a r d , a n d h a d a t t e m p t e d t o g e t a b o a r d t o h i s
l i k i n g a p p o i n t e d i n s t e a d . S i m i l a r l y , t h e S t a t e D e p a r t m e n t l a t e r o p p o s e d B e r n a r d B a r u c h ’ s b o a r d o f c o n s u l t a n t s
( w h i c h e v e n t u a l l y h e l p e d f o r m u l a t e t h e B a r u c h P l a n ) a n d a l s o a t t e m p t e d t o g e t i t s o w n b o a r d a p p o i n t e d .
1 2 6
I n t e r s e r v i c e r i v a l r y h a d a l a r g e i m p a c t o n d e b a t e s a n d p o l i c i e s . E a c h m i l i t a r y s e r v i c e a t t e m p t e d t o i n c r e a s e i t s
i n f l u e n c e o n a t o m i c w e a p o n s a n d a t o m i c p o l i c y , a n d d e c r e a s e t h a t o f t h e o t h e r s e r v i c e s .
P e r h a p s t h e m o s t
1 2 7
n o t o r i o u s e p i s o d e i n t h i s r e g a r d w a s t h e i n f a m o u s “ A d m i r a l s ’ R e v o l t ” o f 1 9 4 9 , i n w h i c h t h e n a v y a t t e m p t e d t o
d i v e r t f u n d i n g a w a y f r o m t h e a i r f o r c e a n d t o w a r d s t h e n a v y . I n a s e r i e s o f h e a r i n g s b e f o r e t h e H o u s e A r m e d
S e r v i c e s C o m m i t t e e , s e n i o r n a v a l o f f i c e r s a r g u e d t h a t t h e a i r f o r c e ’ s s t r a t e g y o f “ a t o m i c b l i t z ” c o u l d n o t w i n a
w a r a g a i n s t t h e S o v i e t U n i o n a n d w a s a n y h o w i m m o r a l . T h e n a v y i n s t e a d s u g g e s t e d ( u l t i m a t e l y u n s u c c e s s f u l l y )
t h a t a i r c r a f t c a r r i e r - b a s e d b o m b e r s l a u n c h t a c t i c a l a t o m i c s t r i k e s a g a i n s t t h e i n v a d i n g S o v i e t a r m i e s .
1 2 8
P u b l i c O p i n i o n a n d t h e A t o m i c S e c r e t
P u b l i c o p i n i o n o n a t o m i c m a t t e r s a n d i n t e r n a t i o n a l c o n t r o l ( w h i c h h a d s i g n i f i c a n t i n f l u e n c e o n p o l i c y m a k i n g ,
s e e e a r l i e r
s e c t i o n
) w a s b a s e d o n o n e c r u c i a l m i s c o n c e p t i o n : t h a t t h e r e e x i s t e d a n “ a t o m i c s e c r e t ” w h i c h w a s t h e
k e y t o t h e p r o d u c t i o n o f t h e b o m b . I t w a s w i d e l y b e l i e v e d t h a t i f t h e S o v i e t s d i s c o v e r e d t h i s s e c r e t , t h e y c o u l d
b u i l d t h e a t o m i c b o m b . T h e r e w a s s i g n i f i c a n t d e b a t e a n d c o n f u s i o n i n p u b l i c d i s c o u r s e o v e r w h a t t h i s s e c r e t
w a s a n d h o w i t c o u l d b e p r o t e c t e d o r s h a r e d .
S c i e n t i s t s s p e n t a s i g n i f i c a n t a m o u n t o f e f f o r t d e b u n k i n g t h e
1 2 9
i d e a t h a t s u c h a n a t o m i c s e c r e t e x i s t e d . T h e y p o i n t e d o u t t h a t t h e b a s i c s c i e n c e w a s a l r e a d y i n t h e p u b l i c
1 2 1
Z a c h a r y ,
E n d l e s s F r o n t i e r
, p . 3 1 3 .
1 2 2
H e r k e n , p p . 3 7 - 8 .
1 2 3
Z a c h a r y ,
E n d l e s s F r o n t i e r
, p . 3 1 0 .
1 2 4
H e r k e n , p p . 7 4 - 7 , 8 2 , 8 5 - 6 , 9 3 . L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a
, p p . 2 1 8 - 2 2 .
1 2 5
H e r k e n , p . 9 7 ; L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a
, p . 2 3 6 .
1 2 6
H e r k e n , p p . 1 5 4 , 1 6 9 .
1 2 7
H e r k e n , p . 1 9 8 - 9 , 2 5 2 - 3 . O n e m a n i f e s t a t i o n o f t h i s w a s t h e 1 9 4 6 a t o m i c b o m b t e s t s ; s e e : L l o y d J . G r a y b a r , “ T h e 1 9 4 6
A t o m i c B o m b T e s t s : A t o m i c D i p l o m a c y o r B u r e a u c r a t i c I n f i g h t i n g ” ,
T h e J o u r n a l o f A m e r i c a n H i s t o r y
7 2 , 4 ( M a r c h 1 9 8 6 ) ,
p p . 8 8 8 - 9 0 7 .
1 2 8
P h i l l i p S . M e i l i n g e r , “ T h e A d m i r a l s ’ R e v o l t o f 1 9 4 9 : L e s s o n s f o r T o d a y ” ,
P a r a m e t e r s
1 9 , 3 ( S e p t e m b e r 1 9 8 9 ) , p p . 8 1 - 9 6 ;
H e r k e n , p p . 3 0 8 - 9 .
1 2 9
D a v i d K a i s e r , “ T h e A t o m i c S e c r e t i n R e d H a n d s ? A m e r i c a n S u s p i c i o n s o f T h e o r e t i c a l P h y s i c i s t s D u r i n g t h e E a r l y C o l d
W a r ” ,
R e p r e s e n t a t i o n s
9 0 , 1 ( S p r i n g 2 0 0 5 ) , p p 2 8 - 6 0 ; H e r k e n , “ ‘ A M o s t D e a d l y I l l u s i o n ’ ” .
3 4
d o m a i n , a n d s t r e s s e d t h e i m p o r t a n c e o f s c i e n t i f i c a n d i n d u s t r i a l e x p e r t i s e a n d r e s o u r c e s , a s w e l l a s f i n a n c i a l
r e s o u r c e s . T h e i r s t a t e m e n t s h a d l i m i t e d i m p a c t : t h e i d e a o f a n a t o m i c s e c r e t w a s d e e p l y e m b e d d e d a n d w a s
m o r e o v e r u s e d a n d r e p e a t e d i n p u b l i c r h e t o r i c b y o p p o n e n t s o f i n t e r n a t i o n a l c o n t r o l s u c h a s G r o v e s .
1 3 0
P e r s o n a l i t y
H i g h - l e v e l p o l i c y m a k i n g w a s h i g h l y d e p e n d e n t o n p e r s o n a l i t y . T h e w a r t i m e p r e s i d e n t , F r a n k l i n D . R o o s e v e l t ,
h a d a r e p u t a t i o n f o r n o t d i s a g r e e i n g w i t h i n d i v i d u a l s i n f a c e - t o - f a c e m e e t i n g s . H e p r e f e r r e d t o r a i s e o b j e c t i o n s
l a t e r w i t h c o n f i d a n t s .
H i s s u c c e s s o r T r u m a n o n l y d i g e s t e d s h o r t b r i e f i n g s .
T r u m a n h a d o p p o r t u n i t i e s t o
1 3 1
1 3 2
m a k e a g r e a t e r i m p a c t o n a t o m i c p o l i c y a n d d i p l o m a c y b u t d i d n o t , f o r n o c l e a r r e a s o n . A t P o t s d a m , f o r
e x a m p l e , h e d o d g e d t h e i s s u e o f i n f o r m i n g t h e S o v i e t U n i o n c l e a r l y a b o u t t h e U . S . a t o m i c p r o g r a m ( t h e n e x t
o p p o r t u n i t y w o u l d a r i s e a f t e r t h e w a r ) .
R o o s e v e l t , i f h e h a d l i v e d , w o u l d p r o b a b l y h a v e c o n d u c t e d p o l i c y
1 3 3
d i f f e r e n t l y t h a n T r u m a n a n d r e a c t e d d i f f e r e n t l y t o p o l i t i c a l d e v e l o p m e n t s .
1 3 4
G r o v e s , B a r u c h , a n d S e c r e t a r y o f S t a t e B y r n e s ( a m o n g s t o t h e r s ) a l s o p u t t h e i r p e r s o n a l s t a m p o n d i p l o m a c y a n d
p o l i c y m a k i n g . E g o a n d p e r s o n a l c a r e e r m o t i v a t i o n s s h a p e d t h e i r a p p r o a c h e s . C r u c i a l l y , o n e r e a s o n f o r B a r u c h
f o r m u l a t i n g h i s o w n d i s t i n c t p r o p o s a l f o r i n t e r n a t i o n a l c o n t r o l — r a t h e r t h a n f o l l o w i n g t h e e a r l i e r
A c h e s o n - L i l i e n t h a l P l a n — w a s e g o . H e r e f u s e d , h e s a i d , t o b e a m e r e “ m e s s e n g e r b o y ” f o r t h e
A c h e s o n - L i l i e n t h a l P l a n .
T r u m a n h i m s e l f w o u l d l a t e r r e c a l l t h a t B a r u c h w a s d r i v e n b y t h e n e e d f o r “ p u b l i c
1 3 5
r e c o g n i t i o n . ”
O n e o f t h e r e a s o n s w h y O p p e n h e i m e r ’ s r o l e i n p o l i c y m a k i n g i n 1 9 4 6 w a s l i m i t e d t o h i s
1 3 6
p a r t i c i p a t i o n i n t h e d r a f t i n g o f t h e A c h e s o n - L i l i e n t h a l R e p o r t w a s b e c a u s e h e m a d e a p o o r p e r s o n a l i m p r e s s i o n
o n T r u m a n d u r i n g t h e i r f i r s t p r i v a t e m e e t i n g .
O n t h e S o v i e t s i d e , a t o m i c d i p l o m a c y a p p e a r s t o h a v e b e e n
1 3 7
l a r g e l y p e r s o n a l l y d e t e r m i n e d b y S t a l i n , w i t h d i p l o m a t s h a v i n g l i t t l e a u t o n o m y .
1 3 8
M i x e d S i g n a l s
1 3 0
H e r k e n , “ ‘ A M o s t D e a d l y I l l u s i o n ’ ” .
1 3 1
A a s e r u d , “ T h e S c i e n t i s t a n d t h e S t a t e s m e n ” .
1 3 2
H e r k e n , p . 1 5 .
1 3 3
W i l s o n D . M i s c a m b l e ,
T h e M o s t C o n t r o v e r s i a l D e c i s i o n : T r u m a n , t h e A t o m i c b o m b s , a n d t h e D e f e a t o f J a p a n
( C a m b r i d g e : C a m b r i d g e U n i v e r s i t y P r e s s , 2 0 1 1 ) ,
p . 7 0 .
1 3 4
C a m p b e l l C r a i g a n d S e r g e y R a d c h e n k o , f o r e x a m p l e , b e l i e v e t h a t “ T r u m a n a t t h e o u t s e t o f h i s p r e s i d e n c y w a s m o r e
d i s p o s e d t o w a r d c o o p e r a t i o n w i t h t h e S o v i e t U n i o n t h a n R o o s e v e l t h a d b e e n a t t h e e n d o f h i s . ” R o o s e v e l t w a s a l s o , t h e y
s u g g e s t , s t r o n g l y i n f l u e n c e d b y C h u r c h i l l i n h i s a t o m i c d i p l o m a c y . C r a i g a n d R a d c h e n k o , p p . 2 9 , 6 5 . T h e r e i s a l s o s o m e
s p e c u l a t i o n o f t h i s i n B e r n s t e i n , “ R o o s e v e l t , T r u m a n , a n d t h e A t o m i c B o m b ” .
1 3 5
H e r k e n , p . 1 6 0 . F o r a l i s t o f t h e k e y d i f f e r e n c e s b e t w e e n t h e B a r u c h P l a n a n d t h e e a r l i e r A c h e s o n - L i l i e n t h a l R e p o r t , s e e
t h e s u m m a r y o f t h e B a r u c h P l a n i n t h e e a r l i e r
s e c t i o n
w h i c h r e v i e w s p r o p o s a l s f o r i n t e r n a t i o n a l c o n t r o l .
1 3 6
H a r r y S . T r u m a n ,
M e m o i r s
V o l u m e I I Y e a r s o f T r i a l a n d H o p e ( N e w Y o r k : D o u b l e d a y a n d C o m p a n y , 1 9 5 6 ) , p . 1 0 ;
J a m e s G r a n t ,
B e r n a r d M . B a r u c h : T h e A d v e n t u r e s o f a W a l l S t r e e t L e g e n d
( N e w Y o r k : J o h n W i l e y & S o n s , 1 9 9 7 ) , p . 2 9 2 .
1 3 7
T r u m a n w o u l d l a t e r c a l l O p p e n h e i m e r a “ c r y - b a b y s c i e n t i s t ” w h o h a d c o m e t o h i s o f f i c e a n d “ s p e n t m o s t o f h i s t i m e
w r i n g i n g h i s h a n d s a n d t e l l i n g m e t h e y h a d b l o o d o n t h e m b e c a u s e o f h i s d i s c o v e r y o f a t o m i c e n e r g y . ” H e k e p t
O p p e n h e i m e r a w a y f r o m p o l i c y m a k i n g f r o m t h a t p o i n t o n . R a y M o n k ,
R o b e r t O p p e n h e i m e r : A L i f e I n s i d e t h e C e n t e r
( N e w Y o r k : D o u b l e d a y , 2 0 1 2 ) , p . 4 9 4 ; B i r d a n d S h e r w i n ,
A m e r i c a n P r o m e t h e u s
, p p . 3 3 1 - 3 , 3 5 0 .
1 3 8
F o r e x a m p l e : H e r k e n , p . 2 0 ; C r a i g a n d R a d c h e n k o ,
T h e A t o m i c B o m b a n d t h e O r i g i n s o f t h e C o l d W a r
, p p . 9 7 - 8 , 1 0 2 ,
1 0 5 , 1 0 9 ; H o l l o w a y ,
S t a l i n a n d t h e B o m b
, c h a p t e r 8 .
3 5
A c t o r s a t t e m p t e d t o u n d e r s t a n d e a c h o t h e r s ’ i n t e n t i o n s , b u t o f t e n s e n t o u t m i x e d s i g n a l s a n d m i s u n d e r s t o o d
e a c h o t h e r s ’ p o s i t i o n s .
W h e n S o v i e t F o r e i g n M i n i s t e r V y a c h e s l a v M o l o t o v a r r i v e d i n N e w Y o r k i n O c t o b e r
1 3 9
1 9 4 6 , h i s e a r l y c o n c i l i a t o r y s t a t e m e n t s l e d U . S . p o l i c y m a k e r s t o b e l i e v e t h a t h e w o u l d t a k e , a t m o s t , a s o f t s t a n c e
a g a i n s t t h e B a r u c h P l a n d u r i n g h i s U . N . G e n e r a l A s s e m b l y s p e e c h . I n s t e a d h e s t u n n e d b o t h U . S . p o l i c y m a k e r s
a n d t h e U . N . b y a g g r e s s i v e l y a t t a c k i n g t h e p l a n a n d e v e n B a r u c h h i m s e l f ( “ c o n c e i t e d ” a n d “ s h o r t - w i t t e d ” ) .
P o l i c y m a k e r s s c r a m b l e d t o f i g u r e o u t a r e s p o n s e .
M i x e d s i g n a l s w e r e a l s o p r e s e n t a t d o m e s t i c m e e t i n g s . B o h r ,
1 4 0
f o r e x a m p l e , m e t R o o s e v e l t i n 1 9 4 4 t o c a l l f o r i n t e r n a t i o n a l c o n t r o l a n d l e f t t h e m e e t i n g t h i n k i n g t h a t h e h a d
m a d e a n i m p a c t . I n f a c t , B o h r h a d l i t t l e i m p a c t o n R o o s e v e l t o t h e r t h a n t o s u g g e s t t o h i m t h a t t h e D a n i s h
p h y s i c i s t m a y b e a s e c u r i t y r i s k .
1 4 1
D i p l o m a t i c M i s s t e p s
A c t o r s a l s o m a d e m i s s t e p s w h e n c a r r y i n g o u t a t o m i c d i p l o m a c y . S e c r e t a r y o f S t a t e B y r n e s ’ e x c l u s i o n o f t h e
F r e n c h f r o m t h e D e c e m b e r 1 9 4 5 M o s c o w c o n f e r e n c e , f o r e x a m p l e , “ n e e d l e s s l y o f f e n d e d t h a t a l l y ” a n d a l l o w e d
S t a l i n t o m a g n a n i m o u s l y i n v i t e t h e F r e n c h t o f u t u r e c o n f e r e n c e s .
S i m i l a r l y , n o t i n c l u d i n g g e n e r a l
1 4 2
d i s a r m a m e n t i n t o t h e B a r u c h P l a n , a s B a r u c h w a n t e d , g a v e t h e S o v i e t s t h e o p p o r t u n i t y t o a n n o u n c e i t a s p a r t o f
t h e i r r i v a l p r o p o s a l l a t e r o n .
B y r n e s o m i t t e d k e y p a s s a g e s o n s t a g e s i n t h e U . S . p r o p o s a l p r e s e n t e d a t M o s c o w
1 4 3
i n D e c e m b e r 1 9 4 5 . T h i s u p s e t T r u m a n , a n d B y r n e s w a s f o r c e d t o i n c l u d e t h e m l a t e r o n . T h i s m i s s t e p f u r t h e r
d a m a g e d h i s r e l a t i o n s h i p w i t h T r u m a n .
B y r n e s a l s o c o u l d h a v e w o r k e d h a r d e r t o b u i l d s u p p o r t i n C o n g r e s s
1 4 4
b e f o r e e m b a r k i n g o n h i s a t o m i c d i p l o m a c y i n l a t e 1 9 4 5 .
1 4 5
M i s u n d e r s t a n d i n g s o f t h e S i g n i f i c a n c e o f I s s u e s
N e g o t i a t o r s a n d p o l i c y m a k e r s m i s u n d e r s t o o d t h e s i g n i f i c a n c e o f v a r i o u s a s p e c t s o f i n t e r n a t i o n a l c o n t r o l . I n
e a r l y U . S . - S o v i e t d i s c u s s i o n s a t t h e U N A E C d u r i n g J u n e t o S e p t e m b e r 1 9 4 5 , B a r u c h a n d h i s d e l e g a t i o n f o c u s e d
o n t h e v e t o . T h i s f i x a t i o n m i s s e d t h e f a c t t h a t t h e v e t o w a s s t r a t e g i c a l l y i r r e l e v a n t . A s c r i t i c s o f B a r u c h ’ s
i n s i s t e n c e o n r e m o v i n g t h e v e t o ( s u c h a s U n d e r S e c r e t a r y o f S t a t e D e a n A c h e s o n a n d p o l i t i c a l c o m m e n t a t o r
W a l t e r L i p p m a n n ) p o i n t e d o u t : i f t h e i n t e r n a t i o n a l c o n t r o l r e g i m e b r o k e d o w n , i t w o u l d l e a d t o a b r e a k d o w n
i n c o o p e r a t i o n i n t h e U n i t e d N a t i o n s a n d a v e t o w o u l d h a v e l i t t l e i m p a c t .
I n a w i d e l y p u b l i c i z e d S e p t e m b e r
1 4 6
1 9 4 6 a t t a c k o n t h e v e t o a s p e c t o f t h e B a r u c h P l a n , S e c r e t a r y o f C o m m e r c e H e n r y W a l l a c e p o i n t e d o u t t h a t “ t h e
p r o p o s a l t o a b o l i s h t h e v e t o . . . h a s n o m e a n i n g w i t h r e s p e c t t o a t r e a t y o n a t o m i c e n e r g y . . . O n c e t h e t r e a t y i s
r a t i f i e d . . . t h e q u e s t i o n o f v e t o b e c o m e s m e a n i n g l e s s . I f a n y n a t i o n v i o l a t e s t h e t r e a t y p r o v i s i o n . . . t h e r e m a i n i n g
s i g n a t o r y n a t i o n s a r e f r e e t o t a k e w h a t a c t i o n t h e y f e e l i s n e c e s s a r y i n c l u d i n g t h e u l t i m a t e s t e p o f d e c l a r i n g w a r . ”
1 3 9
C r a i g a n d R a d c h e n k o ,
T h e A t o m i c B o m b a n d t h e O r i g i n s o f t h e C o l d W a r
, p . x .
1 4 0
L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a
, p p . 3 6 5 - 6 .
1 4 1
A a s e r u d , “ T h e S c i e n t i s t a n d t h e S t a t e s m e n ” .
1 4 2
H e r k e n , p . 7 4 . B r i t i s h d i p l o m a t s w e r e a l s o u p s e t a t B y r n e s ’ d i s m i s s a l o f d i p l o m a t i c p r o t o c o l . J a m e s L . G o r m l y , “ T h e
W a s h i n g t o n D e c l a r a t i o n a n d t h e ‘ P o o r R e l a t i o n ’ : A n g l o - A m e r i c a n A t o m i c D i p l o m a c y , 1 9 4 5 - 4 6 ” ,
D i p l o m a t i c H i s t o r y
8 , 2
( A p r i l 1 9 8 4 ) , p p . 1 2 5 - 1 4 3 .
1 4 3
H e r k e n , p . 1 8 6 ; L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a
, p . 2 9 0 .
1 4 4
H e r k e n , p . 8 2 .
1 4 5
H e r k e n , p p . 9 1 - 2 ; A r n o l d A . O f f n e r ,
A n o t h e r S u c h V i c t o r y : P r e s i d e n t T r u m a n a n d t h e C o l d W a r , 1 9 4 5 - 1 9 5 3
( S t a n f o r d :
S t a n f o r d U n i v e r s i t y P r e s s , 2 0 0 2 ) , p p . 1 1 6 - 8 .
1 4 6
H e r k e n , p . 1 6 4 . W a l t e r L i p p m a n n , “ M r . B a r u c h a n d t h e V e t o ” ,
W a s h i n g t o n P o s t
( 2 0 J u n e 1 9 4 6 ) .
3 6
B a r u c h m a y a l s o h a v e b e c o m e t r a p p e d b y h i s o w n r h e t o r i c : b y o v e r s e l l i n g t h e i m p o r t a n c e o f t h e v e t o t o t h e
1 4 7
U . S . p u b l i c a n d p o l i t i c a l e l i t e s , h e l e f t h i m s e l f l i t t l e r o o m t o c o m p r o m i s e o n i t .
1 4 8
I n c o n g r u o u s S t a n c e s
E v e n o n i s s u e s s u c h a s s e c r e c y ( w h e r e o n e m i g h t i m a g i n e c l e a r b a t t l e l i n e s b e t w e e n a d v o c a t e s f o r a n d o p p o n e n t s
o f i n c r e a s e d s e c r e c y ) , t h e r e w e r e s e e m i n g l y i n c o n g r u o u s s t a n c e s . T h e S m y t h R e p o r t , r e l e a s e d o n t h e 1 2 t h o f
A u g u s t 1 9 4 5 , s o o n a f t e r t h e a t o m i c b o m b i n g s , i s a c a s e i n p o i n t . T h i s o f f i c i a l p u b l i s h e d p r i m e r o n t h e a t o m i c
b o m b a n d t h e M a n h a t t a n P r o j e c t , w r i t t e n b y p h y s i c i s t H e n r y D e W o l f S m y t h , p r o v i d e d a “ s e m i - t e c h n i c a l ”
( S m y t h ’ s w o r d s ) e x p l a n a t i o n o f h o w t h e a t o m i c b o m b w o r k e d a n d h o w i t w a s d e v e l o p e d . I t p r o v e d t o b e
e s p e c i a l l y h e l p f u l f o r t h e S o v i e t p r o j e c t : t h e N K V D ( t h e S o v i e t s e c r e t p o l i c e ) h a d i t t r a n s l a t e d a n d p u b l i s h e d
w i t h a p r i n t r u n o f 3 0 , 0 0 0 c o p i e s . G i v e n t h e s e c r e c y a r o u n d t h e p r o j e c t , h o w d i d i t c o m e t o b e p u b l i s h e d ? A s i t
h a p p e n s , G r o v e s s u p p o r t e d i t s p u b l i c a t i o n — h e w a n t e d t o j u s t i f y t h e l a r g e e x p e n d i t u r e o n t h e M a n h a t t a n
P r o j e c t a n d c r e a t e a “ s e c u r i t y f e n c e ” h i g h l i g h t i n g w h a t c o u l d b e r e l e a s e d t o t h e p u b l i c o n a t o m i c e n e r g y . G r o v e s
a l s o m i s t a k e n l y b e l i e v e d t h a t i t w o u l d b e o f l i t t l e u s e t o t h e S o v i e t s — a n a s t o n i s h i n g o v e r s i g h t f o r a n i n d i v i d u a l
s o f o c u s e d o n s e c r e c y a n d t h e S o v i e t t h r e a t . O n t h e o t h e r h a n d , m a n y o p p o n e n t s o f s e c r e c y , s u c h a s L e o S z i l a r d ,
D a v i d L i l i e n t h a l , a n d S e c r e t a r y o f C o m m e r c e H e n r y W a l l a c e , a r g u e d t h a t t h e r e p o r t g a v e a w a y i m p o r t a n t
i n f o r m a t i o n t o t h e S o v i e t s . T h e y d i d t h i s i n o r d e r t o d i s c r e d i t t h e m i l i t a r y ' s m a n a g e m e n t o f a t o m i c e n e r g y a n d
t o p u s h f o r c i v i l i a n c o n t r o l .
M o r e o v e r , t h e r e w e r e m i s h a p s t o o w i t h t h e m a n a g e m e n t o f i n f o r m a t i o n r e v e a l e d
1 4 9
i n t h e S m y t h R e p o r t . T h e f i r s t e d i t i o n i n c l u d e d a r e f e r e n c e t o t h e u n f o r e s e e n p o i s o n i n g o f t h e H a n f o r d
r e a c t o r s , w h i c h G r o v e s h a d e x c i s e d f r o m t h e s e c o n d ( a n d m o r e w i d e l y c i r c u l a t e d ) e d i t i o n w h e n h e b e c a m e a w a r e
o f i t . T h e r e v e l a t i o n w a s p i c k e d u p b y t h e S o v i e t g o v e r n m e n t w h e n t h e y c o m p a r e d t h e f i r s t a n d s e c o n d e d i t i o n s
a n d m a y h a v e b e e n u s e f u l f o r t h e S o v i e t p r o g r a m .
1 5 0
D o m e s t i c P o l i t i c s S h a p e d D e b a t e s o n I n t e r n a t i o n a l C o n t r o l
D o m e s t i c p a r t i s a n p o l i t i c s s h a p e d s u p p o r t f o r i n t e r n a t i o n a l c o n t r o l a n d r e l a t e d p o l i c y . T h e r e a r e t w o k e y
e x a m p l e s o f t h i s . F i r s t , d e b a t e s o v e r i n t e r n a t i o n a l c o n t r o l b e c a m e i n t e r t w i n e d w i t h d e b a t e s o v e r l e g i s l a t i o n o n
d o m e s t i c g o v e r n a n c e o f n a t i o n a l a t o m i c f a c i l i t i e s a n d p o l i c i e s . T h e p o l i t i c a l b a t t l e t o s e t u p a d o m e s t i c
o r g a n i z a t i o n a l a n d l e g i s l a t i v e f r a m e w o r k t o g o v e r n a t o m i c e n e r g y b e g a n a t t h e e n d o f t h e w a r , a n d e n d e d w i t h
t h e s i g n i n g o f t h e A t o m i c E n e r g y A c t a n d t h e c r e a t i o n o f t h e A t o m i c E n e r g y C o m m i s s i o n i n A u g u s t 1 9 4 6 . T h e
p o l i t i c a l f i g h t s o v e r t h i s l e g i s l a t i o n p o l a r i z e d i n t o t h o s e w a n t i n g m o r e m i l i t a r y o v e r s i g h t a n d t h o s e a r g u i n g f o r
l e s s . S u p p o r t f o r b o t h d o m e s t i c a n d i n t e r n a t i o n a l c o n t r o l w a s d i v i d e d a l o n g p o l i t i c a l l i n e s : o n d o m e s t i c i s s u e s ,
l i b e r a l a n d p r o g r e s s i v e e l i t e s g e n e r a l l y s u p p o r t e d c i v i l i a n o v e r s i g h t o n d o m e s t i c a t o m i c p o l i c y , w h e r e a s
R e p u b l i c a n s a n d c o n s e r v a t i v e s p r e f e r r e d m i l i t a r y o v e r s i g h t . S i m i l a r l y , D e m o c r a t s w e r e m o r e s u p p o r t i v e o f
1 4 7
L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a ,
p . 3 4 8 .
1 4 8
H e r k e n , p p . 1 7 4 - 7 7 .
1 4 9
G o r d i n ,
R e d C l o u d a t D a w n
, p p . 9 1 - 1 0 4 . T h e S m y t h R e p o r t h a s b e e n r e p u b l i s h e d a s : H e n r y D e W o l f S m y t h ,
A t o m i c
E n e r g y f o r M i l i t a r y P u r p o s e s : T h e O f f i c i a l R e p o r t o n t h e D e v e l o p m e n t o f t h e A t o m i c B o m b U n d e r t h e A u s p i c e s o f t h e U n i t e d
S t a t e s G o v e r n m e n t , 1 9 4 0 - 1 9 4 5
( S t a n f o r d : S t a n f o r d U n i v e r s i t y P r e s s , 1 9 8 9 ) . A c c o r d i n g t o G o r d i n , t h e S m y t h R e p o r t
“ r e v e a l e d t h e s c a l e o f t h e e f f o r t a n d t h e s h e e r q u a n t i t y o f r e s o u r c e s , a n d a l s o h i n t e d a t s o m e o f t h e p a t h s t h a t m i g h t w o r k
a n d , b y o m i s s i o n , s o m e t h a t p r o b a b l y w o u l d n o t . ” F r o m G o r d i n ,
R e d C l o u d a t D a w n
, p . 1 0 3 .
1 5 0
R i c h a r d R h o d e s ,
D a r k S u n : T h e M a k i n g o f t h e H y d r o g e n B o m b
( N e w Y o r k : S i m o n a n d S c h u s t e r , 1 9 9 5 ) , p p . 2 1 5 - 7 .
3 7
i n t e r n a t i o n a l c o n t r o l t h a n R e p u b l i c a n s . T h e m i l i t a r y o p p o s e d i n t e r n a t i o n a l c o n t r o l a n d o p p o s e d c i v i l i a n
c o n t r o l o f d o m e s t i c a t o m i c e n e r g y .
1 5 1
A s e c o n d e x a m p l e i s T r u m a n ’ s a p p o i n t m e n t o f B e r n a r d B a r u c h a s U . S . r e p r e s e n t a t i v e o n t h e U N A E C i n M a r c h
1 9 4 6 . B a r u c h w a s a p p o i n t e d w i t h t h e p o w e r t o m a k e p o l i c y . W i t h t h i s p o w e r f u l a p p o i n t m e n t , T r u m a n
u n d e r c u t b o t h h i s e x i s t i n g a t o m i c e x p e r t s a n d t h e S t a t e D e p a r t m e n t . H e d i d t h i s l a r g e l y b e c a u s e o f B a r u c h ’ s
a c c e p t a b i l i t y t o C o n g r e s s a n d b r o a d e r c o n s e r v a t i v e o p i n i o n . B a r u c h , i n t u r n , a p p o i n t e d f r i e n d s a n d a s s o c i a t e s
w i t h l i t t l e k n o w l e d g e o f a t o m i c m a t t e r s a s h i s a d v i s o r s . T r u m a n s o o n r e g r e t t e d h i s d e c i s i o n .
1 5 2
T h e i n t e r t w i n i n g o f d o m e s t i c p o l i t i c s w i t h i n t e r n a t i o n a l c o n t r o l h a d , o n t h e w h o l e , a n e g a t i v e i m p a c t o n
i n t e r n a t i o n a l c o n t r o l . T h e e f f e c t s w e r e t h a t , f i r s t , i t p r o b a b l y m a d e i n t e r n a t i o n a l c o n t r o l m o r e p o l a r i z e d .
R e p u b l i c a n s a n d c o n s e r v a t i v e s u p p o r t f o r t h e m i l i t a r y o n d o m e s t i c l e g i s l a t i o n l e d t h e m a l s o t o a l l y t h e m s e l v e s t o
t h e m i l i t a r y o n i n t e r n a t i o n a l c o n t r o l . T h e s e d i v i s i o n s m a y n o t h a v e b e e n s o s t a r k i f t h e p o l i c y m a k i n g p r o c e s s e s
h a d n o t r u n c o n c u r r e n t l y o r i f t h e d e b a t e o n d o m e s t i c c o n t r o l h a d t a k e n p l a c e a f t e r t h e d e b a t e o n i n t e r n a t i o n a l
c o n t r o l .
S e c o n d , s c i e n t i s t s , w h o w e r e a k e y l o b b y g r o u p f o r i n t e r n a t i o n a l c o n t r o l , w e r e f o r c e d t o s p e n d t i m e
1 5 3
a n d e n e r g y l o b b y i n g o n d o m e s t i c l e g i s l a t i o n , l e a v i n g t h e m l e s s t i m e a n d e n e r g y t o t h i n k a b o u t i n t e r n a t i o n a l
c o n t r o l .
T h e d e b a t e o n d o m e s t i c a t o m i c l e g i s l a t i o n a l s o e x a c e r b a t e d d i v i d e s a m o n g s t s c i e n t i s t s . T h i s m a y h a v e
1 5 4
l e s s e n e d t h e i r a b i l i t y t o w o r k t o g e t h e r o n i n t e r n a t i o n a l c o n t r o l .
T h i r d , p o l i c y d e c i s i o n s o n i n t e r n a t i o n a l
1 5 5
c o n t r o l w e r e m a d e w i t h d o m e s t i c r e s u l t s i n m i n d . T h e s e d e c i s i o n s m a y n o t h a v e l e d t o t h e b e s t o u t c o m e s f o r
i n t e r n a t i o n a l c o n t r o l i t s e l f . O n e p r o m i n e n t e x a m p l e o f t h i s w a s t h e a p p o i n t m e n t o f B e r n a r d B a r u c h a s c h i e f
p o l i c y m a k e r o n i n t e r n a t i o n a l c o n t r o l . T h e a p p o i n t m e n t o f s o m e o n e w i t h b e t t e r k n o w l e d g e o f a n d i n t e r e s t i n
i n t e r n a t i o n a l c o n t r o l m a y h a v e b e e n a b e t t e r c h o i c e .
1 5 6
1 5 1
O n t h e l e g i s l a t i v e h i s t o r y o f t h e A t o m i c E n e r g y A c t , s e e R i c h a r d G . H e w l e t t a n d O s c a r E . A n d e r s o n , J r . ,
A H i s t o r y o f t h e
U n i t e d S t a t e s A t o m i c E n e r g y C o m m i s s i o n
v o l u m e 1 T h e N e w W o r l d 1 9 3 9 / 1 9 4 6 ( U n i v e r s i t y P a r k , P A : T h e U n i v e r s i t y o f
P e n n s y l v a n i a P r e s s , 1 9 6 2 ) , p p . 4 2 8 - 4 5 5 , 4 8 2 - 5 3 0 . A l s o , o n p o l i t i c a l d i v i s i o n s , s e e H e r k e n , p . 2 6 3 .
1 5 2
H e r k e n , p p . 1 5 8 - 6 0 ; B e r n s t e i n , “ T h e Q u e s t f o r S e c u r i t y ” ; G r a n t ,
B e r n a r d M . B a r u c h
, p . 2 9 2 .
1 5 3
H o g a n ,
A C r o s s o f I r o n ,
p p . 2 3 5 - 2 5 2 .
1 5 4
A c c o r d i n g t o h i s t o r i a n B a r t o n B e r n s t e i n “ A m e r i c a n s c i e n t i s t s d e v o t e d f a r m o r e e n e r g y a n d t h o u g h t i n 1 9 4 6 t o g a i n i n g
c i v i l i a n ( r a t h e r t h a n m i l i t a r y ) c o n t r o l o f a t o m i c e n e r g y t h a n t h e y d i d t o a n a l y z i n g A m e r i c a n p l a n s f o r i n t e r n a t i o n a l
c o n t r o l . ” T h i s d a m p e n e d t h e i r c r i t i c a l e n g a g e m e n t w i t h t h e A c h e s o n - L i l i e n t h a l P l a n . B e r n s t e i n , “ T h e Q u e s t f o r S e c u r i t y ” .
1 5 5
S e n i o r s c i e n t i s t s B u s h , C o n a n t , a n d O p p e n h e i m e r s u p p o r t e d t h e M a y - J o h n s o n B i l l , w h e r e a s m o s t r a n k a n d f i l e a t o m i c
s c i e n t i s t s o p p o s e d i t . W a n g ,
A m e r i c a n S c i e n t i s t s i n a n A g e o f A n x i e t y
, p p . 1 4 - 1 5 .
1 5 6
H e r k e n , p p . 1 5 8 - 6 0 .
3 8
4 . 8 V i a b i l i t y o f I n t e r n a t i o n a l C o n t r o l
L e s s o n s
A c h i e v i n g a g r e e m e n t o n a w o r k a b l e s c h e m e f o r i n t e r n a t i o n a l c o n t r o l i s d i f f i c u l t , e v e n i f t h e p o l i t i c a l a t m o s p h e r e
i s c o n d u c i v e o r n e g o t i a t o r s a r e m o r e w i l l i n g t o c o m p r o m i s e . T h e r e m a y b e f u n d a m e n t a l s t r u c t u r a l s t r a t e g i c
o b s t a c l e s , s u c h a s a n i n s u r m o u n t a b l e t r a n s p a r e n c y - s e c u r i t y t r a d e - o f f .
T h u s , w h i l e t h e r e m a y b e i n t e r v e n t i o n s
1 5 7
t o i m p r o v e t h e c h a n c e s o f s u c c e s s f u l i n t e r n a t i o n a l c o n t r o l , s u c h a s m a k i n g p o l i c y m a k i n g m o r e i n f o r m e d , i t m a y
n o t b e e n o u g h t o a c h i e v e s u c c e s s f u l i n t e r n a t i o n a l c o n t r o l . I t m a y b e t h a t r a d i c a l l y d i f f e r e n t p o l i t i c a l o r s o c i a l
c i r c u m s t a n c e s w o u l d b e r e q u i r e d f o r i t s s u c c e s s .
H i s t o r i c a l C a s e
I m p r o v i n g p r o c e s s e s , w i t h c l e a r e r , m o r e t r a n s p a r e n t , a n d m o r e i n f o r m e d p o l i c y m a k i n g w o u l d n o t l i k e l y h a v e
l e d t o s u c c e s s f u l i n t e r n a t i o n a l c o n t r o l i n 1 9 4 5 / 4 6 . T h i s i s o n l y l i k e l y t o h a v e b e e n a c h i e v e d u n d e r r a d i c a l l y
d i f f e r e n t h i s t o r i c a l c i r c u m s t a n c e s . T h i s i s b e c a u s e o f i m p o r t a n t u n d e r l y i n g f a c t o r s w o r k i n g a g a i n s t i n t e r n a t i o n a l
c o n t r o l . W e h a v e g r o u p e d t h e s e f a c t o r s b e l o w i n t o f u n d a m e n t a l s t r u c t u r a l s t r a t e g i c o b s t a c l e s , a w i d e r l a c k o f
s u p p o r t i n t h e U . S . a n d t h e S o v i e t U n i o n , a n d c o n f u s i o n i n p o l i c y m a k i n g . T h e s e f a c t o r s w e r e i n t e r r e l a t e d : t h e
f u n d a m e n t a l s t r u c t u r a l o b s t a c l e s t e n d e d t o l e a d t o c o n f u s i o n i n p o l i c y m a k i n g a n d a l a c k o f s u p p o r t f o r
i n t e r n a t i o n a l c o n t r o l i n t h e U . S . a n d t h e S o v i e t U n i o n .
F u n d a m e n t a l S t r u c t u r a l S t r a t e g i c O b s t a c l e s
T h e r e a r e f u n d a m e n t a l s t r a t e g i c o b s t a c l e s t o a r m s c o n t r o l . F o r e x a m p l e , i n t e r n a t i o n a l c o n t r o l r e q u i r e d s o m e
s t a g e d p r o c e s s . B u t d i f f e r e n t s t a g e s b e s t o w a d v a n t a g e s a n d d i s a d v a n t a g e s t o d i f f e r e n t s i d e s . T h u s i t c a n b e
f u n d a m e n t a l l y c h a l l e n g i n g t o d e v e l o p a p o l i c y t h a t b a l a n c e s t h e s e a d v a n t a g e s a n d d i s a d v a n t a g e s , a n d m o r e s o i f
t h e t w o s i d e s d i s a g r e e a b o u t t h e i r p o w e r a s w e l l a s t h e s i z e o f t h e a d v a n t a g e s a n d d i s a d v a n t a g e s b e s t o w e d b y t h e
s t a g e s . S o , f o r e x a m p l e , t h e B a r u c h P l a n s u g g e s t e d a s u r v e y o f S o v i e t r e s o u r c e s a s a f i r s t s t e p , w i t h o u t t h e U . S .
r e c i p r o c a t i n g a t t h a t p o i n t . T h i s g a v e t h e U . S . t h e o p t i o n t o a b o r t i n t e r n a t i o n a l c o n t r o l a t t h a t s t a g e , h a v i n g
g a i n e d s e n s i t i v e i n f o r m a t i o n a b o u t t h e S o v i e t a t o m i c p r o g r a m w i t h o u t e q u i v a l e n t r e c i p r o c a t i o n .
1 5 8
A s e c o n d o b s t a c l e i s t h a t a n y m o n i t o r i n g ( o r c o n t r o l ) s c h e m e w h i c h w o u l d b e t r a n s p a r e n t e n o u g h t o a s s u r e t h e
o t h e r p a r t y t h e y w i l l n o t b e c a u g h t o f f g u a r d b y a s e c r e t a r m a m e n t s p r o g r a m w i l l b e t o o i n v a s i v e f o r t h e
1 5 7
T h i s o b s t a c l e t o a r m s c o n t r o l i s w e l l d e v e l o p e d i n : A n d r e w J . C o e a n d J a n e V a y n m a n , “ W h y A r m s C o n t r o l I s S o R a r e ” ,
A m e r i c a n P o l i t i c a l S c i e n c e R e v i e w
1 1 4 , 2 ( 2 0 2 0 ) , p p . 3 4 2 – 5 5 .
1 5 8
“ T h e A m e r i c a n P r o p o s a l f o r I n t e r n a t i o n a l C o n t r o l P r e s e n t e d b y B e r n a r d B a r u c h ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
1 & 2
( 1 J u l y 1 9 4 6 ) p p . 3 - 5 , 1 0 . A l s o a t :
h t t p : / / w w w . a t o m i c a r c h i v e . c o m / D o c s / D e t e r r e n c e / B a r u c h P l a n . s h t m l
. A c c e s s e d 2 5 A p r i l
2 0 1 9 . S e c r e t a r y o f C o m m e r c e H e n r y W a l l a c e m a d e t h i s v e r y o b s e r v a t i o n a n d c r i t i c i s m o f t h e B a r u c h P l a n , n o t i n g t h a t t h e
S o v i e t U n i o n h a d o n l y “ t w o c a r d s w h i c h s h e c a n u s e i n n e g o t i a t i n g w i t h u s : ( l ) o u r l a c k o f i n f o r m a t i o n o n t h e s t a t e o f h e r
s c i e n t i f i c a n d t e c h n i c a l p r o g r e s s o n a t o m i c e n e r g y a n d ( 2 ) o u r i g n o r a n c e o f h e r u r a n i u m a n d t h o r i u m r e s o u r c e s . T h e s e
c a r d s a r e n o t h i n g l i k e a s p o w e r f u l a s o u r c a r d s - - - a s t o c k p i l e o f b o m b s , m a n u f a c t u r i n g p l a n t s i n a c t u a l p r o d u c t i o n , B - 2 9 s
a n d B - 3 6 s , a n d o u r b a s e s c o v e r i n g h a l f t h e g l o b e . Y e t w e a r e i n e f f e c t a s k i n g h e r t o r e v e a l h e r o n l y t w o c a r d s i m m e d i a t e l y -
t e l l i n g h e r t h a t a f t e r w e h a v e s e e n h e r c a r d s w e w i l l d e c i d e w h e t h e r w e w a n t t o c o n t i n u e t o p l a y t h e g a m e . ” H e r k e n , p p .
1 8 1 - 8 2 .
3 9
m o n i t o r e d p a r t y , e x p o s i n g t h e m t o s e c u r i t y r i s k s .
I n t h i s c a s e , t h i s t r a d e - o f f b i t h a r d . T h e U . S . m i l i t a r y s o u g h t
1 5 9
i n f o r m a t i o n o n S o v i e t c a p a b i l i t i e s , a n d w a s i n f a c t c a u g h t o f f g u a r d b y t h e i r d e v e l o p m e n t o f t h e a t o m i c b o m b .
I n t u r n , t h e S o v i e t U n i o n p e r c e i v e d i t t o b e a f u n d a m e n t a l r i s k t o a l l o w f o r e i g n a c t o r s t o h a v e e x t e n s i v e a c c e s s t o
t h e S o v i e t U n i o n .
I t ’ s h a r d t o i m a g i n e a n y m o n i t o r i n g s c h e m e w h i c h w o u l d h a v e p e r m i t t e d t h e U n i t e d S t a t e s
1 6 0
t o h a v e s u f f i c i e n t i n f o r m a t i o n , w h i c h w o u l d n ’ t h a v e b e e n p e r c e i v e d a s a n e x i s t e n t i a l t h r e a t t o t h e S o v i e t U n i o n .
L a c k o f S u p p o r t i n t h e U . S .
M a n y U . S . p o l i c y m a k i n g e l i t e s w e r e w e a k s u p p o r t e r s o r e v e n o p p o n e n t s o f i n t e r n a t i o n a l c o n t r o l . T h i s w a s
b e c a u s e , f i r s t , t h e y b e l i e v e d t h a t s o l o n g a s t h e r i s k o f w a r e x i s t e d w i t h t h e S o v i e t s , t h e a t o m i c b o m b w a s n e e d e d
f o r U . S . d e t e r r e n c e a n d d e f e n s e . W i t h s m a l l e r p o s t w a r m i l i t a r y f o r c e s u n a b l e t o d e f e n d E u r o p e a g a i n s t a S o v i e t
i n v a s i o n , a t o m i c b o m b s w e r e s e e n a s a n i m p o r t a n t e l e m e n t o f t h e c o u n t r y ’ s m i l i t a r y a r s e n a l . O p p o n e n t s o f
i n t e r n a t i o n a l c o n t r o l s u c h a s G r o v e s h e l d t h a t t h e a t o m i c b o m b c o u l d o n l y b e r e l i n q u i s h e d o n c e w a r i t s e l f w a s
e n f o r c e a b l y o u t l a w e d , w h i c h t h e y l a r g e l y d i d n o t b e l i e v e p o s s i b l e .
S e c o n d , p o l i c y m a k e r s b e l i e v e d t h a t a
1 6 1
m o n o p o l y o n a t o m i c w e a p o n s ( n o m a t t e r h o w s h o r t l i v e d ) , g a v e t h e U n i t e d S t a t e s c o e r c i v e d i p l o m a t i c p o w e r .
T h u s , t h e a t o m i c b o m b w a s d e s c r i b e d b y s o m e a s “ t h e w i n n i n g w e a p o n . ”
1 6 2
T h i r d , p o l i c y m a k e r s w e r e a l s o i n s u f f i c i e n t l y w o r r i e d a b o u t a p o s s i b l e a r m s r a c e . M o s t w e r e n o t l o o k i n g m o r e
t h a n a d e c a d e o r s o a h e a d . I n t h a t t i m e f r a m e , m a n y w e r e c e r t a i n t h e y c o u l d k e e p a h e a d i n a n y a r m s r a c e a n d
a n y w a y b e l i e v e d t h a t t h e y w o u l d h a v e a m o n o p o l y o n a t o m i c w e a p o n s f o r m a n y y e a r s .
T h i s t h i n k i n g w a s
1 6 3
b u i l t o n t h e a s s u m p t i o n t h a t S o v i e t i n d u s t r i a l , s c i e n t i f i c , a n d e c o n o m i c r e s o u r c e s w e r e s i g n i f i c a n t l y p o o r e r t h a n
t h o s e o f t h e U . S . a n d c e r t a i n l y n o t e n o u g h f o r t h e U . S . S . R . t o c a t c h u p w i t h t h e U . S . i n t e r m s o f a t o m i c r e s e a r c h
a n d d e v e l o p m e n t . F o r e x a m p l e , e a r l y a n a l y s e s o f t h e S o v i e t b o m b p r o g r a m a s s u m e d t h a t t h e S o v i e t U n i o n h a d a
l i m i t e d a b i l i t y t o w o r k i n p a r a l l e l o n t h e v a r i o u s e l e m e n t s n e e d e d t o m a k e a b o m b .
S i m i l a r l y , t h e U . S . A i r
1 6 4
1 5 9
S e e f o r e x a m p l e : C o e a n d V a y n m a n , “ W h y A r m s C o n t r o l I s S o R a r e ” .
1 6 0
C r a i g a n d R a d c h e n k o ,
T h e A t o m i c B o m b a n d t h e O r i g i n s o f t h e C o l d W a r
, p p . 1 3 6 , 1 3 9 . H e r k e n , p . 1 7 7 .
1 6 1
O n m i l i t a r y o p p o s i t i o n , s e e L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a
, p p . 2 8 6 - 9 . F o r G r o v e ’ s a r g u m e n t s : N o r r i s ,
R a c i n g f o r t h e B o m b
, p p . 4 7 1 - 2 .
1 6 2
H e r k e n , p p . 4 - 8 .
1 6 3
H e r k e n , p . 7 . A l s o s e e p r e v i o u s l e s s o n o n
S e c r e c y a n d S e c u r i t y
.
1 6 4
Z i e g l e r , “ I n t e l l i g e n c e A s s e s s m e n t s o f S o v i e t A t o m i c C a p a b i l i t y , 1 9 4 5 - 1 9 4 9 ” ; G o r d i n ,
R e d C l o u d a t D a w n
, p p . 7 2 - 4 . I n
t h e
S a t u r d a y E v e n i n g P o s t
i n 1 9 4 8 , G r o v e s w r o t e t h a t t h e S o v i e t U n i o n " s i m p l y d o e s n o t h a v e e n o u g h p r e c i s i o n i n d u s t r y ,
t e c h n i c a l s k i l l o r s c i e n t i f i c n u m e r i c a l s t r e n g t h t o c o m e e v e n c l o s e t o d u p l i c a t i n g t h e m a g n i f i c e n t a c h i e v e m e n t o f t h e
A m e r i c a n i n d u s t r i a l i s t s , s k i l l e d l a b o r , e n g i n e e r s a n d s c i e n t i s t s w h o m a d e t h e M a n h a t t a n P r o j e c t a s u c c e s s . I n d u s t r i a l l y ,
R u s s i a i s , p r i m a r i l y , a h e a v y - i n d u s t r y n a t i o n ; s h e u s e s a x l e g r e a s e w h e r e w e u s e f i n e l u b r i c a t i n g o i l s . I t i s a n
o x c a r t - v e r s u s - a u t o m o b i l e s i t u a t i o n . " F r o m R h o d e s ,
D a r k S u n
,
p . 2 1 1 . D a v i d L i l i e n t h a l w o u l d l a t e r n o t e t h a t G r o v e s w e n t
t o o f a r i n t h i n k i n g o f t h e S o v i e t U n i o n a s a n “ i g n o r a n t , c l u m s y , b a c k w a r d c o u n t r y ” . S e e Z i e g l e r , “ I n t e l l i g e n c e A s s e s s m e n t s
o f S o v i e t A t o m i c C a p a b i l i t y , 1 9 4 5 - 1 9 4 9 ” . T h i s c h a r a c t e r i z a t i o n w a s n o t l i m i t e d t o G r o v e s , b u t w i d e l y h e l d i n g o v e r n m e n t
c i r c l e s . R i c h a r d R h o d e s n o t e s o n e p o p u l a r j o k e i n W a s h i n g t o n , D . C . , d u r i n g t h e w a r : “ T h e R u s s i a n s c o u l d n ' t d e l i v e r a n
a t o m i c b o m b i n a s u i t c a s e , t h e j o k e w e n t , b e c a u s e t h e y d i d n ' t k n o w h o w t o m a k e a s u i t c a s e . ” R h o d e s ,
D a r k S u n
, p . 2 1 1 .
T h e S o v i e t a t o m i c p r o g r a m w a s n e v e r t h e l e s s a m a s s i v e a n d d i f f i c u l t u n d e r t a k i n g f o r t h e S o v i e t U n i o n . O n e h i s t o r i a n
e s t i m a t e s t h a t t h e p r o g r a m c o s t m o r e i n a b s o l u t e t e r m s t h a n t h e U . S . a t o m i c p r o g r a m : V l a d i s l a v M . Z u b o k , “ S t a l i n a n d t h e
N u c l e a r A g e ” , i n J o h n L e w i s G a d d i s e t a l ( e d s . ) ,
C o l d W a r S t a t e s m e n C o n f r o n t t h e B o m b : N u c l e a r D i p l o m a c y S i n c e 1 9 4 5
( O x f o r d : O x f o r d U n i v e r s i t y P r e s s , 1 9 9 9 ) , p p . 3 9 - 6 1 .
4 0
F o r c e b e l i e v e d i n 1 9 4 6 t h a t i t w o u l d b e m a n y y e a r s b e f o r e t h e S o v i e t U n i o n b u i l t a f l e e t o f h e a v y b o m b e r s
c a p a b l e o f c a r r y i n g a t o m i c b o m b s .
1 6 5
T h e b e l i e f t h a t t h e S o v i e t U n i o n w o u l d n o t b e a b l e t o d e v e l o p a n a t o m i c b o m b i n t h e f o r t i e s w a s a l s o f o u n d e d
o n t h e c o n c l u s i o n t h a t t h e U . S . S . R . d i d n o t h a v e a c c e s s t o s u f f i c i e n t h i g h - g r a d e u r a n i u m . T h e r a p i d e x p l o i t a t i o n
o f t h e S o v i e t U n i o n ’ s ( s u p p o s e d ) l o w - g r a d e u r a n i u m w o u l d r e q u i r e , t h o u g h t G r o v e s , a “ r e v o l u t i o n i n e x t r a c t i o n
t e c h n i q u e s ” t h a t w a s b e y o n d t h e S o v i e t U n i o n ’ s c u r r e n t t e c h n i c a l c a p a b i l i t i e s .
S o m e ( s u c h a s t h e d i p l o m a t
1 6 6
a n d a n a l y s t G e o r g e K e n n a n ) a l s o t h o u g h t t h a t i n t e r n a t i o n a l d i p l o m a c y , c o n f l i c t , o r p e r h a p s e v e n a S o v i e t
c o l l a p s e c o u l d w o r k t o s l o w o r h a l t t h e S o v i e t a t o m i c d e v e l o p m e n t .
T h i s c o n f i d e n c e w a s b o o s t e d b y t h e
1 6 7
a s s u m p t i o n t h a t t h e U . S . c o u l d k e e p a h e a d b y a c c e l e r a t i n g r e s e a r c h a n d d e v e l o p m e n t i n a t o m i c w e a p o n s .
N o r
1 6 8
w e r e t h e s e d e v e l o p m e n t s n e c e s s a r i l y i n c r e m e n t a l : s o m e p h y s i c i s t s ( i n p a r t i c u l a r E d w a r d T e l l e r ) b e l i e v e d t h a t i t
w a s p o s s i b l e f o r t h e U n i t e d S t a t e s t o d e v e l o p t h e f a r m o r e p o w e r f u l t h e r m o n u c l e a r “ s u p e r b o m b . ”
1 6 9
M u c h o f t h e s u p p o r t f o r i n t e r n a t i o n a l c o n t r o l w a s t r a n s i e n t . F o r e x a m p l e , c y n i c a l r e a l i s t s s u c h a s B a r u c h a n d h i s
t e a m w e r e o n l y c o m m i t t e d t o a n i n t e r n a t i o n a l c o n t r o l d e a l i n w h i c h t h e U . S . w o u l d n o t n e e d t o m a k e a n y
s u b s t a n t i a l c o m p r o m i s e s . O n c e t h e i r p r e f e r r e d d e a l s w e r e r e j e c t e d , t h e y r e j e c t e d i n t e r n a t i o n a l c o n t r o l a l t o g e t h e r .
O t h e r s , s u c h a s s e n i o r R e p u b l i c a n S e n a t o r A r t h u r V a n d e n b e r g , o n l y s w i t c h e d t o s u p p o r t t h e U n i t e d N a t i o n s
1 7 0
w h e n i t w a s p o l i t i c a l l y e x p e d i e n t ( a n d p o s s i b l y e l e c t o r a l l y p o p u l a r ) .
S u p p o r t f r o m k e y p o l i c y m a k e r s ( e . g . , t h e
1 7 1
w a r t i m e S e c r e t a r y o f W a r H e n r y L . S t i m s o n a n d S e c r e t a r y o f S t a t e J a m e s F . B y r n e s ) a l s o w a x e d a n d w a n e d
d e p e n d i n g o n t h e i r a s s e s s m e n t o f S o v i e t f l e x i b i l i t y a n d U . S . p r o g r e s s i n a t o m i c w e a p o n s .
T h e p u b l i c w a s i t s e l f
1 7 2
e a s i l y a l a r m e d b y s e c u r i t y a n d s e c r e c y c o n c e r n s ( s e e p r e v i o u s l e s s o n
S e c r e c y a n d S e c u r i t y
) .
L a c k o f S u p p o r t i n t h e S o v i e t U n i o n
S o v i e t a t o m i c p o l i c y w a s d e t e r m i n e d b y S t a l i n , a n d h i s t o r i a n s b e l i e v e t h a t h e w a s u n l i k e l y t o a g r e e t o a n y f o r m
o f i n t e r n a t i o n a l c o n t r o l . A f t e r A u g u s t 1 9 4 5 , h e w a s f u l l y c o m m i t t e d t o t h e d e v e l o p m e n t o f t h e S o v i e t a t o m i c
b o m b .
S t a l i n d i d n o t a p p e a r t o f e a r t h e d e s t r u c t i v e e f f e c t s o f t h e a t o m i c b o m b o r a s u b s e q u e n t a r m s r a c e , a n d
1 7 3
t h o u g h t t h a t h e c o u l d c a t c h u p w i t h t h e U . S . a t o m i c p r o g r a m o r c o m p e n s a t e t h r o u g h o t h e r m e a n s ( e . g . , l a r g e r
1 6 5
T h e a i r f o r c e e v e n d i s m i s s e d r e p o r t s i n t h e B e r l i n p r e s s i n N o v e m b e r 1 9 4 6 t h a t t h e S o v i e t U n i o n h a d b e g u n w o r k o n a n
i n d i g e n o u s c o p y o f t h e B - 2 9 S u p e r f o r t r e s s . C o n c e r n o v e r S o v i e t d e l i v e r y c a p a b i l i t y w a s o n l y r a i s e d f o l l o w i n g t h e u n v e i l i n g
o f t h r e e T u - 4 h e a v y b o m b e r s a t t h e 1 9 4 7 T u s h i n o d i s p l a y . S t e v e Z a l o g a ,
T a r g e t A m e r i c a : t h e S o v i e t U n i o n a n d t h e S t r a t e g i c
A r m s R a c e , 1 9 4 5 - 1 9 6 4
( N o v a t o , C A : P r e s i d o , 1 9 9 3 ) , p . 7 3 .
1 6 6
Z i e g l e r , “ I n t e l l i g e n c e A s s e s s m e n t s o f S o v i e t A t o m i c C a p a b i l i t y , 1 9 4 5 - 1 9 4 9 ” ; G o r d i n ,
R e d C l o u d a t D a w n
, p p . 7 2 - 4 .
1 6 7
H e r k e n , p p . 1 8 7 , 2 1 2 , 2 1 6 , 2 3 5 - 3 6 ; D a v i d M a y e r s ,
G e o r g e K e n n a n a n d t h e D i l e m m a s o f U S F o r e i g n P o l i c y
( O x f o r d :
O x f o r d U n i v e r s i t y P r e s s , 1 9 8 8 ) , p p . 1 2 0 , 3 0 3 - 5 , 3 0 7 .
1 6 8
Z a c h a r y ,
E n d l e s s F r o n t i e r
, p p . 2 5 0 - 5 1 .
1 6 9
R h o d e s ,
D a r k S u n
, p p . 2 0 6 - 7 .
1 7 0
F o r e x a m p l e , B a r u c h . H e r k e n , p . 1 7 0 .
1 7 1
H e r k e n , p . 7 5 .
1 7 2
O n S t i m s o n : H e r k e n , p . 1 9 ; R o n a l d E . P o w a s k i ,
M a r c h t o A r m a g e d d o n : T h e U n i t e d S t a t e s a n d t h e N u c l e a r A r m s R a c e ,
1 9 3 9 t o t h e P r e s e n t
( O x f o r d : O x f o r d U n i v e r s i t y P r e s s , 1 9 8 7 ) , p p . 2 9 - 3 0 . O n B y r n e s : H e r k e n , p p . 6 1 , 7 1 , 7 2 .
1 7 3
C r a i g a n d R a d c h e n k o ,
T h e A t o m i c B o m b a n d t h e O r i g i n s o f t h e C o l d W a r
, p p . 9 9 - 1 0 1 , 1 0 5 - 0 6 , 1 0 9 - 1 1 0 ; H o l l o w a y ,
S t a l i n a n d t h e B o m b
, p . 1 6 2 ; V l a d i s l a v Z u b o k a n d C o n s t a n t i n e P l e s h a k o v ,
I n s i d e t h e K r e m l i n ’ s C o l d W a r : F r o m S t a l i n t o
K h r u s h c h e v
( C a m b r i d g e , M A : H a r v a r d U n i v e r s i t y P r e s s , 1 9 9 6 ) , p p . 4 0 - 4 6 ; Z u b o k , “ S t a l i n a n d t h e N u c l e a r A g e ” .
4 1
c o n v e n t i o n a l f o r c e s ) .
H e h a d a c y n i c a l a t t i t u d e t o c o o p e r a t i o n ; “ s c i e n t i f i c e x c h a n g e , ” f o r e x a m p l e , m e a n t
1 7 4
e x t r a c t i n g s c i e n t i f i c a n d o t h e r i n s i g h t s f r o m t h e U . S .
S t a l i n w a s a l s o s e t a g a i n s t f o r e i g n m i s s i o n s i n t h e S o v i e t
1 7 5
U n i o n , f o r e x a m p l e , f o r i n s p e c t i o n s o r m o n i t o r i n g .
S o v i e t d e l e g a t e t o t h e U N A E C A n d r e i G r o m y k o w o u l d
1 7 6
r e f l e c t f o r t y y e a r s l a t e r t h a t “ I a m c e r t a i n t h a t S t a l i n w o u l d n o t h a v e g i v e n u p t h e c r e a t i o n o f h i s o w n a t o m i c
b o m b . H e w e l l u n d e r s t o o d t h a t T r u m a n w o u l d n o t g i v e u p a t o m i c w e a p o n s . ”
1 7 7
T e n s i o n s b e t w e e n t h e U . S . a n d t h e S o v i e t U n i o n
G r o w i n g t e n s i o n s b e t w e e n t h e U . S . a n d t h e S o v i e t U n i o n u n d e r m i n e d c o n f i d e n c e i n e a c h o t h e r , i n c r e a s e d t h e
a l l u r e o f a t o m i c w e a p o n s , a n d n o d o u b t m a d e a g r e e m e n t h a r d e r t o r e a c h . I t l e d t o , i n G a d d i s ’ w o r d s , a
“ g r o w i n g s e n s e o f i n s e c u r i t y ” i n 1 9 4 5 a n d 1 9 4 6 .
W e k n o w m o s t a b o u t U . S . a t t i t u d e s , w h e r e h i s t o r i a n s h a v e
1 7 8
n o t e d t h a t d i s a g r e e m e n t s o v e r p o s t w a r E u r o p e , f o r e x a m p l e o v e r i s s u e s i n P o l a n d , R o m a n i a , a n d G e r m a n y , a n d
l a t e r o v e r I r a n , d e c r e a s e d U . S . t r u s t i n t h e S o v i e t U n i o n f r o m 1 9 4 5 o n w a r d s . T h e d i s c o v e r y o f S o v i e t a t o m i c
e s p i o n a g e f r o m 1 9 4 3 d e t r a c t e d f r o m c o n f i d e n c e b u i l d i n g , t h o u g h h i s t o r i a n s a r e d i v i d e d o n h o w m u c h o f a n
i m p a c t i t m a y h a v e h a d o n U . S . p o l i c y t o w a r d s t h e S o v i e t U n i o n .
O n t h e S o v i e t s i d e , t h e r e i s s o m e i n d i c a t i o n
1 7 9
t h a t c e r t a i n U . S . p o l i c i e s , s u c h a s T r u m a n ’ s c a n c e l l a t i o n o f l e n d - l e a s e t o t h e S o v i e t U n i o n , m a y h a v e h a r d e n e d
S o v i e t a t t i t u d e s t o w a r d s t h e U . S . , l e a d i n g i n t h i s c a s e t o a n i n c r e a s e i n “ u n i l a t e r a l i s t ” t e n d e n c i e s .
1 8 0
P r o c e s s
T h e r e w a s s i g n i f i c a n t c o n f u s i o n a n d m u d d l i n g t h r o u g h i n p o l i c y m a k i n g ( s e e t h e l e s s o n
M u d d l e d
P o l i c y m a k i n g
) . G i v e n t h a t o n e s i d e h a d a t o m i c w e a p o n s a n d t h e o t h e r d i d n o t , i t w a s d i f f i c u l t t o c r e a t e a
t r a n s i t i o n a l p r o c e s s w h i c h w o u l d p r o v i d e s e c u r i t y t o b o t h s i d e s a n d n o t l e a v e t h e m v u l n e r a b l e . T h i s w a s
e s p e c i a l l y p r o b l e m a t i c g i v e n t h e i n c r e a s i n g t r u s t d e f i c i t b e t w e e n t h e t w o c o u n t r i e s i n l a t e 1 9 4 5 a n d 1 9 4 6 .
1 8 1
E v e n i f a n i n t e r n a t i o n a l c o n t r o l a g r e e m e n t h a d b e e n a c h i e v e d , i t m a y h a v e q u i c k l y b r o k e n d o w n a s a l l t y p e s o f
m o n i t o r i n g , i n p r a c t i c e , w e r e d i f f i c u l t
a n d c h a n g i n g i n t e r n a t i o n a l r e l a t i o n s , d o m e s t i c p o l i t i c s , o r p u b l i c
1 8 2
o p i n i o n m a y h a v e l e d o n e c o u n t r y o r t h e o t h e r t o a b a n d o n t h e a g r e e m e n t .
1 8 3
D i s t a n t C o u n t e r f a c t u a l s
I m p r o v i n g p r o c e s s e s , w i t h c l e a r e r , m o r e t r a n s p a r e n t , a n d m o r e i n f o r m e d p o l i c y m a k i n g w o u l d p r o b a b l y n o t
h a v e l e d t o s u c c e s s f u l i n t e r n a t i o n a l c o n t r o l . O n l y v e r y r a d i c a l l y d i f f e r e n t h i s t o r i c a l c i r c u m s t a n c e s , w h i c h w o u l d
1 7 4
C r a i g a n d R a d c h e n k o ,
T h e A t o m i c B o m b a n d t h e O r i g i n s o f t h e C o l d W a r
, p . 1 0 9 - 1 0 .
1 7 5
C r a i g a n d R a d c h e n k o ,
T h e A t o m i c B o m b a n d t h e O r i g i n s o f t h e C o l d W a r
, p p . 1 3 6 , 1 3 9 .
1 7 6
H e r k e n , p . 1 7 7 .
1 7 7
C i t e d i n : H o l l o w a y , “ T h e S o v i e t U n i o n a n d t h e B a r u c h P l a n ” .
1 7 8
J o h n L e w i s G a d d i s ,
T h e C o l d W a r
( L o n d o n : A l l e n L a n e , 2 0 0 5 ) , 2 7 .
1 7 9
L y n n E t h e r i d g e D a v i s , T h e C o l d W a r B e g i n s : S o v i e t - A m e r i c a n C o n f l i c t O v e r E a s t e r n E u r o p e ( P r i n c e t o n , N J : P r i n c e t o n
U n i v e r s i t y P r e s s , 1 9 7 4 ) , c h a p t e r s 7 , 8 , a n d 9 . R a i n e , “ T h e I r a n i a n C r i s i s o f 1 9 4 6 a n d t h e O r i g i n s o f t h e C o l d W a r ” . F o r t h e
a r g u m e n t t h a t i t u n d e r m i n e d t r u s t , s e e : K a t h e r i n e A . S . S i b l e y ,
R e d S p i e s i n A m e r i c a : S t o l e n S e c r e t s a n d t h e D a w n o f t h e
C o l d W a r
( L a w r e n c e , K S : U n i v e r s i t y o f K a n s a s P r e s s , 2 0 0 4 ) . p . 1 7 4 - 1 8 0 . F o r t h e o p p o s i t e p o i n t o f v i e w s e e : H e r k e n , p p .
1 3 5 - 3 6 .
1 8 0
V l a d i s l a v M . Z u b o k ,
A F a i l e d E m p i r e : T h e S o v i e t U n i o n i n t h e C o l d W a r f r o m S t a l i n t o G o r b a c h e v
( C h a p e l H i l l , N C :
U n i v e r s i t y o f N o r t h C a r o l i n a P r e s s , 2 0 0 7 ) , p . 1 5 .
1 8 1
C r a i g a n d R a d c h e n k o ,
T h e A t o m i c B o m b a n d t h e O r i g i n s o f t h e C o l d W a r
, p . 1 3 8 .
1 8 2
T o a p p r e c i a t e t h e d i f f i c u l t i e s i n m o n i t o r i n g , s e e G o r d i n ,
R e d C l o u d a t D a w n
, c h a p t e r 5 .
1 8 3
F o r a s e n s e o f t h e r a p i d l y c h a n g i n g i n t e r n a t i o n a l p o l i t i c s o f t h e l a t e f o r t i e s , s e e G a d d i s ,
T h e C o l d W a r
, c h a p t e r 1 .
4 2
h a v e c h a n g e d t h e u n d e r l y i n g p o l i t i c a l a n d s o c i a l d y n a m i c s , m a y h a v e l e d t o i n t e r n a t i o n a l c o n t r o l . T h e s e m o s t
d i s t a n t c o u n t e r f a c t u a l s c o u l d b e :
●
I f t h e S o v i e t U n i o n w a s w i l l i n g t o s e t t l e f o r a j u n i o r s t a t u s i n t e r n a t i o n a l l y v i s - a - v i s t h e U n i t e d S t a t e s .
T h i s m i g h t h a v e c o m e a b o u t i n a v a r i e t y o f w a y s , i n c l u d i n g , o n e h i s t o r i a n h a s s p e c u l a t e d , t h r o u g h
S t a l i n ’ s d e a t h i n t h e f a l l o f 1 9 4 5 , i n w h i c h h i s s u c c e s s o r s “ m i g h t h a v e c h o s e n a m o r e a c c o m m o d a t i n g
c o u r s e t o w a r d t h e U n i t e d S t a t e s ”
.
1 8 4
●
I f t h e r e h a d b e e n l e s s c o n f l i c t o f i n t e r e s t b e t w e e n t h e S o v i e t U n i o n a n d t h e U . S . i n E u r o p e a n d A s i a .
F o r e x a m p l e , i f t h e S o v i e t s w e r e l e s s e x p a n s i o n i s t a n d / o r t h e U . S . m o r e a c c e p t i n g o f S o v i e t
e x p a n s i o n i s m .
●
I f t h e r e h a d b e e n b o t h c o n v e n t i o n a l a n d n u c l e a r p o w e r p a r i t y b e t w e e n t h e t w o s u p e r p o w e r s , m a k i n g i t
e a s i e r f o r t h e m t o i d e n t i f y a s y m m e t r i c a r m s c o n t r o l b a r g a i n .
●
I f p o l i c y m a k e r s w e r e r a d i c a l l y m o r e l o n g - t e r m i s t , w i t h m o r e f o r e s i g h t a n d p e r c e p t i o n o f t h e d a n g e r s o f
a n u c l e a r a r m s r a c e , o r o t h e r w i s e r a d i c a l l y m o r e c o n c e r n e d a b o u t t h e n u c l e a r a r m s r a c e .
●
I f t h e U . S . c a r r i e d o u t a s u c c e s s f u l p r e v e n t i v e s t r i k e o n t h e S o v i e t U n i o n , t h u s s t a l l i n g t h e S o v i e t a t o m i c
p r o g r a m , o r o t h e r w i s e w a s a b l e t o c o e r c i v e l y s t o p t h e S o v i e t p r o g r a m .
1 8 4
Z u b o k ,
A F a i l e d E m p i r e
, p . 5 1 .
4 3
4 . 9 R i s k y C o o p e r a t i o n
L e s s o n s
W h i l e , i n t h e a b s t r a c t , i n t e r n a t i o n a l c o o p e r a t i o n i s d e s i r a b l e , i n p r a c t i c e , s t e p s f o r c o o p e r a t i o n c a n i n c u r
s u b s t a n t i a l s t r a t e g i c , d i p l o m a t i c , p o l i t i c a l , a n d t e c h n o l o g i c a l l o s s e s . E l i t e s m a y a l s o p e r s o n a l l y l o s e s u p p o r t o r
p o l i t i c a l c a p i t a l b y s u p p o r t i n g i n t e r n a t i o n a l c o n t r o l . A d v o c a t e s o f c o o p e r a t i o n w o u l d d o w e l l t o f u l l y
u n d e r s t a n d t h e p o t e n t i a l r i s k s f r o m c o o p e r a t i o n s o a s t o m i t i g a t e t h e m . R i s k s a r e o f t e n n o t e x p l i c i t l y s t a t e d b y
t h e c o n c e r n e d p a r t i e s , a n d s o m e r i s k s a r e d e e p l y e m b e d d e d i n t h e c i r c u m s t a n c e s , i n s t i t u t i o n s , o r w o r l d v i e w .
1 8 5
H i s t o r i c a l C a s e
R i s k s f o r t h e U n i t e d S t a t e s
E n t e r i n g i n t o o r e v e n d i s c u s s i n g i n t e r n a t i o n a l c o n t r o l w i t h t h e S o v i e t U n i o n c a r r i e d s u b s t a n t i a l s t r a t e g i c a n d
d i p l o m a t i c r i s k s f o r t h e U n i t e d S t a t e s a n d p o l i t i c a l r i s k s f o r t h e i n c u m b e n t a d m i n i s t r a t i o n .
S i m p l y d i s c u s s i n g i n t e r n a t i o n a l c o n t r o l w i t h t h e S o v i e t U n i o n w o u l d l e s s e n t h e d i p l o m a t i c l e v e r a g e t h a t a t o m i c
b o m b s m i g h t b r i n g w h i l s t t h e d i s c u s s i o n s w e r e o n g o i n g .
D i s c u s s i o n s m a y a l s o h a v e r i s k e d r e v e a l i n g
1 8 6
i n f o r m a t i o n a b o u t k e y a s p e c t s o f t h e U . S . a t o m i c p r o g r a m , s u c h a s t h e h i g h l y s e c r e t e f f o r t t o m o n o p o l i z e g l o b a l
h i g h - g r a d e u r a n i u m d e p o s i t s .
N e g o t i a t i o n s c o n s e q u e n t l y c a r r i e d t h e r i s k o f s l o w i n g d o w n t h e U . S . a t o m i c
1 8 7
p r o g r a m a n d p o s s i b l y e v e n a l l o w i n g t h e S o v i e t s t o c a t c h u p .
1 8 8
I n t e r n a t i o n a l c o n t r o l c o u l d a l s o l e a d t o t h e s h a r i n g o f t e c h n i c a l o r s t r a t e g i c i n f o r m a t i o n ( e . g . , h o w f e w b o m b s
t h e U . S . h a d ) t h a t w o u l d a l l o w a n a c c e l e r a t i o n o f t h e S o v i e t b o m b p r o g r a m o r o t h e r s t r a t e g i c a d v a n t a g e o v e r t h e
U n i t e d S t a t e s . S t a l i n h a d d e c i d e d t h a t t h e S o v i e t U n i o n s h o u l d a t t e m p t t o c o p y t h e U . S . b o m b - m a k i n g p r o c e s s ,
s o a n y t e c h n i c a l i n f o r m a t i o n g l e a n e d m a y h a v e b e e n u s e f u l . M o r e o v e r , i n 1 9 4 6 , t h e U . S . S . R . w a s s t r u g g l i n g w i t h
t h e c o n s t r u c t i o n o f i t s f i r s t e x p e r i m e n t a l r e a c t o r a n d t h e l a r g e - s c a l e s e p a r a t i o n o f u r a n i u m , a n d e v e n t h e R u s s i a n
t r a n s l a t i o n o f t h e S m y t h R e p o r t h e l p e d w i t h t h e S o v i e t p r o g r a m i n e a r l y 1 9 4 6 .
1 8 9
T h e r e w e r e a l s o r i s k s a s s o c i a t e d w i t h p u b l i c r e a c t i o n . F i r s t , s t a r t i n g n e g o t i a t i o n s o n i n t e r n a t i o n a l c o n t r o l r a i s e d
p u b l i c e x p e c t a t i o n s t h a t a f a v o r a b l e a g r e e m e n t m a y e v e n t u a l l y b e r e a c h e d . Y e t s u c h e x p e c t a t i o n s m a y n o t b e
m e t , w h i c h w o u l d r e f l e c t u n f a v o r a b l y o n t h e a d m i n i s t r a t i o n . T h a t w a s o n e o f t h e r e a s o n s w h y i t w a s i m p o r t a n t
f o r B a r u c h a n d o t h e r U . S . p o l i c y m a k e r s t o e n s u r e t h a t t h e S o v i e t U n i o n w o u l d b e b l a m e d i f / o n c e i n t e r n a t i o n a l
1 8 5
F o r e x a m p l e , a r e d u c t i o n i n t h e p e r c e i v e d u s e f u l n e s s o f t h e a t o m i c b o m b f o r t h e S o v i e t U n i o n m a y h a v e r e d u c e d t h e
r i s k i n e s s o f t h e c o u n t r y e n t e r i n g i n t o i n t e r n a t i o n a l c o n t r o l . S e e t h e H i s t o r i c a l C a s e d i s c u s s i o n . F o r a t h e o r e t i c a l d i s c u s s i o n
o n r i s k p e r c e p t i o n , s e e H y e - J i n P a e k a n d T h o m a s H o v e , “ R i s k P e r c e p t i o n s a n d R i s k C h a r a c t e r i s t i c s ” ,
O x f o r d R e s e a r c h
E n c y c l o p e d i a s : C o m m u n i c a t i o n
. A v a i l a b l e a t :
h t t p s : / / o x f o r d r e . c o m / c o m m u n i c a t i o n / v i e w / 1 0 . 1 0 9 3 / a c r e f o r e / 9 7 8 0 1 9 0 2 2 8 6 1 3 . 0 0 1 . 0 0 0 1 / a c r e f o r e - 9 7 8 0 1 9 0 2 2 8 6 1 3 - e - 2 8 3
, a c c e s s e d 1 J u n e 2 0 1 9 .
1 8 6
H e r k e n
,
p p . 4 - 8 ; B e r n s t e i n , “ T h e Q u e s t f o r S e c u r i t y ” .
1 8 7
H e r k e n , p . 1 8 8 .
1 8 8
T h i s c o n c e r n w a s e n c a p s u l a t e d i n t h e d e b a t e o v e r s t a g e s , s e e H e r k e n , p p . 1 5 7 .
1 8 9
G o r d i n ,
R e d C l o u d a t D a w n ,
p . 1 5 2 . H o l l o w a y ,
S t a l i n a n d t h e B o m b
, p p . 1 7 8 - 9 .
4 4
c o n t r o l n e g o t i a t i o n s f a i l e d .
S e c o n d , i n t e r n a t i o n a l c o n t r o l d i s c u s s i o n s r i s k e d f u r t h e r s e n s i t i z i n g t h e U . S . p u b l i c
1 9 0
t o t h e d e s t r u c t i v e n e s s o f t h e a t o m i c b o m b , m a k i n g i t h a r d e r t o m o b i l i z e t h e p u b l i c f o r m i l i t a r y a c t i o n .
1 9 1
T h e r e w e r e a l s o r i s k s a s s o c i a t e d w i t h c o o p e r a t i n g w i t h a l l i e s . D o i n g s o r i s k e d i n c r e a s e d l e a k s t o t h e S o v i e t
U n i o n , a n d i n d e e d , B r i t i s h s c i e n t i s t s s u c h a s K l a u s F u c h s d i d p a s s o n i n f o r m a t i o n f r o m t h e M a n h a t t a n P r o j e c t
t o t h e S o v i e t U n i o n .
T h e U . S . a l s o r i s k e d o f f e n d i n g a l l i e s b y d i r e c t b a r g a i n i n g w i t h S o v i e t U n i o n . F o r
1 9 2
e x a m p l e , t h e B r i t i s h w i s h e d t o b e i n c l u d e d i n a n y n e g o t i a t i o n s o n i n t e r n a t i o n a l c o n t r o l . W h e n S e c r e t a r y o f S t a t e
B y r n e s d i d d i r e c t l y b a r g a i n w i t h t h e S o v i e t U n i o n i n l a t e 1 9 4 5 , B r i t a i n a n d e s p e c i a l l y F r a n c e w e r e o f f e n d e d , a n d
i t h a r m e d t h e a l l i a n c e .
1 9 3
R i s k s f o r t h e S o v i e t U n i o n
N e g o t i a t i n g o r s t a r t i n g t h e p r o c e s s o f i n t e r n a t i o n a l c o n t r o l a l s o c a r r i e d s i g n i f i c a n t r i s k s f o r t h e S o v i e t U n i o n .
I n t e r n a t i o n a l i n s p e c t i o n s c o u l d r e v e a l r a w m a t e r i a l s a n d f a c i l i t i e s , a n d f a c i l i t a t e a p r e v e n t i v e a t t a c k o n t h e S o v i e t
p r o g r a m .
I n s p e c t i o n s a n d o p e n n e s s c o u l d u n d e r m i n e r e g i m e s t a b i l i t y . U . S . p r o p o s a l s a d v o c a t e d a g r a d u a l
1 9 4
s t a g e d p r o c e s s f o r t h e i n s t i t u t i o n o f i n t e r n a t i o n a l c o n t r o l . F r o m t h e S o v i e t p o i n t o f v i e w , t h i s g a v e t h e U . S . a n
a d v a n t a g e i n t h e e a r l i e r s t a g e s o f i n t e r n a t i o n a l c o n t r o l . T h e B u s h P l a n , f o r e x a m p l e , s u g g e s t e d t h r e e s t a g e s : ( 1 )
b a s i c i n f o r m a t i o n s h a r i n g , ( 2 ) i n s p e c t i o n s ( a t w h i c h p o i n t e a c h c o u n t r y w o u l d r e v e a l i t s a t o m i c f a c i l i t i e s a n d
r e s o u r c e s ) , a n d ( 3 ) t r a n s f e r o f r e s o u r c e s a n d m a t e r i a l .
T h e A c h e s o n - L i l i e n t h a l P l a n w a s m o r e n u a n c e d . B u t
1 9 5
t h a t t o o s t i p u l a t e d t h a t t h e U . S . a n d t h e S o v i e t U n i o n w o u l d g i v e u p i n f o r m a t i o n o n t h e i r a t o m i c f a c i l i t i e s i n
g r a d u a l s t a g e s . S i m i l a r l y , t h e f o r m a l h a n d o v e r o f a t o m i c f a c i l i t i e s t o t h e U . N . w o u l d o c c u r i n s t a g e s . C r u c i a l l y ,
t h e h a n d o v e r o f a t o m i c b o m b s t h e m s e l v e s w o u l d o n l y o c c u r a t t h e e n d . T h e p l a n d i d n o t s t i p u l a t e w h e n t h e
U . S . w o u l d s t o p m a n u f a c t u r i n g b o m b s .
T h e f i r s t d e t a i l e d d e n u n c i a t i o n o f t h e B a r u c h P l a n i n t h e S o v i e t p r e s s
1 9 6
f o c u s e d o n t h e i s s u e o f s t a g e s . W h y , a s k e d
P r a v d a
, d i d t h e p l a n a l l o w t h e U . S . t o c o n t i n u e m a k i n g a t o m i c
b o m b s t h r o u g h o u t m o s t o f t h e i n t e r n a t i o n a l c o n t r o l i n s t i t u t i o n p r o c e s s ?
1 9 7
T h e U . N . a n d p a r t i c u l a r l y t h e S e c u r i t y C o u n c i l w a s d o m i n a t e d b y t h e U . S . a n d i t s a l l i e s , a n d s o w e r e a l s o
s u s p e c t t o t h e S o v i e t s . A n y c o n t r o l l i n g i n t e r n a t i o n a l o r g a n i z a t i o n , n o t e d t h e S o v i e t p h y s i c i s t ( a n d a d v i s o r t o t h e
S o v i e t d e l e g a t i o n a t t h e U N A E C ) D . V . S k o b e l ’ t s y n i n 1 9 4 6 , w o u l d b e “ i n r e a l i t y , p r o b a b l y … A m e r i c a n . ”
1 9 8
1 9 0
C r a i g a n d R a d c h e n k o ,
T h e A t o m i c B o m b a n d t h e O r i g i n s o f t h e C o l d W a r
, p . 1 3 0 .
1 9 1
T h e p u b l i c w a s a l r e a d y w a r w e a r y a n d s o m e w h a t f e a r f u l o f a t o m i c w a r . H e r k e n , p p . 2 1 4 - 1 6 . B o y e r ,
B y t h e B o m b ’ s E a r l y
L i g h t ,
p p . 1 3 - 1 5 .
1 9 2
H e r k e n , p . 2 6 5 . R o b e r t C h a d w e l l W i l l i a m s ,
K l a u s F u c h s : A t o m S p y
( C a m b r i d g e , M A : H a r v a r d U n i v e r s i t y P r e s s , 1 9 8 7 ) .
1 9 3
S u s a n n a S c h r a f s t e t t e r , “ ‘ L o q u a c i o u s … a n d p o i n t l e s s a s e v e r ’ ? B r i t a i n , t h e U n i t e d S t a t e s a n d t h e U n i t e d N a t i o n s
N e g o t i a t i o n s o n I n t e r n a t i o n a l C o n t r o l o f N u c l e a r E n e r g y , 1 9 4 5 - 4 8 ” ,
C o n t e m p o r a r y B r i t i s h H i s t o r y
1 6 , 4 ( W i n t e r 2 0 0 2 ) ,
8 7 - 1 0 8 .
1 9 4
O n S o v i e t c o n c e r n s , s e e D a v i d H o l l o w a y , “ T h e S o v i e t U n i o n a n d t h e c r e a t i o n o f t h e I n t e r n a t i o n a l A t o m i c E n e r g y
A g e n c y ” ,
C o l d W a r H i s t o r y
1 6 , 2 ( 2 0 1 6 ) , p p . 1 7 7 - 9 3 . T h i s c o n c e r n w a s a l s o v o i c e d b y S e c r e t a r y o f C o m m e r c e H e n r y
W a l l a c e t o T r u m a n i n J u l y 1 9 4 6 , s e e H e r k e n , p . 1 8 1 .
1 9 5
U n i t e d S t a t e s D e p a r t m e n t o f S t a t e ,
F o r e i g n R e l a t i o n s o f t h e U n i t e d S t a t e s : D i p l o m a t i c P a p e r s , 1 9 4 5 . G e n e r a l : P o l i t i c a l
a n d E c o n o m i c M a t t e r s
V o l u m e I I ( W a s h i n g t o n , D C : G o v e r n m e n t P r i n t i n g O f f i c e , 1 9 4 6 ) , p p . 6 9 - 7 3 . A v a i l a b l e a t :
h t t p : / / d i g i c o l l . l i b r a r y . w i s c . e d u / c g i - b i n / F R U S / F R U S - i d x ? t y p e = a r t i c l e & d i d = F R U S . F R U S 1 9 4 5 v 0 2 . i 0 0 0 5 & i d = F R U S . F R
U S 1 9 4 5 v 0 2 & i s i z e = M
. A c c e s s e d 2 5 A p r i l 2 0 1 9 .
1 9 6
T h a t , f o r e x a m p l e , w a s W a l l a c e ’ s c r i t i c i s m o f t h e B a r u c h P l a n : H e r k e n , p . 1 8 1 .
1 9 7
L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a
, p . 3 1 2 .
1 9 8
H o l l o w a y ,
S t a l i n a n d t h e B o m b
, p . 1 6 4 .
4 5
S u c h a n o r g a n i z a t i o n , a n d s o i n t e r n a t i o n a l c o n t r o l i t s e l f , c o u l d b e u s e d a g a i n s t t h e S o v i e t U n i o n . T h i s c o n c e r n
c a n b e s e e n i n t h e S o v i e t i n s i s t e n c e t h a t ( 1 ) t h e U N A E C r e p o r t t o t h e U . N . S e c u r i t y C o u n c i l a n d n o t t h e
G e n e r a l A s s e m b l y ( w h i c h w a s p e r c e i v e d t o b e e v e n m o r e b i a s e d t o w a r d s t h e U . S . ) a n d ( 2 ) t h a t t h e v e t o a p p l y t o
t h e d e l i b e r a t i o n s o f t h e U N A E C a n d t o i n t e r n a t i o n a l c o n t r o l .
1 9 9
L a s t l y , b o t h t h e U . S . a n d S o v i e t U n i o n c o u l d l o s e b a r g a i n i n g l e v e r a g e b y o f f e r i n g c o o p e r a t i o n f i r s t . A c c o r d i n g
t o h i s t o r i a n J o h n L e w i s G a d d i s , f o r e x a m p l e , T r u m a n l o s t b a r g a i n i n g a d v a n t a g e b y u n i l a t e r a l l y c o m m i t t i n g t h e
U . S . t o i n t e r n a t i o n a l c o n t r o l i n l a t e 1 9 4 5 .
2 0 0
1 9 9
H e r k e n , p p . 8 4 , 1 7 4 .
2 0 0
J o h n L e w i s G a d d i s ,
S t r a t e g i e s o f C o n t a i n m e n t : A C r i t i c a l A p p r a i s a l o f A m e r i c a n N a t i o n a l S e c u r i t y P o l i c y d u r i n g t h e
C o l d W a r
2 n d e d . ( O x f o r d : O x f o r d U n i v e r s i t y P r e s s , 2 0 0 5 ) , p . 1 7 .
4 6
4 . 1 0 P r e v e n t i v e S t r i k e
L e s s o n
E v e n t h e m o s t v i o l e n t o f s o l u t i o n s , s u c h a s a p r e v e n t i v e w a r o r a p r e v e n t i v e s t r i k e , m a y g a i n t r a c t i o n .
H i s t o r i c a l C a s e
D u r i n g t h e y e a r s o f t h e U . S . a t o m i c m o n o p o l y , 1 9 4 5 – 1 9 4 9 , m a n y i n t h e U . S . a n d B r i t a i n f a v o r e d r e s o r t i n g t o
f o r c e t o p r e v e n t t h e S o v i e t U n i o n f r o m o b t a i n i n g a t o m i c w e a p o n s . G e n e r a l s s u c h a s H e n r y H . “ H a p ” A r n o l d ,
C a r l S p a a t z , a n d C u r t i s L e M a y a r g u e d t h i s , a s d i d i n t e r n a t i o n a l i s t s s u c h a s E l y C u l b e r t s o n .
H a w k i s h
2 0 1
i n t e l l e c t u a l s p u b l i s h e d b o o k s s u g g e s t i n g p r e v e n t i v e w a r . I n
I f R u s s i a S t r i k e s
( 1 9 4 9 ) , G e o r g e E l i o t , a l e a d i n g
m i l i t a r y a n a l y s t , c a l l e d f o r t h e U . S . t o p r e s e n t a n u l t i m a t u m t o M o s c o w : a c c e p t t h e B a r u c h P l a n o r t h e U . S .
w o u l d u s e a t o m i c b o m b s a g a i n s t S o v i e t a t o m i c f a c i l i t i e s . P o l i t i c a l s c i e n t i s t J a m e s B u r n h a m , i n h i s 1 9 4 7
S t r u g g l e
f o r t h e W o r l d
,
c a l l e d f o r p o l i t i c a l s u b v e r s i o n t o d e s t r o y t h e C o m m u n i s t s t a t e , o r i f t h a t f a i l e d , t h e n a i r s t r i k e s o n
S o v i e t m i l i t a r y t a r g e t s .
N e w Y o r k H e r a l d T r i b u n e
j o u r n a l i s t s J o s e p h a n d S t e w a r t A l s o p c a l l e d f o r p r e v e n t i v e w a r
i n t h e i r c o l u m n s .
E v e n p a c i f i s t a n d s o c i a l i s t - f r i e n d l y B e r t r a n d R u s s e l l c a l l e d f o r t h e U . S . t o t h r e a t e n w a r a s
2 0 2
“ p a r t o f a p l a n h e h a d d e v e l o p e d t o p r o m o t e g l o b a l p e a c e . ”
A s r e c o u n t e d b y a m e m b e r o f t h e a u d i e n c e , i n o n e
2 0 3
s p e e c h R u s s e l l a r g u e d t h a t :
T h e S o v i e t U n i o n d i d n o t y e t p o s s e s s a n u c l e a r c a p a b i l i t y b u t t h a t i t w o u l d v e r y s o o n d o s o , a f t e r w h i c h
a l l h i s t o r y m a d e i t c l e a r t h a t s o o n e r o r l a t e r t h e r e w o u l d b e a w a r b e t w e e n t h e t w o s u p e r p o w e r s t h a t
w o u l d b e i n f i n i t e l y m o r e d e v a s t a t i n g t h a n e i t h e r o f t h e t w o w o r l d w a r s t h r o u g h w h i c h h e h a d l i v e d .
T h e o n l y w a y o f p r e v e n t i n g t h i s A r m a g e d d o n , h e c o n c l u d e d w i t h r e m o r s e l e s s i f u n p a l a t a b l e l o g i c , w a s
f o r A m e r i c a t o l a u n c h a n u c l e a r a t t a c k o n t h e S o v i e t U n i o n b e f o r e i t a c q u i r e d t h e b o m b : a f t e r t h a t i t
w o u l d b e t o o l a t e .
2 0 4
T h e f o l l o w i n g a r e , i n o u r e s t i m a t i o n , t h e r e a s o n s w h y a p r e e m p t i v e s t r i k e o r a p r e v e n t i v e w a r w a s n o t l a u n c h e d
b y t h e U . S . i n t h e l a t e 4 0 s .
I n t e n t a n d A p p e t i t e
T h e r e w a s n o p u b l i c a p p e t i t e f o r a n o t h e r m a j o r w a r d u e t o s i g n i f i c a n t w a r w e a r i n e s s . T h e p u b l i c w a n t e d
w i d e - s c a l e d e m o b i l i z a t i o n a n d e x p r e s s e d c o n c e r n a t c o n t i n u e d o v e r s e a s d e p l o y m e n t s . T r u m a n a n d o t h e r
2 0 1
R u s s e l l D . B u h i t e a n d W M . C h r i s t o p h e r H a m e l , “ W a r f o r P e a c e : T h e Q u e s t i o n o f a n A m e r i c a n P r e v e n t i v e W a r a g a i n s t
t h e S o v i e t U n i o n , 1 9 4 5 – 1 9 5 5 ” ,
D i p l o m a t i c H i s t o r y
1 4 , 3 ( 1 9 9 0 ) , p p . 3 6 7 – 8 4 . M e l v y n P . L e f f l e r , “ S t r a t e g y , D i p l o m a c y , a n d
t h e C o l d W a r : T h e U n i t e d S t a t e s , T u r k e y , a n d N A T O , 1 9 4 5 - 1 9 5 2 ” ,
T h e J o u r n a l o f A m e r i c a n H i s t o r y
7 1 , 4 ( M a r c h 1 9 8 5 ) ,
p p . 8 0 7 - 2 5 . S e c r e t a r y o f D e f e n s e J a m e s F o r r e s t a l w o u l d w r i t e i n 1 9 4 7 t h a t “ T h e y e a r s b e f o r e a n y p o s s i b l e p o w e r c a n
a c h i e v e t h e c a p a b i l i t y e f f e c t i v e l y t o a t t a c k u s w i t h w e a p o n s o f m a s s d e s t r u c t i o n a r e o u r y e a r s o f o p p o r t u n i t y . ” Q u o t e d i n
J o h n L e w i s G a d d i s ,
S t r a t e g i e s o f C o n t a i n m e n t : A C r i t i c a l A p p r a i s a l o f P o s t w a r A m e r i c a n N a t i o n a l S e c u r i t y P o l i c y
( N e w
Y o r k : O x f o r d U n i v e r s i t y P r e s s , 1 9 8 2 ) , p . 6 2 . S e e a l s o G e o r g e H . Q u e s t e r ,
N u c l e a r M o n o p o l y
( N e w B r u n s w i c k : T r a n s a c t i o n
P u b l i s h e r s , 2 0 0 0 ) , c h a p t e r 4 .
2 0 2
B u h i t e a n d H a m e l , “ W a r f o r P e a c e ” .
2 0 3
D a v i d B l i t z , “ D i d R u s s e l l A d v o c a t e P r e v e n t i v e A t o m i c W a r A g a i n s t t h e U S S R ? ” ,
T h e J o u r n a l o f B e r t r a n d R u s s e l l
S t u d i e s
2 2 ( s u m m e r 2 0 0 2 ) , p p . 5 - 4 5 .
2 0 4
Q u o t e d i n B l i t z , “ D i d R u s s e l l A d v o c a t e P r e v e n t i v e A t o m i c W a r A g a i n s t t h e U S S R ? ” .
4 7
p o l i t i c i a n s r e s p o n d e d t o t h i s p u b l i c p r e s s u r e .
A n o v e r t a c t o f a g g r e s s i o n w o u l d n o t h a v e p l a y e d w e l l w i t h
2 0 5
p u b l i c o p i n i o n a n d w o u l d h a v e b e e n d i s a g r e e a b l e t o s o m e p o l i c y m a k e r s w h o s a w i t a s b e i n g a g a i n s t A m e r i c a n
p r i n c i p l e s .
I t w a s o n l y a f t e r 1 9 5 0 t h a t s u p p o r t f o r a p r e v e n t i v e w a r a g a i n s t t h e S o v i e t U n i o n b e g a n t o g r o w
2 0 6
a p p r e c i a b l y a m o n g s t t h e U . S . p u b l i c . T h i s , i n t u r n , s p u r r e d t a l k o f p r e v e n t i v e w a r a m o n g s t p o l i c y m a k e r s . “ F o r
t h e f i r s t t i m e , ” n o t e d
N e w s w e e k
i n F e b r u a r y 1 9 5 0 , “ s o m e m e m b e r s o f C o n g r e s s w e r e b e g i n n i n g t o s p e c u l a t e o n
w h a t h a d f o r m e r l y b e e n a n a l m o s t f o r b i d d e n s u b j e c t - p r e v e n t i v e w a r . ”
2 0 7
A l t h o u g h m a n y p o t e n t i a l s u p p o r t e r s o f a p r e v e n t i v e s t r i k e w e r e h a w k s , t h e r e w e r e a l s o p o l i c y m a k e r s w h o w e r e
i n s t i n c t i v e l y m u c h l e s s h o s t i l e t o t h e S o v i e t U n i o n . I n 1 9 5 0 a n d 1 9 5 1 , f o r e x a m p l e , w h e n a p r e v e n t i v e w a r w a s
m o r e o p e n l y d i s c u s s e d i n p o l i c y m a k i n g c i r c l e s , k e y p o l i c y m a k e r s s u c h a s S e c r e t a r y o f S t a t e D e a n A c h e s o n
e x p r e s s e d s i g n i f i c a n t c o n c e r n .
H a w k i s h a t t i t u d e s m a y h a v e b e e n d a m p e n e d b y t h e f a c t t h a t t h e U . S . m i l i t a r y
2 0 8
a n d m a n y i n g o v e r n m e n t w e r e o v e r c o n f i d e n t a b o u t t h e i r a t o m i c l e a d o v e r t h e S o v i e t U n i o n a n d h a d
u n d e r e s t i m a t e d t h e p r o g r e s s b e i n g m a d e b y t h e S o v i e t p r o g r a m , p a r t i c u l a r l y t h e S o v i e t U n i o n ’ s a b i l i t y t o
a c q u i r e h i g h - g r a d e u r a n i u m .
M a n y p o l i c y m a k e r s w e r e s o c o n f i d e n t o f t h e U . S . l e a d t h a t t h e y e v e n o p p o s e d
2 0 9
t h e d e v e l o p m e n t o f a p r o g r a m f o r d e t e c t i n g S o v i e t n u c l e a r t e s t s .
T h e U . S . g o v e r n m e n t , i n f a c t , h a d l i t t l e
2 1 0
i n f o r m a t i o n a b o u t t h e S o v i e t b o m b p r o g r a m . I f i t h a d h a d c r e d i b l e i n f o r m a t i o n a b o u t t h e p r o g r e s s o f t h e
p r o g r a m , t h i s c o u l d h a v e a c t e d a s a f o c u s i n g e v e n t f o r a w a r .
2 1 1
H a w k i s h a t t i t u d e s m a y a l s o h a v e b e e n d a m p e n e d b y t h e b e l i e f , h e l d b y m a n y i n t h e U . S . a d m i n i s t r a t i o n , t h a t
S o v i e t e x p a n s i o n i s m ( m o r e a c o n c e r n t h a n t h e S o v i e t a t o m i c p r o g r a m i n t h e 4 0 s ) c o u l d b e c o n t a i n e d t h r o u g h
d i p l o m a c y , a l l i a n c e b u i l d i n g , a n d i n i t i a t i v e s s u c h a s t h e M a r s h a l l P l a n . T h e r e w a s a g e n e r a l c o n s e n s u s i n t h e U . S .
m i l i t a r y e s t a b l i s h m e n t i n t h e l a t e 4 0 s t h a t t h e S o v i e t U n i o n w i s h e d t o a v o i d m i l i t a r y e n g a g e m e n t s .
2 1 2
T a r g e t e d s t r i k e s m a y h a v e b e e n u n p a l a t a b l e b e c a u s e t h e y w o u l d p r o b a b l y h a v e l e d t o g e n e r a l w a r , w h i c h w o u l d
b e v e r y c o s t l y : E u r o p e a n d p a r t s o f A s i a w o u l d l i k e l y h a v e b e e n i n v a d e d a n d o c c u p i e d b y t h e S o v i e t U n i o n f o r a t
2 0 5
H e r k e n , p p . 2 1 4 - 1 6 ; L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a
, p p . 2 2 7 - 3 0 ; H a r r y S . T r u m a n ,
T h e M e m o i r s o f H a r r y
S . T r u m a n
v o l u m e 1 ( N e w Y o r k : G a r d e n C i t y , 1 9 5 5 ) , p p . 4 4 3 , 4 5 3 .
2 0 6
B u h i t e a n d H a m e l , “ W a r f o r P e a c e ” ; H e r k e n , p . 2 2 3 . O n w a r w e a r i n e s s a n d p u b l i c p r e s s u r e , s e e R . A l t o n L e e , “ T h e
A r m y ‘ M u t i n y ’ o f 1 9 4 6 ” ,
T h e A m e r i c a n J o u r n a l o f A m e r i c a n H i s t o r y
5 3 , 3 ( D e c e m b e r 1 9 6 6 ) , p p . 5 5 5 - 7 1 .
2 0 7
S t e v e n C a s e y , “ S e l l i n g N S C - 6 8 : T h e T r u m a n A d m i n i s t r a t i o n , P u b l i c O p i n i o n , a n d t h e P o l i t i c s o f M o b i l i z a t i o n ,
1 9 5 0 – 5 1 ” ,
D i p l o m a t i c H i s t o r y
2 9 , 4 ( S e p t e m b e r 2 0 0 5 ) , p p . 6 5 5 - 9 0 . A l s o : “ S i g n i f i c a n c e o f t h e H - B o m b , a n d A m e r i c a ’ s
D i l e m m a ” ,
N e w s w e e k
( 1 2 F e b r u a r y 1 9 5 0 ) , p . 2 0 , c i t e d i n C a s e y , “ S e l l i n g N S C - 6 8 ” .
2 0 8
T i m o t h y J . B o t t i ,
A c e i n t h e H o l e : W h y t h e U n i t e d S t a t e s D i d N o t U s e N u c l e a r W e a p o n s i n t h e C o l d W a r , 1 9 4 5 t o 1 9 6 5
( W e s t p o r t , C T : G r e e n w o o d P r e s s , 1 9 9 6 ) , p p . 2 8 - 9 . O n p r e v e n t i v e w a r d i s c u s s i o n s m o r e b r o a d l y , s e e M a r c T r a c h t e n b e r g ,
H i s t o r y a n d S t r a t e g y
( P r i n c e t o n , N J : P r i n c e t o n U n i v e r s i t y P r e s s , 1 9 9 1 ) , p p . 1 0 3 - 7 .
2 0 9
T h e P i n c h e r w a r p l a n , f o r e x a m p l e , a s s u m e d t h a t t h e U . S . w o u l d m o s t p r o b a b l y r e t a i n a n a t o m i c m o n o p o l y u n t i l 1 9 5 6 :
H e r k e n , p . 2 2 9 . A l s o : L a w r e n c e F r e e d m a n ,
T h e E v o l u t i o n o f N u c l e a r S t r a t e g y
3 r d e d . ( L o n d o n : P a l g r a v e M a c m i l l a n , 2 0 0 3 ) ,
p . 2 6 .
2 1 0
H e r k e n , p . 3 0 0 .
2 1 1
I n 1 9 5 3 f o r e x a m p l e , T r u m a n b r i e f l y c o n s i d e r e d a p r e v e n t i v e s t r i k e a f t e r h e a r i n g o f t h e f i r s t S o v i e t h y d r o g e n b o m b t e s t .
B u h i t e a n d H a m e l , “ W a r f o r P e a c e ” .
2 1 2
B u h i t e a n d H a m e l , “ W a r f o r P e a c e ” . M e l v y n P . L e f f l e r , “ T h e A m e r i c a n C o n c e p t i o n o f N a t i o n a l S e c u r i t y a n d t h e
B e g i n n i n g s o f t h e C o l d W a r , 1 9 4 5 - 4 8 ” ,
T h e A m e r i c a n H i s t o r i c a l R e v i e w
8 9 , 2 ( A p r i l 1 9 8 4 ) , p p . 3 4 6 - 8 1 .
4 8
l e a s t a y e a r , a n d i t w o u l d h a v e l e d t o m i l l i o n s o f S o v i e t d e a t h s , h u n d r e d s o f t h o u s a n d s i f n o t m i l l i o n s o f U . S .
a l l i e d d e a t h s , a n d t e n s o f t h o u s a n d s o f U . S . c o m b a t d e a t h s .
2 1 3
M i l i t a r y C a p a b i l i t i e s
F o l l o w i n g d e m o b i l i z a t i o n , U . S . a r m e d f o r c e s w e r e i n a d e q u a t e f o r a m i l i t a r y d e f e a t o f t h e S o v i e t U n i o n , a n d t h e
U . S . w a s w e l l a w a r e o f t h i s . S o v i e t a i r f o r c e s , r a d a r , a n d a n t i - a i r c r a f t g u n s c o n t i n u e d t o i m p r o v e t h r o u g h t o t h e
f i f t i e s .
I t i s a l s o u n c l e a r w h e t h e r t h e U . S . h a d t h e c a p a b i l i t i e s t o c a r r y o u t a s u f f i c i e n t l y e f f e c t i v e p r e v e n t i v e
2 1 4
s t r i k e . S o v i e t a t o m i c f a c i l i t i e s w e r e w i d e s p r e a d a n d n o t e a s i l y a t t a c k e d b y t h e U n i t e d S t a t e s . L a c k o f i n t e l l i g e n c e
m e a n t t h a t U . S . m i l i t a r y p l a n n e r s h a d i m m e n s e d i f f i c u l t y s e l e c t i n g t a r g e t s f o r a e r i a l a t t a c k ( c o n v e n t i o n a l o r
a t o m i c ) i n t h e l a t e 4 0 s .
T h e U . S . h a d a p o o r l e v e l o f n u c l e a r r e a d i n e s s . T h e y h a d v e r y f e w a t o m i c b o m b s : b y
2 1 5
t h e e n d o f 1 9 4 7 , t h e U . S . o n l y h a d 1 3 , a n d b y 1 9 4 8 , o n l y 5 0 . T h e r e w e r e a l s o i s s u e s w i t h b o m b a s s e m b l y a n d
d e l i v e r y c a p a b i l i t i e s .
2 1 6
C o u l d T h e r e H a v e B e e n a P r e v e n t i v e S t r i k e ?
C o u n t e r f a c t u a l l y , t h e n , a p r e v e n t i v e s t r i k e w o u l d h a v e b e c o m e a r e a l i s t i c o p t i o n i f a c e r t a i n n u m b e r o f f a c t o r s
h a d b e e n p r e s e n t : a m o n g s t p o l i c y m a k e r s , a m o r e a l a r m i s t a s s e s s m e n t o f t h e S o v i e t U n i o n ’ s a t o m i c p r o g r a m a n d
i t s p r o g r e s s , a n d b e t t e r i n t e l l i g e n c e a b o u t S o v i e t a t o m i c f a c i l i t i e s . F o r e x a m p l e , m o r e s e r i o u s c r i s e s i n T u r k e y a n d
t h e M i d d l e E a s t , t h e n o f s i g n i f i c a n t c o n c e r n f o r t h e U . S . , m a y a l s o h a v e h e l p e d m a k e a s t r o n g e r c a s e f o r a s t r i k e
a g a i n s t S o v i e t a t o m i c f a c i l i t i e s .
I n 1 9 5 3 , f o l l o w i n g t h e S o v i e t d e t o n a t i o n o f a n H - b o m b , T r u m a n b r i e f l y
2 1 7
t h o u g h t a b o u t a p r e v e n t i v e s t r i k e . W i t h t h e a p p r o p r i a t e i n t e l l i g e n c e , a n d i f t h e c i r c u m s t a n c e s h a d b e e n
f a v o r a b l e , h e m a y h a v e c o n s i d e r e d a s t r i k e e a r l i e r , p r i o r t o t h e S o v i e t a t o m i c b o m b t e s t i n 1 9 4 9 .
F a c t o r s s u c h
2 1 8
2 1 3
O n t h e
P i n c h e r
p l a n s f o r a g e n e r a l w a r i n E u r o p e a n d t h e s e c r u d e f a t a l i t y e s t i m a t e s , s e e R o s s ,
A m e r i c a n W a r P l a n s
, p p .
2 5 - 5 2
.
U n f o r t u n a t e l y , w e d o n o t h a v e a c l e a r r e f e r e n c e f o r t h e e s t i m a t e d f a t a l i t i e s f r o m d i f f e r e n t w a r p l a n s .
2 1 4
A l t h o u g h t h e S o v i e t U n i o n h a d d e m o b i l i z e d a s w e l l , i t s t i l l h a d s u f f i c i e n t g r o u n d f o r c e s t o d e f e n d i t s e l f f r o m a U . S .
a t t a c k . T h e U n i t e d S t a t e s w a s a w a r e o f t h i s : a J o i n t I n t e l l i g e n c e C o m m i t t e e o f t h e J o i n t C h i e f s o f S t a f f n o t e d i n N o v e m b e r
1 9 4 5 t h a t i n a w a r i n E u r o p e o r m a i n l a n d A s i a , “ t h e S o v i e t s w o u l d e n j o y a g r e a t p r e p o n d e r a n c e i n n u m b e r s o f m e n a g a i n s t
t h e U n i t e d S t a t e s o r e v e n a g a i n s t t h e U n i t e d S t a t e s , G r e a t B r i t a i n a n d F r a n c e . ” Q u o t e d i n : H o l l o w a y ,
S t a l i n a n d t h e B o m b
,
p . 2 3 1 . N S C 1 0 0 , i s s u e d i n J a n u a r y 1 9 5 1 , e s t i m a t e d t h a t U . S . m i l i t a r y s t r e n g t h w a s 1 0 % o f t h a t o f t h e S o v i e t U n i o n a t t h a t
t i m e , t h o u g h c o u l d b e i n c r e a s e d t o 2 0 % i n t h e n e a r f u t u r e . N S C 1 0 0 , " R e c o m m e n d e d P o l i c i e s a n d A c t i o n s i n L i g h t o f t h e
G r a v e W o r l d S i t u a t i o n ” , 1 1 J a n u a r y 1 9 5 1 , a v a i l a b l e a t :
h t t p s : / / w w w . m t h o l y o k e . e d u / a c a d / i n t r e l / k o r e a / n s c 1 0 0 . h t m
.
A c c e s s e d 2 0 A p r i l 2 0 1 9 . T h e R e d A r m y ’ s 5 0 0 d i v i s i o n s a t t h e e n d o f t h e w a r w e r e r e d u c e d t o 1 7 5 t h r o u g h d e m o b i l i z a t i o n ,
b u t t h e s e r e m a i n i n g d i v i s i o n s w e r e s t r e n g t h e n e d w i t h g r e a t e r f i r e p o w e r a n d m o b i l i t y : H o l l o w a y ,
S t a l i n a n d t h e B o m b
,
p . 2 3 1 . M o r e d e t a i l s o n S o v i e t p o s t w a r a r m y s t r e n g t h a r e i n M a t t h e w A . E v a n g e l i s t a , “ S t a l i n ' s P o s t w a r A r m y R e a p p r a i s e d ” ,
I n t e r n a t i o n a l S e c u r i t y
7 , 3 ( W i n t e r 1 9 8 2 - 1 9 8 3 ) , p p . 1 1 0 - 3 8 . T h e W e s t , b y c o n t r a s t , h a d 3 7 5 , 0 0 0 o c c u p y i n g s o l d i e r s i n
G e r m a n y a n d A u s t r i a , a n d a n o t h e r 4 0 0 , 0 0 0 i n W e s t e r n E u r o p e ( e x c l u d i n g B r i t a i n ) . S e e H o l l o w a y ,
S t a l i n a n d t h e B o m b
,
p . 2 3 2 . T h e r e w e r e a l s o r a p i d i m p r o v e m e n t s i n e a r l y w a r n i n g r a d a r , a n t i - a i r c r a f t g u n s , j e t f i g h t e r s , a n d s t r a t e g i c b o m b e r s
( i n c l u d i n g t h e T u - 4 ) i n t h e l a t e f o r t i e s . H o l l o w a y ,
S t a l i n a n d t h e B o m b
, p p . 2 3 5 - 6 , 2 4 3 ; B u h i t e a n d H a m e l , “ W a r f o r
P e a c e ” .
2 1 5
K e n Y o u n g , “ U S ' A t o m i c C a p a b i l i t y ' a n d t h e B r i t i s h F o r w a r d B a s e s i n t h e E a r l y C o l d W a r ” ,
J o u r n a l o f C o n t e m p o r a r y
H i s t o r y
4 2 , 1 ( J a n u a r y 2 0 0 7 ) , p p . 1 1 7 - 3 6 .
2 1 6
R o b e r t S . N o r r i s a n d H a n s M . K r i s t e n s e n , “ G l o b a l N u c l e a r W e a p o n s I n v e n t o r i e s , 1 9 4 5 – 2 0 1 0 ” ,
B u l l e t i n o f t h e A t o m i c
S c i e n t i s t s
6 6 , 4 ( 2 0 1 0 ) , p p . 7 7 - 8 3 . D a v i d A l a n R o s e n b e r g , “ A m e r i c a n A t o m i c S t r a t e g y a n d t h e H y d r o g e n B o m b D e c i s i o n ” ,
T h e J o u r n a l o f A m e r i c a n H i s t o r y
6 6 , 1 ( J u n e 1 9 7 9 ) , p p . 6 2 - 8 7 .
2 1 7
O n U . S . c o n c e r n s o v e r S o v i e t e x p a n s i o n i s m i n A s i a , s e e M e l v y n P . L e f f l e r , “ S t r a t e g y , D i p l o m a c y , a n d t h e C o l d W a r :
T h e U n i t e d S t a t e s , T u r k e y , a n d N A T O , 1 9 4 5 - 1 9 5 2 ” ,
T h e J o u r n a l o f A m e r i c a n H i s t o r y
7 1 , 4 ( M a r c h 1 9 8 5 ) , p p . 8 0 7 - 2 5 .
2 1 8
B u h i t e a n d H a m e l , “ W a r f o r P e a c e ” .
4 9
a s l e s s d e m o b i l i z a t i o n o f m i l i t a r y f o r c e s a n d l e s s w a r w e a r i n e s s a m o n g s t t h e p u b l i c m a y a l s o h a v e i n c r e a s e d t h e
c h a n c e s o f a p r e v e n t i v e s t r i k e a s a r e a l i s t i c o p t i o n .
5 0
5 . C o n c l u s i o n a n d E x t e n s i o n s
O u r s t u d y o f a t t e m p t s a t t h e i n t e r n a t i o n a l c o n t r o l o f a t o m i c e n e r g y i n 1 9 4 5 – 4 6 s u g g e s t s t h a t r a d i c a l s c h e m e s
f o r i n t e r n a t i o n a l g o v e r n a n c e c a n g e t w i d e s p r e a d s u p p o r t , e v e n f r o m s k e p t i c s , b u t t h a t t h e s u p p o r t c a n b e
t e n u o u s a n d f l e e t i n g . T e c h n i c a l e x p e r t s c a n b o l s t e r s u p p o r t , b u t m u d d l e d p o l i c y m a k i n g , s e c r e c y , a n d c o n c e r n s
o v e r s e c u r i t y c a n u n d e r m i n e i t .
O u r l e s s o n s p o i n t t o t h e d i f f i c u l t i e s i n h e r e n t i n a t t e m p t i n g t o a c h i e v e
i n t e r n a t i o n a l c o n t r o l a n d t o t h e d e e p i n t e r t w i n i n g o f t e c h n i c a l a n d p o l i t i c a l i s s u e s . I t i s , i n f a c t , a m a z i n g t h a t
d e b a t e s o n i n t e r n a t i o n a l c o n t r o l g o t a s f a r a s t h e y d i d i n 1 9 4 6 ( t h a t i s , i n f a c t , o u r f i r s t l e s s o n ) . T h e r e a r e ,
h o w e v e r , o p p o r t u n i t i e s f o r t h o s e p u s h i n g t o w a r d s i n t e r n a t i o n a l g o v e r n a n c e : e v e n c y n i c s c a n s u p p o r t p r o p o s a l s ,
a n d p u b l i c o p i n i o n a n d t e c h n i c a l e x p e r t i s e c a n b e p o w e r f u l s o u r c e s o f s u p p o r t .
T h e h i s t o r y o f a t o m i c i n t e r n a t i o n a l c o n t r o l i s t o o r i c h a n d b r o a d t o b e f u l l y c a p t u r e d i n t h i s r e p o r t , a n d c a n
p r o v i d e m a n y o t h e r l e s s o n s f o r f u t u r e p o w e r f u l t e c h n o l o g i e s . Q u e s t i o n s a n d t o p i c s w o r t h y o f f u r t h e r i n q u i r y
i n c l u d e :
●
T h e r o l e o f a c t i v i s t s a n d a c t i v i s m . H o w d i d a c t i v i s t s f o r m a n d m a i n t a i n t h e i r o r g a n i z a t i o n s ? H o w w e r e
t h e y f u n d e d , a n d d i d t h a t m a t t e r ? W h a t t a c t i c s a n d o r g a n i z a t i o n s w e r e e s p e c i a l l y s u c c e s s f u l ?
●
H o w a r e t h e p o l i t i c s o f i n t e r n a t i o n a l c o n t r o l i m p a c t e d b y t r a d i t i o n a l l y i m p o r t a n t f e a t u r e s o f t h e
p o l i t i c a l l a n d s c a p e , s u c h a s p a r t i s a n d i v i d e s , t h e j u d i c i a r y , s t r o n g e x e c u t i v e s , l a m e - d u c k p r e s i d e n t s ,
s t r o n g o r w e a k i n c u m b e n t s , u p c o m i n g e l e c t i o n s , e t c . ?
●
C a n w e s a y m o r e a b o u t t h e r o l e a n d d y n a m i c s o f t h e p u b l i c s p h e r e ? W h a t i d e a s o r f r a m i n g s w e r e m o s t
l i k e l y t o r e s o n a t e ? W h a t c o m m u n i c a t i o n s w e r e m o s t i m p a c t f u l ( e . g . , l e c t u r e s , r a d i o t a l k s , i n t e r v i e w s ,
p r e s i d e n t i a l s p e e c h e s ) ?
●
T h e m i l i t a r y r e s p o n d e d i n c o m p l e x a n d v a r i e d w a y s t o a t o m i c w e a p o n s . M o r e w o r k s h o u l d b e d o n e t o
u n d e r s t a n d t h e e x t e n t t o w h i c h t h e s e r e s p o n s e s w e r e s h a p e d b y o r g a n i z a t i o n a l i n t e r e s t , i d e a s , p e r s o n a l
i d i o s y n c r a s i e s , a n d o t h e r f a c t o r s .
O t h e r m o m e n t s i n t h e g l o b a l p o l i t i c s o f n u c l e a r w e a p o n s a n d o f o t h e r p o w e r f u l t e c h n o l o g i e s a l s o w a r r a n t s t u d y .
T h e s e i n c l u d e :
●
E a r l i e r n e g o t i a t i o n s r e v o l v i n g a r o u n d t h e a b o l i t i o n o f l a r g e c l a s s e s o f t e c h n o l o g i e s , e s p e c i a l l y t h e n a v a l
a r m s t r e a t i e s o f t h e 1 9 2 0 s , t h e 1 9 2 5 G e n e v a P r o t o c o l o n t h e p r o h i b i t i o n o f t h e u s e o f c h e m i c a l
w e a p o n s , a n d t h e d i s c u s s i o n s a t t h e 1 9 3 2 G e n e v a d i s a r m a m e n t c o n f e r e n c e ( e . g . , a v i a t i o n ) m a y g i v e
u s e f u l i n s i g h t s i n t o p o s s i b l e d i r e c t i o n s f o r m o d e r n t r a n s f o r m a t i v e t e c h n o l o g i e s .
●
N e g o t i a t i o n s o v e r t h e S t r a t e g i c A r m s L i m i t a t i o n T a l k s a g r e e m e n t ( s i g n e d 1 9 7 2 ) , t h e A n t i - B a l l i s t i c
M i s s i l e T r e a t y ( 1 9 7 2 ) , t h e B i o l o g i c a l W e a p o n s C o n v e n t i o n ( 1 9 7 2 ) , t h e S t r a t e g i c A r m s L i m i t a t i o n
T a l k s I I a g r e e m e n t ( 1 9 7 9 ) , t h e S t r a t e g i c A r m s R e d u c t i o n T r e a t y ( 1 9 9 1 ) , a n d t h e C h e m i c a l W e a p o n s
C o n v e n t i o n ( 1 9 9 3 ) m a y a l s o b e r e l e v a n t .
2 1 9
2 1 9
T h e r e i s a v e r y l a r g e l i t e r a t u r e o n t h e s e a g r e e m e n t s a n d e f f o r t s . F o r a n o v e r v i e w o f t h e s e , s e e R o b e r t E . W i l l i a m s , J r . , a n d
P a u l R . V i o t t i ,
A r m s C o n t r o l : H i s t o r y , T h e o r y , a n d P o l i c y
v o l u m e 1 : F o u n d a t i o n s o f A r m s C o n t r o l ( S a n t a B a r b a r a , C A :
A B C - C L I O , 2 0 1 2 ) , c h a p t e r s 1 4 , 1 5 , 1 6 , 1 7 , 1 8 .
5 1
F i n a l l y , i t i s w o r t h a r e m i n d e r t h a t a l l o f t h e s e h i s t o r i c a l e p i s o d e s p r o v i d e o n l y c i r c u m s c r i b e d l e s s o n s f o r f u t u r e
p o w e r f u l t e c h n o l o g i e s , s u c h a s A I . N o n e o f t h e m o f f e r a c l e a n a n a l o g y . R a t h e r , t h e y a r e s o u r c e s o f i n s p i r a t i o n ,
i n s i g h t i n t o m e c h a n i s m s a n d d y n a m i c s , a n d e x a m p l e s o f h o w p o l i t i c s c a n p l a y o u t .
5 2
6 . R e f e r e n c e s
U n i t e d S t a t e s O f f i c i a l D o c u m e n t s
B u s h , V a n n e v a r , a n d J a m e s B . C o n a n t , “ M e m o r a n d u m ” , 3 0 S e p t e m b e r 1 9 4 4 . A v a i l a b l e a t :
h t t p s : / / n s a r c h i v e 2 . g w u . e d u / / N S A E B B / N S A E B B 1 6 2 / 1 . p d f . A c c e s s e d 2 2 S e p t e m b e r 2 0 1 8 .
B a r n a r d , C h e s t e r I . , J . R . O p p e n h e i m e r , C h a r l e s A . T h o m a s e t a l . ,
A R e p o r t o n t h e I n t e r n a t i o n a l C o n t r o l o f
A t o m i c E n e r g y
( W a s h i n g t o n , D C : T h e S t a t e D e p a r t m e n t , 1 9 4 6 ) .
“ M e m o r a n d u m b y t h e D i r e c t o r o f t h e O f f i c e o f S c i e n t i f i c R e s e a r c h a n d D e v e l o p m e n t ( B u s h ) t o t h e S e c r e t a r y
o f S t a t e ” , 5 N o v e m b e r 1 9 4 5 , i n U . S . D e p a r t m e n t o f S t a t e ,
F o r e i g n R e l a t i o n s o f t h e U n i t e d S t a t e s : D i p l o m a t i c
P a p e r s , 1 9 4 5
. G e n e r a l : P o l i t i c a l a n d E c o n o m i c M a t t e r s V o l u m e I I ( W a s h i n g t o n , D C : G o v e r n m e n t P r i n t i n g
O f f i c e , 1 9 4 6 ) , p p . 6 9 - 7 3 . A v a i l a b l e a t :
h t t p : / / d i g i c o l l . l i b r a r y . w i s c . e d u / c g i - b i n / F R U S / F R U S - i d x ? t y p e = a r t i c l e & d i d = F R U S . F R U S 1 9 4 5 v 0 2 . i 0 0 0 5 & i d =
F R U S . F R U S 1 9 4 5 v 0 2 & i s i z e = M . A c c e s s e d 2 4 A p r i l 2 0 1 9 .
M e m o r a n d u m , “ D r a f t P r o p o s a l s o n A t o m i c E n e r g y f o r S u b m i s s i o n t o S o v i e t G o v e r n m e n t ” , 1 0 D e c e m b e r
1 9 4 5 , i n U . S . D e p a r t m e n t o f S t a t e ,
F o r e i g n R e l a t i o n s o f t h e U n i t e d S t a t e s : D i p l o m a t i c P a p e r s , 1 9 4 5
. G e n e r a l :
P o l i t i c a l a n d E c o n o m i c M a t t e r s V o l u m e I I ( W a s h i n g t o n , D C : G o v e r n m e n t P r i n t i n g O f f i c e , 1 9 4 6 ) . A v a i l a b l e a t :
h t t p s : / / h i s t o r y . s t a t e . g o v / h i s t o r i c a l d o c u m e n t s / f r u s 1 9 4 5 v 0 2 / d 3 8 . A c c e s s e d 2 2 A p r i l 2 0 1 9 .
“ M e m o r a n d u m b y t h e C o m m a n d i n g G e n e r a l , M a n h a t t a n E n g i n e e r D i s t r i c t ( G r o v e s ) ” , 2 J a n u a r y 1 9 4 6 , i n U . S .
D e p a r t m e n t o f S t a t e ,
F o r e i g n R e l a t i o n s o f t h e U n i t e d S t a t e s , 1 9 4 6
. G e n e r a l ; T h e U n i t e d N a t i o n s , V o l u m e I
( W a s h i n g t o n , D C : G o v e r n m e n t P r i n t i n g O f f i c e , 1 9 4 7 ) , d o c u m e n t 6 0 0 . A v a i l a b l e a t :
h t t p s : / / h i s t o r y . s t a t e . g o v / h i s t o r i c a l d o c u m e n t s / f r u s 1 9 4 6 v 0 1 / d 6 0 0 . A c c e s s e d 2 4 A p r i l 2 0 1 9 .
N o t e s o f t h e I n t e r i m C o m m i t t e e M e e t i n g , T h u r s d a y 3 1 M a y 1 9 4 5
( 3 1 M a y 1 9 4 5 ) . A v a i l a b l e a t :
h t t p : / / w w w . n u c l e a r f i l e s . o r g / m e n u / k e y - i s s u e s / n u c l e a r - w e a p o n s / h i s t o r y / p r e - c o l d - w a r / i n t e r i m - c o m m i t t e e / i n t e r
i m - c o m m i t t e e - i n f o r m a l - n o t e s _ 1 9 4 5 - 0 5 - 3 1 . h t m
, a c c e s s e d 1 5 O c t o b e r 2 0 1 9 .
N S C 1 0 0 , " R e c o m m e n d e d P o l i c i e s a n d A c t i o n s i n L i g h t o f t h e G r a v e W o r l d S i t u a t i o n ” , 1 1 J a n u a r y 1 9 5 1 ,
a v a i l a b l e a t : h t t p s : / / w w w . m t h o l y o k e . e d u / a c a d / i n t r e l / k o r e a / n s c 1 0 0 . h t m . A c c e s s e d 2 0 A p r i l 2 0 1 9 .
S t i m s o n , H e n r y ,
M e m o r a n d u m D i s c u s s e d w i t h t h e P r e s i d e n t
( 2 5 A p r i l 1 9 4 5 ) . A v a i l a b l e a t :
h t t p : / / w w w . n u c l e a r f i l e s . o r g / m e n u / l i b r a r y / c o r r e s p o n d e n c e / s t i m s o n - h e n r y / c o r r _ s t i m s o n _ 1 9 4 5 - 0 4 - 2 5 . h t m
,
a c c e s s e d 1 5 O c t o b e r 2 0 1 9 .
H e n r y S t i m s o n ,
M e m o r a n d u m o n t h e E f f e c t s o f t h e A t o m i c B o m b
( 1 1 S e p t e m b e r 1 9 4 5 ) . A v a i l a b l e a t :
h t t p : / / w w w . n u c l e a r f i l e s . o r g / m e n u / l i b r a r y / c o r r e s p o n d e n c e / s t i m s o n - h e n r y / c o r r _ s t i m s o n _ 1 9 4 5 - 0 9 - 1 1 . h t m
,
a c c e s s e d 1 6 O c t o b e r 2 0 1 9 .
5 3
U . S . D e p a r t m e n t o f S t a t e ,
D o c u m e n t s o n D i s a r m a m e n t
v o l u m e 1 1 9 4 5 - 1 9 5 6 ( W a s h i n g t o n , D C : D e p a r t m e n t o f
S t a t e , 1 9 6 0 ) , p p . 1 7 - 2 4 .
U . S . D e p a r t m e n t o f S t a t e ,
U n i t e d S t a t e s A t o m i c E n e r g y P r o p o s a l s
( W a s h i n g t o n , D C : U . S . D e p a r t m e n t o f S t a t e ,
1 9 4 6 ) .
O t h e r P r i m a r y S o u r c e s
“ R e p o r t o f t h e M e e t i n g o f t h e M i n i s t e r s o f F o r e i g n A f f a i r s o f t h e U n i o n o f S o v i e t S o c i a l i s t R e p u b l i c s , t h e
U n i t e d S t a t e s o f A m e r i c a , t h e U n i t e d K i n g d o m ” , 2 7 D e c e m b e r 1 9 4 5 . A t
h t t p : / / a v a l o n . l a w . y a l e . e d u / 2 0 t h _ c e n t u r y / d e c a d e 1 9 . a s p . A c c e s s e d 2 4 A p r i l 2 0 1 9 .
“ S i g n i f i c a n c e o f t h e H - B o m b , a n d A m e r i c a ’ s D i l e m m a ” ,
N e w s w e e k
( 1 2 F e b r u a r y 1 9 5 0 ) .
“ T h e C a r n e g i e D r a f t C o n v e n t i o n ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
1 & 2 , 2 ( 1 J u l y 1 9 4 6 ) , p p . 1 5 - 1 9 .
“ T h e A m e r i c a n P r o p o s a l f o r I n t e r n a t i o n a l C o n t r o l P r e s e n t e d b y B e r n a r d B a r u c h ” ,
B u l l e t i n o f t h e A t o m i c
S c i e n t i s t s
1 & 2 ( 1 J u l y 1 9 4 6 ) p p . 3 - 5 , 1 0 . A l s o a t :
h t t p : / / w w w . a t o m i c a r c h i v e . c o m / D o c s / D e t e r r e n c e / B a r u c h P l a n . s h t m l . A c c e s s e d 2 5 A p r i l 2 0 1 9 .
“ T h e R u s s i a n P r o p o s a l f o r I n t e r n a t i o n a l C o n t r o l ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
1 & 2 , 2 ( 1 J u l y 1 9 4 6 ) , p p .
8 - 1 0 .
“ U n i t e d N a t i o n s a n d A t o m i c E n e r g y … . ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
3 , 2 ( 1 9 4 7 ) , p p . 5 5 - 6 .
A t o m i c S c i e n t i s t s o f C h i c a g o ,
T h e A t o m i c B o m b : F a c t s a n d I m p l i c a t i o n s
( C h i c a g o : T h e A t o m i c S c i e n t i s t s o f
C h i c a g o , 1 9 4 6 ) .
B o h r , N i e l s , “ N i e l s B o h r ' s M e m o r a n d u m t o P r e s i d e n t R o o s e v e l t ” , J u l y 1 9 4 4 . A v a i l a b l e a t :
h t t p : / / w w w . a t o m i c a r c h i v e . c o m / D o c s / M a n h a t t a n P r o j e c t / B o h r m e m o . s h t m l . A c c e s s e d 2 2 S e p t e m b e r 2 0 1 8 .
C a n t r i l , H a d l e y ,
P u b l i c O p i n i o n 1 9 3 5 - 1 9 4 6
( P r i n c e t o n , N J : P r i n c e t o n U n i v e r s i t y P r e s s , 1 9 5 1 ) .
L e v i e r o , A n t h o n y , “ D e n a t u r i n g G u a r d o n A t o m N o t S u r e ” ,
N e w Y o r k T i m e s
( 1 0 A p r i l 1 9 4 6 ) , p . 1 6 .
L i p p m a n n , W a l t e r , “ M r . B a r u c h a n d t h e V e t o ” ,
W a s h i n g t o n P o s t
( 2 0 J u n e 1 9 4 6 ) .
O p p e n h e i m e r , J . R . , “ T h e A t o m B o m b a s a G r e a t F o r c e f o r P e a c e ” ,
T h e N e w Y o r k T i m e s
( 9 J u n e 1 9 4 6 ) .
S m y t h , H e n r y D e W o l f ,
A t o m i c E n e r g y f o r M i l i t a r y P u r p o s e s : T h e O f f i c i a l R e p o r t o n t h e D e v e l o p m e n t o f t h e
A t o m i c B o m b U n d e r t h e A u s p i c e s o f t h e U n i t e d S t a t e s G o v e r n m e n t , 1 9 4 0 - 1 9 4 5
( S t a n f o r d : S t a n f o r d U n i v e r s i t y
P r e s s , 1 9 8 9 ) .
S z i l a r d , L e o , “ C a l l i n g f o r a C r u s a d e ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
3 , 4 - 5 ( M a y 1 9 4 7 ) , p p . 1 0 2 - 6 , 1 2 5 .
T r u m a n , H a r r y S . ,
T h e M e m o i r s o f H a r r y S . T r u m a n
v o l u m e 1 ( N e w Y o r k : G a r d e n C i t y , 1 9 5 5 ) .
5 4
T r u m a n , H a r r y S . ,
M e m o i r s
V o l u m e I I Y e a r s o f T r i a l a n d H o p e ( N e w Y o r k : D o u b l e d a y a n d C o m p a n y , 1 9 5 6 ) .
W r i g h t , Q u i n c y , “ D r a f t f o r a C o n v e n t i o n o n A t o m i c E n e r g y ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
1 ( 1 A p r i l 1 9 4 6 ) ,
p p . 1 1 - 1 3 .
S e c o n d a r y S o u r c e s
A a s e r u d , F i n n , “ T h e S c i e n t i s t a n d t h e S t a t e s m e n : N i e l s B o h r ' s P o l i t i c a l C r u s a d e d u r i n g W o r l d W a r I I ” ,
H i s t o r i c a l S t u d i e s i n t h e P h y s i c a l a n d B i o l o g i c a l S c i e n c e s
3 0 , 1 ( 1 9 9 9 ) , p p . 1 - 4 7 .
B a e r , G e o r g e W . ,
O n e H u n d r e d Y e a r s o f S e a P o w e r : T h e U . S . N a v y , 1 8 9 0 - 1 9 9 0
( S t a n f o r d : S t a n f o r d U n i v e r s i t y
P r e s s , 1 9 9 6 ) .
B e i s n e r , R o b e r t L . ,
D e a n A c h e s o n : A L i f e i n t h e C o l d W a r
( O x f o r d : O x f o r d U n i v e r s i t y P r e s s , 2 0 0 6 ) .
B e r n s t e i n , B a r t o n J . , “ T h e Q u e s t f o r S e c u r i t y : A m e r i c a n F o r e i g n P o l i c y a n d I n t e r n a t i o n a l C o n t r o l o f A t o m i c
E n e r g y , 1 9 4 2 - 1 9 4 6 ” ,
T h e J o u r n a l o f A m e r i c a n H i s t o r y
6 0 , 4 ( M a r c h 1 9 7 4 ) , 1 0 0 3 - 1 0 4 4 .
B e r n s t e i n , B a r t o n J . , “ R o o s e v e l t , T r u m a n , a n d t h e A t o m i c B o m b , 1 9 4 1 - 1 9 4 5 : A R e i n t e r p r e t a t i o n ” ,
P o l i t i c a l
S c i e n c e Q u a r t e r l y
9 0 , 1 ( S p r i n g 1 9 7 5 ) , p p . 2 3 - 6 9 .
B e r n s t e i n , B a r t o n J . , “ T h e U n e a s y A l l i a n c e : R o o s e v e l t , C h u r c h i l l , a n d t h e A t o m i c B o m b , 1 9 4 0 - 1 9 4 5 ” ,
T h e
W e s t e r n P o l i t i c a l Q u a r t e r l y
2 9 , 2 ( J u n e 1 9 7 6 ) , p p . 2 0 2 - 3 0 .
B e r n s t e i n , B a r t o n J . , “ S c i e n t i s t s a n d N u c l e a r W e a p o n s i n W o r l d W a r I I ” , i n T h o m a s W . Z e i l e r a n d D a n i e l M .
D u B o i s ( e d s . ) ,
A C o m p a n i o n t o W o r l d W a r I I
v o l u m e 1 ( O x f o r d : W i l e y - B l a c k w e l l , 2 0 1 3 ) , p p . 5 1 6 - 4 8 .
B i r d , K a i , a n d M a r t i n J . S h e r w i n ,
A m e r i c a n P r o m e t h e u s : T h e T r i u m p h a n d T r a g e d y o f J . R o b e r t O p p e n h e i m e r
( N e w Y o r k : V i n t a g e B o o k s , 2 0 0 6 ) .
B l i t z , D a v i d , “ D i d R u s s e l l A d v o c a t e P r e v e n t i v e A t o m i c W a r A g a i n s t t h e U S S R ? ” ,
T h e J o u r n a l o f B e r t r a n d
R u s s e l l S t u d i e s
2 2 ( s u m m e r 2 0 0 2 ) , p p . 5 - 4 5 .
N i c k B o s t r o m , A l l a n D a f o e , a n d C a r r i c k F l y n n , “ P o l i c y D e s i d e r a t a f o r S u p e r i n t e l l i g e n t A I : A V e c t o r F i e l d
A p p r o a c h ” ( 2 0 1 8 ) . A v a i l a b l e a t :
h t t p s : / / w w w . n i c k b o s t r o m . c o m / p a p e r s / a i p o l i c y . p d f
, a c c e s s e d 2 8 A p r i l 2 0 2 0 .
B o s t r o m , N i c k , “ T h e V u l n e r a b l e W o r l d H y p o t h e s i s ” ,
G l o b a l P o l i c y
( 2 0 1 9 ) , d o i : 1 0 . 1 1 1 1 / 1 7 5 8 - 5 8 9 9 . 1 2 7 1 8 .
B o t t i , T i m o t h y J . ,
A c e i n t h e H o l e : W h y t h e U n i t e d S t a t e s D i d N o t U s e N u c l e a r W e a p o n s i n t h e C o l d W a r , 1 9 4 5
t o 1 9 6 5
( W e s t p o r t , C T : G r e e n w o o d P r e s s , 1 9 9 6 ) .
B o y e r , P a u l ,
B y t h e B o m b ’ s E a r l y L i g h t : A m e r i c a n T h o u g h t a n d C u l t u r e a t t h e D a w n o f t h e A t o m i c A g e
( N e w
Y o r k : P a n t h e o n B o o k s , 1 9 8 5 ) .
5 5
B r o s c i o u s , D a v i d , “ L o n g i n g f o r I n t e r n a t i o n a l C o n t r o l , B a n k i n g o n A m e r i c a n S u p e r i o r i t y : H a r r y S T r u m a n ’ s
A p p r o a c h t o N u c l e a r W e a p o n s ” , i n J o h n L e w i s G a d d i s e t a l ( e d s . ) ,
C o l d W a r S t a t e s m e n C o n f r o n t t h e B o m b :
N u c l e a r D i p l o m a c y S i n c e 1 9 4 5
( O x f o r d : O x f o r d U n i v e r s i t y P r e s s , 1 9 9 9 ) , p p . 1 5 - 3 8 .
B u h i t e , R u s s e l l D . , a n d W M . C h r i s t o p h e r H a m e l , “ W a r f o r P e a c e : T h e Q u e s t i o n o f a n A m e r i c a n P r e v e n t i v e
W a r a g a i n s t t h e S o v i e t U n i o n , 1 9 4 5 – 1 9 5 5 ” ,
D i p l o m a t i c H i s t o r y
1 4 , 3 ( 1 9 9 0 ) , p p . 3 6 7 – 8 4 .
C a m e r o n , J a m e s ,
T h e D o u b l e G a m e T h e D e m i s e o f A m e r i c a ' s F i r s t M i s s i l e D e f e n s e S y s t e m a n d t h e R i s e o f
S t r a t e g i c A r m s L i m i t a t i o n
( N e w Y o r k : O x f o r d U n i v e r s i t y P r e s s , 2 0 1 7 ) .
C a r n e g i e , A l l i s o n , a n d A u s t i n C a r s o n , “ T h e D i s c l o s u r e D i l e m m a : N u c l e a r I n t e l l i g e n c e a n d I n t e r n a t i o n a l
O r g a n i z a t i o n s ” ,
T h e A m e r i c a n J o u r n a l o f P o l i t i c a l S c i e n c e
6 3 , 2 ( 2 0 1 9 ) , p p . 2 6 9 - 2 8 5 .
C a r s o n , A . , “ F a c i n g O f f a n d S a v i n g F a c e : C o v e r t I n t e r v e n t i o n a n d E s c a l a t i o n M a n a g e m e n t i n t h e K o r e a n W a r ” ,
I
n t e r n a t i o n a l O r g a n i z a t i o n
7 0 , 1 ( 2 0 1 6 ) , p p . 1 0 3 - 3 1 .
C a s e y , S t e v e n , “ S e l l i n g N S C - 6 8 : T h e T r u m a n A d m i n i s t r a t i o n , P u b l i c O p i n i o n , a n d t h e P o l i t i c s o f
M o b i l i z a t i o n , 1 9 5 0 – 5 1 ” ,
D i p l o m a t i c H i s t o r y
2 9 , 4 ( S e p t e m b e r 2 0 0 5 ) , p p . 6 5 5 - 9 0 .
C i h o n , P e t e r , M a t t h i j s M . M a a s , a n d L u k e K e m p , “ S h o u l d A r t i f i c i a l I n t e l l i g e n c e G o v e r n a n c e b e C e n t r a l i s e d ?
D e s i g n L e s s o n s f r o m H i s t o r y ” ,
P r o c e e d i n g s o f t h e 2 0 2 0 A A A I / A C M C o n f e r e n c e o n A I , E t h i c s , a n d S o c i e t y ( A I E S
’ 2 0 ) , F e b r u a r y 7 – 8 , 2 0 2 0 , N e w Y o r k , N Y , U S A
( 2 0 2 0 ) . h t t p s : / / d o i . o r g / 1 0 . 1 1 4 5 / 3 3 7 5 6 2 7 . 3 3 7 5 8 5 7 .
C o e , A n d r e w J . , a n d J a n e V a y n m a n . “ W h y A r m s C o n t r o l I s S o R a r e ” ,
A m e r i c a n P o l i t i c a l S c i e n c e R e v i e w
1 1 4 , 2
( 2 0 2 0 ) , p p . 3 4 2 – 5 5 .
C u r a t o l a , J o h n M . ,
B i g g e r B o m b s f o r a B r i g h t e r T o m o r r o w : S t r a t e g i c A i r C o m m a n d a n d A m e r i c a n W a r P l a n s
a t t h e D a w n o f t h e A t o m i c A g e , 1 9 4 5 - 1 9 5 0
( J e f f e r s o n , N C : M a c F a r l a n d , 2 0 1 6 ) .
D a f o e , A l l a n , A
I G o v e r n a n c e : A R e s e a r c h A g e n d a
( O x f o r d : F u t u r e o f H u m a n i t y I n s t i t u t e , 2 0 1 8 ) . A v a i l a b l e a t :
h t t p s : / / w w w . f h i . o x . a c . u k / w p - c o n t e n t / u p l o a d s / G o v A I A g e n d a . p d f , a c c e s s e d 2 7 M a y 2 0 1 9 .
D a v i s , L y n n E t h e r i d g e ,
T h e C o l d W a r B e g i n s : S o v i e t - A m e r i c a n C o n f l i c t O v e r E a s t e r n E u r o p e
( P r i n c e t o n , N J :
P r i n c e t o n U n i v e r s i t y P r e s s , 1 9 7 4 ) , c h a p t e r s 7 , 8 , a n d 9 .
D a v i s , Z a c h a r y S . , “ T h e R e a l i s t N u c l e a r R e g i m e ” ,
S e c u r i t y S t u d i e s
2 , 3 - 4 ( 1 9 9 3 ) , p p . 7 9 - 9 9 .
D e V o l p i , A . , “ D e n a t u r i n g F i s s i l e M a t e r i a l s ” ,
P r o g r e s s i n N u c l e a r E n e r g y
1 0 , 2 ( 1 9 8 2 ) , p p . 1 6 1 - 2 2 0 .
D i v i n e , R o b e r t A . ,
S e c o n d C h a n c e : T h e T r i u m p h o f I n t e r n a t i o n a l i s m i n A m e r i c a D u r i n g W o r l d W a r I I
( N e w
Y o r k : A t h e n e u m , 1 9 7 1 ) .
D o n n e l l y , J a c k ,
R e a l i s m a n d I n t e r n a t i o n a l R e l a t i o n s
( C a m b r i d g e : C a m b r i d g e U n i v e r s i t y P r e s s , 2 0 0 4 ) .
D o n o v a n , R o b e r t J . ,
C o n f l i c t a n d C r i s i s : T h e P r e s i d e n c y o f H a r r y S . T r u m a n , 1 9 4 5 - 1 9 4 8
( C o l u m b i a , M O :
U n i v e r s i t y o f M i s s o u r i P r e s s , 1 9 9 6 ) .
5 6
E v a n g e l i s t a , M a t t h e w A . , “ S t a l i n ' s P o s t w a r A r m y R e a p p r a i s e d ” ,
I n t e r n a t i o n a l S e c u r i t y
7 , 3 ( W i n t e r , 1 9 8 2 - 1 9 8 3 ) ,
p p . 1 1 0 - 3 8 .
F a r m e l o , G r a h a m ,
C h u r c h i l l ’ s B o m b : H o w t h e U n i t e d S t a t e s O v e r t o o k B r i t a i n i n t h e F i r s t N u c l e a r A r m s R a c e
( N e w Y o r k : B a s i c B o o k s , 2 0 1 3 ) .
F r e e d m a n , L a w r e n c e ,
T h e E v o l u t i o n o f N u c l e a r S t r a t e g y
3 r d e d . ( L o n d o n : P a l g r a v e M a c m i l l a n , 2 0 0 3 ) .
G a d d i s , J o h n L e w i s ,
S t r a t e g i e s o f C o n t a i n m e n t : A C r i t i c a l A p p r a i s a l o f P o s t w a r A m e r i c a n N a t i o n a l S e c u r i t y
P o l i c y
( N e w Y o r k : O x f o r d U n i v e r s i t y P r e s s , 1 9 8 2 ) .
G a d d i s , J o h n L e w i s ,
T h e C o l d W a r
( L o n d o n : A l l e n L a n e , 2 0 0 5 ) .
G a d d i s , J o h n L e w i s ,
S t r a t e g i e s o f C o n t a i n m e n t : A C r i t i c a l A p p r a i s a l o f A m e r i c a n N a t i o n a l S e c u r i t y P o l i c y
d u r i n g t h e C o l d W a r
2 n d e d . ( O x f o r d : O x f o r d U n i v e r s i t y P r e s s , 2 0 0 5 ) .
G o r d i n , M i c h a e l ,
R e d C l o u d a t D a w n : T r u m a n , S t a l i n , a n d t h e E n d o f t h e A t o m i c M o n o p o l y
( N e w Y o r k :
F a r r a r , S t r a u s a n d G i r o u x , 2 0 0 9 ) .
G o r m l y , J a m e s L . , “ T h e W a s h i n g t o n D e c l a r a t i o n a n d t h e ‘ P o o r R e l a t i o n ’ : A n g l o - A m e r i c a n A t o m i c D i p l o m a c y ,
1 9 4 5 - 4 6 ” ,
D i p l o m a t i c H i s t o r y
8 , 2 ( A p r i l 1 9 8 4 ) , p p . 1 2 5 - 1 4 3 .
G r a n t , J a m e s ,
B e r n a r d M . B a r u c h : T h e A d v e n t u r e s o f a W a l l S t r e e t L e g e n d
( N e w Y o r k : J o h n W i l e y & S o n s ,
1 9 9 7 ) .
G r a y b a r , L l o y d J . , “ T h e 1 9 4 6 A t o m i c B o m b T e s t s : A t o m i c D i p l o m a c y o r B u r e a u c r a t i c I n f i g h t i n g ” ,
T h e J o u r n a l
o f A m e r i c a n H i s t o r y
7 2 , 4 ( M a r c h 1 9 8 6 ) , p p . 8 8 8 - 9 0 7 .
H a w k i n s , H e l e n S . e t a l ( e d s . ) ,
T o w a r d a L i v a b l e W o r l d : L e o S z i l a r d a n d t h e C r u s a d e f o r N u c l e a r A r m s C o n t r o l
( C a m b r i d g e , M A : T h e M I T P r e s s , 1 9 8 7 ) .
H e l m r e i c h , J o n a t h a n E . ,
G a t h e r i n g R a r e O r e s : T h e D i p l o m a c y o f U r a n i u m A c q u i s i t i o n , 1 9 4 3 - 1 9 5 4
( P r i n c e t o n ,
N J : P r i n c e t o n U n i v e r s i t y P r e s s , 1 9 8 6 ) .
H e r k e n , G r e g g ,
T h e W i n n i n g W e a p o n : T h e A t o m i c B o m b i n t h e C o l d W a r 1 9 4 5 - 1 9 5 0
2 n d e d i t i o n ( P r i n c e t o n ,
N J : P r i n c e t o n U n i v e r s i t y P r e s s , 1 9 8 8 ) .
H e r k e n , G r e g g , “ ‘ A M o s t D e a d l y I l l u s i o n ’ : T h e A t o m i c S e c r e t a n d A m e r i c a n N u c l e a r W e a p o n s P o l i c y ,
1 9 4 5 - 1 9 5 0 ” ,
P a c i f i c H i s t o r i c a l R e v i e w
4 9 , 1 ( F e b r u a r y 1 9 8 0 ) , p p . 5 1 - 7 6 .
H e r s h b e r g , J a m e s ,
J a m e s B . C o n a n t : H a r v a r d t o H i r o s h i m a a n d t h e M a k i n g o f t h e N u c l e a r A g e
( N e w Y o r k :
A l f r e d A . K n o p f , 1 9 9 3 ) .
H e w l e t t , R i c h a r d G . , a n d O s c a r E . A n d e r s o n , J r . ,
A H i s t o r y o f t h e U n i t e d S t a t e s A t o m i c E n e r g y C o m m i s s i o n
v o l u m e 1 T h e N e w W o r l d 1 9 3 9 / 1 9 4 6 ( U n i v e r s i t y P a r k , P A : T h e U n i v e r s i t y o f P e n n s y l v a n i a P r e s s , 1 9 6 2 ) .
5 7
H o g a n , M i c h a e l J . ,
A C r o s s o f I r o n : H a r r y S . T r u m a n a n d t h e O r i g i n s o f t h e N a t i o n a l S e c u r i t y S t a t e
( C a m b r i d g e : C a m b r i d g e U n i v e r s i t y P r e s s , 2 0 0 0 ) .
H o g a n , J . , a n d S . F e e n e y , “ T h e R o l e o f t h e P o l i t i c a l E n t r e p r e n e u r i n t h e C o n t e x t o f P o l i c y C h a n g e a n d C r i s i s ” ,
M i d w e s t P o l i t i c a l S c i e n c e A s s o c i a t i o n A n n u a l C o n f e r e n c e , C h i c a g o , A p r i l 1 4 t h 2 0 1 3 . A v a i l a b l e a t
h t t p s : / / a r r o w . d i t . i e / c g i / v i e w c o n t e n t . c g i ? a r t i c l e = 1 0 1 5 & c o n t e x t = b u s c h g r a c o n . A c c e s s e d 2 2 A p r i l 2 0 1 9 .
H o l l o w a y , D a v i d , “ T h e S o v i e t U n i o n a n d t h e c r e a t i o n o f t h e I n t e r n a t i o n a l A t o m i c E n e r g y A g e n c y ” ,
C o l d W a r
H i s t o r y
1 6 , 2 ( 2 0 1 6 ) , p p . 1 7 7 - 9 3 .
D a v i d H o l l o w a y , “ T h e S o v i e t U n i o n a n d t h e B a r u c h P l a n ” .
H t t p s : / / w w w . w i l s o n c e n t e r . o r g / b l o g - p o s t / s o v i e t - u n i o n - a n d - b a r u c h - p l a n
, a c c e s s e d 1 3 J u n e 2 0 2 0 .
K a i s e r , D a v i d , “ T h e A t o m i c S e c r e t i n R e d H a n d s ? A m e r i c a n S u s p i c i o n s o f T h e o r e t i c a l P h y s i c i s t s D u r i n g t h e
E a r l y C o l d W a r ” ,
R e p r e s e n t a t i o n s
9 0 , 1 ( S p r i n g 2 0 0 5 ) , p p 2 8 - 6 0 .
K a p l a n , E d w a r d ,
T o K i l l N a t i o n s : A m e r i c a n S t r a t e g y i n t h e A i r - A t o m i c A g e a n d t h e R i s e o f M u t u a l l y A s s u r e d
D e s t r u c t i o n
( I t h a c a , N Y : C o r n e l l U n i v e r s i t y P r e s s , 2 0 1 5 ) .
K a p l a n , F r e d ,
T h e W i z a r d s o f A r m a g e d d o n
2 n d e d i t i o n ( S t a n f o r d : S t a n f o r d U n i v e r s i t y P r e s s , 1 9 9 1 ) ,
K i m b a l l - S m i t h , A l i c e ,
A P e r i l a n d a H o p e : T h e S c i e n t i s t s ’ M o v e m e n t i n A m e r i c a : 1 9 4 5 - 4 7
( C h i c a g o : T h e
U n i v e r s i t y o f C h i c a g o P r e s s , 1 9 6 5 ) .
L a n o u e t t e , W i l l i a m , a n d B e l a S i l a r d ,
G e n i u s i n t h e S h a d o w s : A B i o g r a p h y o f L e o S z i l a r d , t h e M a n B e h i n d t h e
B o m b
( N e w Y o r k : C . S c r i b n e r ’ s S o n s , 1 9 9 2 ) .
L e e , R . A l t o n , “ T h e A r m y ‘ M u t i n y ’ o f 1 9 4 6 ” ,
T h e A m e r i c a n J o u r n a l o f A m e r i c a n H i s t o r y
5 3 , 3 ( D e c e m b e r
1 9 6 6 ) , p p . 5 5 5 - 7 1 .
L e f f l e r , M e l v y n P . , “ S t r a t e g y , D i p l o m a c y , a n d t h e C o l d W a r : T h e U n i t e d S t a t e s , T u r k e y , a n d N A T O ,
1 9 4 5 - 1 9 5 2 ” ,
T h e J o u r n a l o f A m e r i c a n H i s t o r y
7 1 , 4 ( M a r c h 1 9 8 5 ) , p p . 8 0 7 - 2 5 .
L e f f l e r , M e l v y n P . , “ T h e A m e r i c a n C o n c e p t i o n o f N a t i o n a l S e c u r i t y a n d t h e B e g i n n i n g s o f t h e C o l d W a r ,
1 9 4 5 - 4 8 ” ,
T h e A m e r i c a n H i s t o r i c a l R e v i e w
8 9 , 2 ( A p r i l 1 9 8 4 ) , p p . 3 4 6 - 8 1 .
L i e b e r m a n , J o s e p h I . ,
T h e S c o r p i o n a n d t h e T a r a n t u l a : T h e S t r u g g l e t o C o n t r o l A t o m i c W e a p o n s 1 9 4 5 - 1 9 4 9
( B o s t o n : H o u g h t o n M i f f l i n C o m p a n y , 1 9 7 0 ) .
L i n d b l o m , C h a r l e s E . , “ T h e S c i e n c e o f ‘ M u d d l i n g T h r o u g h ’ ” ,
P u b l i c A d m i n i s t r a t i o n R e v i e w
1 9 , 2 ( S p r i n g
1 9 5 9 ) , p p . 7 9 - 8 8 .
L i p s e y , R i c h a r d G . , K e n n e t h I . C a r l a w , a n d C l i f f o r d T . B e k a r ,
E c o n o m i c T r a n s f o r m a t i o n s : G e n e r a l P u r p o s e
T e c h n o l o g i e s a n d L o n g - T e r m E c o n o m i c G r o w t h
( O x f o r d : O x f o r d U n i v e r s i t y P r e s s , 2 0 0 5 ) .
M a a s , M a t t h i j s M . , “ H o w V i a b l e i s I n t e r n a t i o n a l A r m s C o n t r o l f o r M i l i t a r y A r t i f i c i a l I n t e l l i g e n c e ? T h r e e
L e s s o n s f r o m N u c l e a r W e a p o n s ” ,
C o n t e m p o r a r y S e c u r i t y P o l i c y
4 0 , 3 ( 2 0 1 9 ) , p p . 2 8 5 - 3 1 1 .
5 8
M a d d o c k , S h a n e J . ,
N u c l e a r A p a r t h e i d : T h e Q u e s t f o r A m e r i c a n A t o m i c S u p r e m a c y f r o m W o r l d W a r I I t o t h e
P r e s e n t
( C h a p e l H i l l , N C : T h e U n i v e r s i t y o f N o r t h C a r o l i n a P r e s s , 2 0 1 0 ) .
M a l l o y , S e a n L . ,
A t o m i c T r a g e d y : H e n r y L . S t i m s o n a n d t h e D e c i s i o n t o U s e t h e B o m b A g a i n s t J a p a n
( I t h a c a , N Y :
C o r n e l l U n i v e r s i t y P r e s s , 2 0 0 8 ) .
M a r k , E d u a r d , " T h e T u r k i s h W a r S c a r e o f 1 9 4 6 , " i n M e l v y n P . L e f f l e r a n d D a v i d S . P a i n t e r ( e d s . ) ,
O r i g i n s o f t h e
C o l d W a r : A n I n t e r n a t i o n a l H i s t o r y
2 n d e d . ( N e w Y o r k : R o u t l e d g e , 2 0 0 5 ) , p p . 1 1 2 - 3 3 .
M a y e r s , D a v i d ,
G e o r g e K e n n a n a n d t h e D i l e m m a s o f U S F o r e i g n P o l i c y
( O x f o r d : O x f o r d U n i v e r s i t y P r e s s ,
1 9 8 8 ) .
M a z o w e r , M a r k ,
G o v e r n i n g t h e W o r l d : T h e H i s t o r y o f a n I d e a
( N e w Y o r k : P e n g u i n , 2 0 1 2 ) .
M e i l i n g e r , P h i l l i p S . , “ T h e A d m i r a l s ’ R e v o l t o f 1 9 4 9 : L e s s o n s f o r T o d a y ” ,
P a r a m e t e r s
1 9 , 3 ( S e p t e m b e r 1 9 8 9 ) ,
p p . 8 1 - 9 6 .
M i l l e r , S t e v e n ,
S t r a t e g y a n d N u c l e a r D e t e r r e n c e : A n I n t e r n a t i o n a l S e c u r i t y R e a d e r
( P r i n c e t o n , N J : P r i n c e t o n
U n i v e r s i t y P r e s s , 1 9 8 4 ) .
M i s c a m b l e , W i l s o n D . ,
T h e M o s t C o n t r o v e r s i a l D e c i s i o n : T r u m a n , t h e A t o m i c b o m b s , a n d t h e D e f e a t o f J a p a n
( C a m b r i d g e : C a m b r i d g e U n i v e r s i t y P r e s s , 2 0 1 1 ) .
M o n k , R a y ,
R o b e r t O p p e n h e i m e r : A L i f e I n s i d e t h e C e n t e r
( N e w Y o r k : D o u b l e d a y , 2 0 1 2 ) .
M o r g a n , P a t r i c k M . , “ E l e m e n t s o f a G e n e r a l T h e o r y o f A r m s C o n t r o l ” , R o b e r t E . W i l l i a m s a n d P a u l R . V i o t t i
( e d s . ) ,
A r m s C o n t r o l : H i s t o r y , T h e o r y , a n d P o l i c y
( S a n t a B a r b a r a , C A : P r a e g e r , 2 0 1 2 ) , p p . 1 5 - 4 0 .
N o r r i s , R o b e r t S . ,
R a c i n g f o r t h e B o m b : T h e T r u e S t o r y o f G e n e r a l L e s l i e R . G r o v e s , t h e M a n b e h i n d t h e B i r t h o f
t h e A t o m i c A g e
( H a n o v e r , N H : S t e e r f o r t h P r e s s , 2 0 0 2 ) .
N o r r i s , R o b e r t S . , a n d H a n s M . K r i s t e n s e n , “ G l o b a l N u c l e a r W e a p o n s I n v e n t o r i e s , 1 9 4 5 – 2 0 1 0 ” ,
B u l l e t i n o f t h e
A t o m i c S c i e n t i s t s
6 6 , 4 ( 2 0 1 0 ) , p p . 7 7 - 8 3 .
O f f n e r , A r n o l d A . ,
A n o t h e r S u c h V i c t o r y : P r e s i d e n t T r u m a n a n d t h e C o l d W a r , 1 9 4 5 - 1 9 5 3
( S t a n f o r d : S t a n f o r d
U n i v e r s i t y P r e s s , 2 0 0 2 ) .
P a e k , H y e - J i n , a n d T h o m a s H o v e , “ R i s k P e r c e p t i o n s a n d R i s k C h a r a c t e r i s t i c s ” ,
O x f o r d R e s e a r c h E n c y c l o p e d i a s :
C o m m u n i c a t i o n
. A v a i l a b l e a t :
h t t p s : / / o x f o r d r e . c o m / c o m m u n i c a t i o n / v i e w / 1 0 . 1 0 9 3 / a c r e f o r e / 9 7 8 0 1 9 0 2 2 8 6 1 3 . 0 0 1 . 0 0 0 1 / a c r e f o r e - 9 7 8 0 1 9 0 2 2
8 6 1 3 - e - 2 8 3 , a c c e s s e d 1 J u n e 2 0 1 9 .
P a y n e , K e n n e t h , “ A r t i f i c i a l I n t e l l i g e n c e : A R e v o l u t i o n i n S t r a t e g i c A f f a i r s ? ” ,
S u r v i v a l : G l o b a l P o l i t i c s a n d
S t r a t e g y
6 0 , 5 ( 2 0 1 8 ) , p p . 7 - 3 2 .
5 9
R a i n e , F e r n a n d e S c h e i d , “ T h e I r a n i a n C r i s i s o f 1 9 4 6 a n d t h e O r i g i n s o f t h e C o l d W a r ” i n M e l v y n P . L e f f l e r a n d
D a v i d S . P a i n t e r ( e d s . ) ,
O r i g i n s o f t h e C o l d W a r : A n I n t e r n a t i o n a l H i s t o r y
2 n d e d . ( N e w Y o r k : R o u t l e d g e ,
2 0 0 5 ) , p p . 9 3 - 1 1 1 .
P l o k h y , S . M . ,
Y a l t a : T h e P r i c e o f P e a c e
( N e w Y o r k : P e n g u i n , 2 0 1 1 ) .
R o s e r , M a x , a n d M o h a m e d N a g d y , “ N u c l e a r W e a p o n s ” ,
O u r W o r l d I n D a t a . o r g
( 2 0 1 3 ) . R e t r i e v e d f r o m :
h t t p s : / / o u r w o r l d i n d a t a . o r g / n u c l e a r - w e a p o n s . A c c e s s e d 2 6 A u g u s t 2 0 2 0 .
P o w a s k i , R o n a l d E . ,
M a r c h t o A r m a g e d d o n : T h e U n i t e d S t a t e s a n d t h e N u c l e a r A r m s R a c e , 1 9 3 9 t o t h e P r e s e n t
( O x f o r d : O x f o r d U n i v e r s i t y P r e s s , 1 9 8 7 ) .
Q u e s t e r , G e o r g e H . ,
N u c l e a r M o n o p o l y
( N e w B r u n s w i c k : T r a n s a c t i o n P u b l i s h e r s , 2 0 0 0 ) .
R h o d e s , R i c h a r d ,
D a r k S u n : T h e M a k i n g o f t h e H y d r o g e n B o m b
( N e w Y o r k : S i m o n a n d S c h u s t e r , 1 9 9 5 ) .
R o s e n b e r g , D a v i d A l a n , “ A m e r i c a n A t o m i c S t r a t e g y a n d t h e H y d r o g e n B o m b D e c i s i o n ” ,
T h e J o u r n a l o f
A m e r i c a n H i s t o r y
6 6 , 1 ( J u n e 1 9 7 9 ) , p p . 6 2 - 8 7 .
R o s e n b e r g , D a v i d A l a n , “ U . S . N u c l e a r S t o c k p i l e , 1 9 4 5 t o 1 9 5 0 ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
3 8 , 5 ( M a y
1 9 8 2 ) , p p . 2 5 - 3 0 .
R o s e n b e r g , D a v i d A l a n , “ T h e O r i g i n s o f O v e r k i l l : N u c l e a r W e a p o n s a n d A m e r i c a n S t r a t e g y , 1 9 4 5 - 1 9 6 0 ” ,
I n t e r n a t i o n a l S e c u r i t y
7 , 4 ( S p r i n g 1 9 8 3 ) , p p . 3 - 7 1 .
R o s e r , M a x , a n d M o h a m e d N a g d y , “ N u c l e a r W e a p o n s ” ,
O u r W o r l d I n D a t a . o r g
( 2 0 1 3 ) .
R o s s , S t e v e n T . ,
A m e r i c a n W a r P l a n s 1 9 4 5 - 1 9 5 0
( N e w Y o r k : R o u t l e d g e , 2 0 1 3 ) .
S a c k s , S a m m , “ T h e R i p p l e E f f e c t s o f t h e C h i n a C h i p H a c k i n g S t o r y ” ,
W a r o n t h e R o c k s
( 1 7 O c t o b e r 2 0 1 8 ) .
h t t p s : / / w a r o n t h e r o c k s . c o m / 2 0 1 8 / 1 0 / t h e - r i p p l e - e f f e c t s - o f - t h e - c h i n a - c h i p - h a c k i n g - s t o r y . A c c e s s e d 2 4 A p r i l
2 0 1 9 .
S c h r a f s t e t t e r , S u s a n n a , “ ‘ L o q u a c i o u s … a n d p o i n t l e s s a s e v e r ’ ? B r i t a i n , t h e U n i t e d S t a t e s a n d t h e U n i t e d N a t i o n s
N e g o t i a t i o n s o n I n t e r n a t i o n a l C o n t r o l o f N u c l e a r E n e r g y , 1 9 4 5 - 4 8 ” ,
C o n t e m p o r a r y B r i t i s h H i s t o r y
1 6 , 4
( W i n t e r 2 0 0 2 ) , 8 7 - 1 0 8 .
S c h r e c k e r , E l l e n ,
M a n y A r e t h e C r i m e s : M c C a r t h y i s m i n A m e r i c a
( P r i n c e t o n , N J : P r i n c e t o n U n i v e r s i t y P r e s s ,
1 9 9 9 ) .
S c h w a r t z , S t e p h e n I . ,
A t o m i c A u d i t : T h e C o s t s a n d C o n s e q u e n c e s o f U . S . N u c l e a r W e a p o n s S i n c e 1 9 4 0
( W a s h i n g t o n , D C : B r o o k i n g s I n s t i t u t i o n P r e s s , 1 9 9 8 ) .
S c h w e b e r , S i l v a n S . ,
E i n s t e i n & O p p e n h e i m e r : T h e M e a n i n g o f G e n i u s
( C a m b r i d g e , M A : H a r v a r d U n i v e r s i t y
P r e s s , 2 0 0 8 ) .
6 0
S h e r w i n , M a r t i n J . , “ N i e l s B o h r a n d t h e F i r s t P r i n c i p l e s o f A r m s C o n t r o l ” , i n H e r m a n F e s h b a c h , T e t s u o
M a t s u i , a n d A l e x a n d r a O l e s o n ( e d s . ) ,
N i e l s B o h r : P h y s i c s a n d t h e W o r l d
( A b i n g d o n : R o u t l e d g e , 1 9 9 8 ) , p p .
3 1 9 - 3 0 .
S h e r w i n , M a r t i n J . ,
A W o r l d D e s t r o y e d : H i r o s h i m a a n d I t s L e g a c i e s
3 r d e d i t i o n ( S t a n f o r d : S t a n f o r d U n i v e r s i t y
P r e s s , 2 0 0 3 ) .
S i b l e y , K a t h e r i n e A . S . ,
R e d S p i e s i n A m e r i c a : S t o l e n S e c r e t s a n d t h e D a w n o f t h e C o l d W a r
( L a w r e n c e , K S :
U n i v e r s i t y o f K a n s a s P r e s s , 2 0 0 4 ) .
S t e u r y , D o n a l d P . , “ H o w t h e C I A M i s s e d S t a l i n ' s B o m b : D i s s e c t i n g S o v i e t A n a l y s i s , 1 9 4 6 - 5 0 ” . A v a i l a b l e a t :
h t t p s : / / w w w . c i a . g o v / l i b r a r y / c e n t e r - f o r - t h e - s t u d y - o f - i n t e l l i g e n c e / c s i - p u b l i c a t i o n s / c s i - s t u d i e s / s t u d i e s / v o l 4 9 n o 1
/ h t m l _ f i l e s / s t a l i n s _ b o m b _ 3 . h t m l , a c c e s s e d 1 J u n e 2 0 1 9 .
T i l l , C . E . , “ D e n a t u r e d F u e l C y c l e s ” , i n J o s e p h L . F o w l e r , C l e l a n d H . J o h n s o n , C h a r l e s D . B o w m a n ( e d s . ) ,
N u c l e a r C r o s s S e c t i o n s f o r T e c h n o l o g y : P r o c e e d i n g s o f t h e I n t e r n a t i o n a l C o n f e r e n c e
v o l 1 3 ( O a k R i d g e , T N : O a k
R i d g e N a t i o n a l L a b o r a t o r y , 1 9 8 0 ) , p p . 1 1 5 - 1 8 .
T r a c h t e n b e r g , M a r c ,
H i s t o r y a n d S t r a t e g y
( P r i n c e t o n , N J : P r i n c e t o n U n i v e r s i t y P r e s s , 1 9 9 1 ) .
W a n g , J e s s i c a ,
A m e r i c a n S c i e n c e i n a n A g e o f A n x i e t y : S c i e n t i s t s , A n t i c o m m u n i s m , a n d t h e C o l d W a r
( C h a p e l
H i l l , N C : U n i v e r s i t y o f N o r t h C a r o l i n a P r e s s , 1 9 9 9 ) .
W h i t a k e r , R e g i n a l d ,
C o l d W a r C a n a d a : T h e M a k i n g o f a N a t i o n a l I n s e c u r i t y S t a t e , 1 9 4 5 - 1 9 5 7
( T o r o n t o :
U n i v e r s i t y o f T o r o n t o P r e s s , 1 9 9 4 ) .
W i l l i a m s , R o b e r t C h a d w e l l ,
K l a u s F u c h s : A t o m S p y
( C a m b r i d g e , M A : H a r v a r d U n i v e r s i t y P r e s s , 1 9 8 7 ) .
W i l l i a m s , J r . , R o b e r t E . , a n d P a u l R . V i o t t i ,
A r m s C o n t r o l : H i s t o r y , T h e o r y , a n d P o l i c y
v o l u m e 1 : F o u n d a t i o n s
o f A r m s C o n t r o l ( S a n t a B a r b a r a , C A : A B C - C L I O , 2 0 1 2 ) .
W i t t n e r , L a w r e n c e S . ,
T h e S t r u g g l e A g a i n s t t h e B o m b
v o l u m e 1 : O n e W o r l d o r N o n e : A H i s t o r y o f t h e W o r l d
N u c l e a r D i s a r m a m e n t M o v e m e n t T h r o u g h 1 9 5 3 ( S t a n f o r d : S t a n f o r d U n i v e r s i t y P r e s s , 1 9 9 3 ) .
Y o u n g , K e n , “ U S ' A t o m i c C a p a b i l i t y ' a n d t h e B r i t i s h F o r w a r d B a s e s i n t h e E a r l y C o l d W a r ” ,
J o u r n a l o f
C o n t e m p o r a r y H i s t o r y
4 2 , 1 ( J a n u a r y 2 0 0 7 ) , p p . 1 1 7 - 3 6 .
Z a c h a r y , G . P a s c a l ,
E n d l e s s F r o n t i e r : V a n n e v a r B u s h , E n g i n e e r o f t h e A m e r i c a n C e n t u r y
( N e w Y o r k : S i m o n a n d
S c h u s t e r , 2 0 1 8 ) .
Z a i d i , W a q a r , ” ’ A v i a t i o n W i l l E i t h e r D e s t r o y o r S a v e O u r C i v i l i z a t i o n ’ : P r o p o s a l s f o r t h e I n t e r n a t i o n a l C o n t r o l
o f A v i a t i o n , 1 9 2 0 - 1 9 4 5 ” ,
J o u r n a l o f C o n t e m p o r a r y H i s t o r y
4 6 , 1 ( 2 0 1 1 ) , p p . 1 5 0 - 7 8 .
Z a l o g a , S t e v e ,
T a r g e t A m e r i c a : t h e S o v i e t U n i o n a n d t h e S t r a t e g i c A r m s R a c e , 1 9 4 5 - 1 9 6 4
( N o v a t o , C A : P r e s i d o ,
1 9 9 3 ) .
6 1
Z i e g l e r , C h a r l e s A . , a n d D a v i d J a c o b s o n ,
S p y i n g W i t h o u t S p i e s : O r i g i n s o f A m e r i c a ' s S e c r e t N u c l e a r S u r v e i l l a n c e
S y s t e m
( W e s t p o r t , C T : P r a e g e r , 1 9 9 5 ) .
Z i e g l e r , C h a r l e s A . , “ I n t e l l i g e n c e A s s e s s m e n t s o f S o v i e t A t o m i c C a p a b i l i t y , 1 9 4 5 - 1 9 4 9 : M y t h s , M o n o p o l i e s a n d
M a s k i r o v k a ” ,
I n t e l l i g e n c e a n d N a t i o n a l S e c u r i t y
1 2 , 4 ( 1 9 9 7 ) , p p . 1 - 2 4 .
Z u b o k , V l a d i s l a v , a n d C o n s t a n t i n e P l e s h a k o v ,
I n s i d e t h e K r e m l i n ’ s C o l d W a r : F r o m S t a l i n t o K h r u s h c h e v
( C a m b r i d g e , M A : H a r v a r d U n i v e r s i t y P r e s s , 1 9 9 6 ) .
Z u b o k , V l a d i s l a v M . , “ S t a l i n a n d t h e N u c l e a r A g e ” , i n J o h n L e w i s G a d d i s e t a l ( e d s . ) ,
C o l d W a r S t a t e s m e n
C o n f r o n t t h e B o m b : N u c l e a r D i p l o m a c y S i n c e 1 9 4 5
( O x f o r d : O x f o r d U n i v e r s i t y P r e s s , 1 9 9 9 ) , p p . 3 9 - 6 1 .
Z u b o k , V l a d i s l a v M . ,
A F a i l e d E m p i r e : T h e S o v i e t U n i o n i n t h e C o l d W a r f r o m S t a l i n t o G o r b a c h e v
( C h a p e l
H i l l , N C : U n i v e r s i t y o f N o r t h C a r o l i n a P r e s s , 2 0 0 7 ) ,
Z w e t s l o o t , R e m c o , a n d A l l a n D a f o e , “ T h i n k i n g A b o u t R i s k s f r o m A I : A c c i d e n t s , M i s u s e a n d S t r u c t u r e ” ,
L a w f a r e ( 1 1 F e b r u a r y 2 0 1 9 ) . A v a i l a b l e a t :
h t t p s : / / w w w . l a w f a r e b l o g . c o m / t h i n k i n g - a b o u t - r i s k s - a i - a c c i d e n t s - m i s u s e - a n d - s t r u c t u r e , a c c e s s e d 2 7 M a y 2 0 1 9 .
6 2
A p p e n d i x A : T i m e l i n e
2 2 0
S . M . P l o k h y ,
Y a l t a : T h e P r i c e o f P e a c e
( N e w Y o r k : P e n g u i n , 2 0 1 1 ) , p . 3 4 1 .
2 2 1
R o b e r t J . D o n o v a n ,
C o n f l i c t a n d C r i s i s : T h e P r e s i d e n c y o f H a r r y S . T r u m a n , 1 9 4 5 - 1 9 4 8
( C o l u m b i a , M O : U n i v e r s i t y o f
M i s s o u r i P r e s s , 1 9 9 6 ) , p . 1 1 .
2 2 2
Z a c h a r y ,
E n d l e s s F r o n t i e r
, p . 2 4 3 .
6 3
1 9 4 1
O c t 9
P r e s i d e n t F . D . R o o s e v e l t g i v e s g o a h e a d f o r t h e d e v e l o p m e n t o f t h e a t o m i c b o m b .
D e c 6
R o o s e v e l t a u t h o r i z e s t h e M a n h a t t a n E n g i n e e r i n g D i s t r i c t f o r t h e c r e a t i o n o f t h e a t o m i c b o m b .
1 9 4 2
S e p t 2 3
C o l o n e l L e s l i e G r o v e s p l a c e d i n c h a r g e o f t h e M a n h a t t a n P r o j e c t . P h y s i c i s t J . R o b e r t
O p p e n h e i m e r a p p o i n t e d S c i e n t i f i c D i r e c t o r .
1 9 4 4
M a y 1 6
E m i n e n t D a n i s h p h y s i c i s t N i e l s B o h r m e e t s B r i t i s h P r i m e M i n i s t e r W i n s t o n C h u r c h i l l i n
L o n d o n i n a n e f f o r t t o c o n v i n c e h i m o f t h e d a n g e r s o f a t o m i c w a r f a r e a n d a n a t o m i c a r m s r a c e
a n d t o s u g g e s t i n t e r n a t i o n a l c o n t r o l o f a t o m i c e n e r g y . T h e m e e t i n g f a i l s .
A u g 2 6
N i e l s B o h r m e e t s R o o s e v e l t i n W a s h i n g t o n , D . C . , i n a n e f f o r t t o c o n v i n c e h i m o f t h e d a n g e r s
o f a t o m i c w a r f a r e a n d a n a t o m i c a r m s r a c e a n d t o s u g g e s t i n t e r n a t i o n a l c o n t r o l o f a t o m i c
e n e r g y . R o o s e v e l t a p p e a r s r e c e p t i v e t o B o h r ’ s a r g u m e n t s .
1 9 4 5
F e b 4 - 1 1
Y a l t a C o n f e r e n c e . A h i g h p o i n t o f S o v i e t - W e s t r e l a t i o n s .
M a r c h
R a p i d d e t e r i o r a t i o n i n U . S . - S o v i e t r e l a t i o n s . “ R e l a t i o n s s a n k t o t h e i r l o w e s t p o i n t s i n c e 1 9 4 1 ”
b e c a u s e t h e S o v i e t s w o u l d n ’ t l e t U . S p l a n e s o r o f f i c e r s i n t o P o l a n d t o r e s c u e A m e r i c a n P O W s
w h o w e r e n o t d o i n g w e l l ; S t a l i n f e a r e d a t t e m p t s t o i n f i l t r a t e a n d s u r v e i l .
O t h e r c r i s e s
2 2 0
d e v e l o p e d o n S o v i e t c o n d u c t i n R o m a n i a a n d P o l a n d ; t h e B e r n c r i s i s e m e r g e d w h e n S t a l i n
f e a r e d t h e W e s t w o u l d m a k e a s e p a r a t e p e a c e . O n M a r c h 2 4 , R o o s e v e l t “ b a n g e d h i s f i s t s o n t h e
a r m s o f h i s w h e e l c h a i r a n d s a i d : ‘ A v e r e l l i s r i g h t ; w e c a n ’ t d o b u s i n e s s w i t h S t a l i n . H e h a s
b r o k e n e v e r y o n e o f h i s p r o m i s e s h e m a d e a t Y a l t a . ’ ”
2 2 1
A p r i l 1 2
R o o s e v e l t d i e s . H a r r y T r u m a n i s n a m e d t h e 3 3 r d P r e s i d e n t o f t h e U n i t e d S t a t e s .
M a y 1 4 , 3 1
a n d J u n e 1
I n t e r i m C o m m i t t e e d e l i b e r a t i o n s . M a y 1 4 : S c i e n t i f i c a d v i s o r s V a n n e v a r B u s h a n d J a m e s B .
C o n a n t s h a r e m e m o w a r n i n g o f a t o m i c a r m s r a c e .
2 2 2
J u n e 2 6
T h e U n i t e d N a t i o n s c h a r t e r i s s i g n e d b y d e l e g a t e s o f f i f t y n a t i o n s a t S a n F r a n c i s c o .
2 2 3
“ S i n c e t h e s t a r t o f t h e c o u n c i l , t h e r e f o r e , i t h a d b e c o m e i n c r e a s i n g l y a p p a r e n t t o B y r n e s t h a t t h e b o m b w a s a b u s t a s a
h o p e d - f o r m a s t e r c a r d i n d i p l o m a c y . A s t h e c o n f e r e n c e c o n t i n u e d , m o r e o v e r , i t b e g a n t o d a w n o n h i m t h a t A m e r i c a ' s
a t o m i c m o n o p o l y , f a r f r o m b e i n g a n a s s e t a t L o n d o n , m i g h t b e i n s t e a d a d i p l o m a t i c l i a b i l i t y t h e r e . ” H e r k e n , p . 4 9 .
2 2 4
P o w a s k i ,
M a r c h t o A r m a g e d d o n ,
p p . 3 0 - 1 .
2 2 5
T h e B u s h m e m o c a n b e f o u n d a t U n i t e d S t a t e s D e p a r t m e n t o f S t a t e ,
F o r e i g n R e l a t i o n s o f t h e U n i t e d S t a t e s : D i p l o m a t i c
P a p e r s , 1 9 4 5 . G e n e r a l : P o l i t i c a l a n d E c o n o m i c M a t t e r s
V o l u m e I I ( W a s h i n g t o n , D C : G o v e r n m e n t P r i n t i n g O f f i c e , 1 9 4 6 ) ,
p p . 6 9 - 7 3 . A v a i l a b l e a t :
h t t p : / / d i g i c o l l . l i b r a r y . w i s c . e d u / c g i - b i n / F R U S / F R U S - i d x ? t y p e = a r t i c l e & d i d = F R U S . F R U S 1 9 4 5 v 0 2 . i 0 0 0 5 & i d = F R U S . F R
U S 1 9 4 5 v 0 2 & i s i z e = M
. A c c e s s e d 2 4 A p r i l 2 0 1 9 . F o r c o n t e x t s e e : B e r n s t e i n , “ T h e Q u e s t f o r S e c u r i t y ; Z a c h a r y ,
E n d l e s s
F r o n t i e r
, p p . 3 0 9 - 1 0 .
6 4
J u l y 1 6
T r i n i t y T e s t o f t h e f i r s t a t o m i c b o m b a t A l a m o g o r d o , N e w M e x i c o .
J u l y 1 7 -
A u g 2
P o t s d a m C o n f e r e n c e .
A u g 6
U . S . d r o p s a t o m i c b o m b o n H i r o s h i m a .
A u g 9
U . S . d r o p s a t o m i c b o m b o n N a g a s a k i .
S e p t 1 1 -
O c t 2
F i r s t m e e t i n g o f C o u n c i l o n F o r e i g n M i n i s t e r s . H e l d i n L o n d o n . S e c r e t a r y o f S t a t e J a m e s F .
B y r n e s r e p r e s e n t s t h e U . S . a n d a t t e m p t s t o c a r r y o u t n e g o t i a t i o n s w i t h o u t i n v o l v i n g o t h e r
p o l i c y m a k i n g e l i t e s i n t h e U . S . , i n c l u d i n g o t h e r s i n t h e S t a t e D e p a r t m e n t . B y r n e s h o p e d t h a t
t h e a t o m i c b o m b w o u l d h e l p i n n e g o t i a t i o n s w i t h S o v i e t c o u n t e r p a r t M o l o t o v , b u t i t d i d n o t .
2 2 3
S e p t 2 1
S e c r e t a r y o f W a r H e n r y L . S t i m s o n p r e s e n t s h i s p r o p o s a l f o r i n t e r n a t i o n a l c o n t r o l t o t h e
C a b i n e t .
2 2 4
O c t
R e p r e s e n t a t i v e A n d r e w M a y ( c h a i r m a n o f t h e H o u s e M i l i t a r y A f f a i r s C o m m i t t e e ) a n d S e n a t o r
E d w i n C . J o h n s o n ( c h a i r m a n o f t h e S e n a t e M i l i t a r y A f f a i r s C o m m i t t e e ) i n t r o d u c e t h e
M a y - J o h n s o n B i l l i n C o n g r e s s o n t h e p r o p o s e d d o m e s t i c a t o m i c e n e r g y a g e n c y . T h e b i l l g o t
s t u c k o v e r t h e c o m i n g w e e k s a n d w a s e v e n t u a l l y s u p p l a n t e d b y t h e M c M a h o n B i l l .
N o v c . 1 - 1 5
V a n n e v a r B u s h p r e s e n t s h i s p l a n f o r i n t e r n a t i o n a l c o n t r o l t o t h e S t a t e D e p a r t m e n t , t h e
s o - c a l l e d B u s h P l a n . T h i s p l a n f o r m e d t h e b a s i s f o r S t a t e D e p a r t m e n t t h i n k i n g o n a t o m i c
c o o p e r a t i o n w i t h t h e S o v i e t U n i o n i n D e c e m b e r 1 9 4 5 .
2 2 5
N o v 1 5
B r i t a i n , C a n a d a , a n d t h e U . S . a g r e e o n t h e T h r e e N a t i o n D e c l a r a t i o n o n a t o m i c e n e r g y .
D e c 1 0
T h e S t a t e D e p a r t m e n t ’ s C o h e n - P a s v o l s k y C o m m i t t e e c o m p l e t e s a p l a n f o r t h e i n t e r n a t i o n a l
c o n t r o l o f a t o m i c e n e r g y . T h i s p l a n w o u l d i n f o r m S e c r e t a r y o f S t a t e B y r n e s ’ n e g o t i a t i o n s w i t h
t h e S o v i e t s i n M o s c o w i n D e c e m b e r 1 9 4 5 .
D e c 2 0
S e n a t o r B r i e n M c M a h o n i n t r o d u c e s i n t o t h e S e n a t e l e g i s l a t i o n f o r a n a l t e r n a t i v e
c i v i l i a n - o r i e n t e d a t o m i c e n e r g y b i l l ( k n o w n a s t h e M c M a h o n B i l l ) .
2 2 6
H e r k e n , p p . 6 9 - 8 5 .
2 2 7
“ R e p o r t o f t h e M e e t i n g o f t h e M i n i s t e r s o f F o r e i g n A f f a i r s o f t h e U n i o n o f S o v i e t S o c i a l i s t R e p u b l i c s , t h e U n i t e d S t a t e s
o f A m e r i c a , t h e U n i t e d K i n g d o m ” , 2 7 D e c e m b e r 1 9 4 5 . A t :
h t t p : / / a v a l o n . l a w . y a l e . e d u / 2 0 t h _ c e n t u r y / d e c a d e 1 9 . a s p
.
A c c e s s e d 2 4 A p r i l 2 0 1 9 . H e r k e n , p p . 8 7 - 8 .
2 2 8
“ M e m o r a n d u m b y t h e C o m m a n d i n g G e n e r a l , M a n h a t t a n E n g i n e e r D i s t r i c t ( G r o v e s ) ” , 2 J a n u a r y 1 9 4 6 , i n U . S .
D e p a r t m e n t o f S t a t e ,
F o r e i g n R e l a t i o n s o f t h e U n i t e d S t a t e s , 1 9 4 6
. G e n e r a l ; T h e U n i t e d N a t i o n s , V o l u m e I ( W a s h i n g t o n ,
D C : G o v e r n m e n t P r i n t i n g O f f i c e , 1 9 4 7 ) , d o c u m e n t 6 0 0 . A v a i l a b l e a t :
h t t p s : / / h i s t o r y . s t a t e . g o v / h i s t o r i c a l d o c u m e n t s / f r u s 1 9 4 6 v 0 1 / d 6 0 0 . A c c e s s e d 2 4 A p r i l 2 0 1 9
2 2 9
Q u i n c y W r i g h t , “ D r a f t f o r a C o n v e n t i o n o n A t o m i c E n e r g y ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
1 ( 1 A p r i l 1 9 4 6 ) , p p . 1 1 - 3 .
2 3 0
“ T h e C a r n e g i e D r a f t C o n v e n t i o n ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
1 - 2 , 2 ( 1 J u l y 1 9 4 6 ) , p p . 1 5 - 1 9 .
2 3 1
J . R . O p p e n h e i m e r , “ T h e A t o m B o m b a s a G r e a t F o r c e f o r P e a c e ” ,
T h e N e w Y o r k T i m e s
( 9 J u n e 1 9 4 6 ) .
6 5
D e c 1 6 - 2 6
T h e s e c o n d C o n f e r e n c e o f F o r e i g n M i n i s t e r s h e l d a t M o s c o w . A t t e n d e d b y t h e U n i t e d S t a t e s
( S e c r e t a r y o f S t a t e J a m e s F . B y r n e s ) , t h e U n i t e d K i n g d o m ( E r n e s t B e v i n ) , a n d t h e S o v i e t U n i o n
( V y a c h e s l a v M o l o t o v ) . B y r n e s a t t e n d s w i t h t h e a i m o f n e g o t i a t i n g a p a t h w a y t o c o o p e r a t i v e
c o n t r o l o f t h e a t o m i c b o m b w i t h t h e S o v i e t U n i o n , b a s e d o n t h e C o h e n - P a s v o l s k y p l a n . T h e
c o n f e r e n c e i s p a r t i a l l y s u c c e s s f u l , l e a d i n g t o s e v e r a l a g r e e d p o i n t s b e t w e e n t h e S o v i e t U n i o n
a n d t h e U n i t e d S t a t e s ( s e e b e l o w ) .
2 2 6
D e c 2 7
C o m m u n i q u e i s s u e d a f t e r t h e M o s c o w C o n f e r e n c e . I n c l u d e s a n a g r e e m e n t o n t h e
e s t a b l i s h m e n t o f a U n i t e d N a t i o n s c o m m i s s i o n f o r t h e c o n t r o l o f a t o m i c e n e r g y .
2 2 7
1 9 4 6
J a n 2
G e n e r a l G r o v e s r e l e a s e s h i s m e m o r a n d u m o n a t o m i c w e a p o n s t o C o n g r e s s :
O u r A r m y o f t h e
F u t u r e — A s I n f l u e n c e d b y A t o m i c W e a p o n s
.
2 2 8
J a n 2 4
U . N . G e n e r a l A s s e m b l y p a s s e s a r e s o l u t i o n c r e a t i n g t h e U . N . A t o m i c E n e r g y C o m m i s s i o n .
F e b 3
C o l u m n i s t D r e w P e a r s o n p u b l i s h e s a r e p o r t r e v e a l i n g t h a t a S o v i e t s p y r i n g , l e d b y B r i t i s h
p h y s i c i s t A l a n N u n n M a y , h a d b e e n d i s c o v e r e d i n C a n a d a .
F e b 2 2
G e o r g e K e n n a n , U . S . c h a r g é d ' a f f a i r e s i n M o s c o w , s e n d s h i s “ L o n g T e l e g r a m ” o n t h e S o v i e t
U n i o n .
M a r c h 5
F o r m e r B r i t i s h P r i m e M i n i s t e r W i n s t o n C h u r c h i l l d e l i v e r s h i s “ I r o n C u r t a i n ” s p e e c h a t
F u l t o n , M i s s o u r i .
M a r c h 1 8
F i n a n c i e r B e r n a r d B a r u c h a p p o i n t e d a s U . S . d e l e g a t e t o t h e U . N . A t o m i c E n e r g y C o m m i s s i o n .
A p r i l 1
P o l i t i c a l s c i e n t i s t Q u i n c y W r i g h t r e l e a s e s h i s p r o p o s a l f o r i n t e r n a t i o n a l c o n t r o l , t h e “ D r a f t f o r
a C o n v e n t i o n o n A t o m i c E n e r g y ” .
2 2 9
J u n e
T h e C a r n e g i e E n d o w m e n t f o r I n t e r n a t i o n a l P e a c e r e l e a s e s i t s p r o p o s a l f o r i n t e r n a t i o n a l
c o n t r o l , t i t l e d “ D r a f t C o n v e n t i o n o f t h e C a r n e g i e E n d o w m e n t C o m m i t t e e o n A t o m i c
E n e r g y ” .
2 3 0
J u n e 9
J . R o b e r t O p p e n h e i m e r p u b l i s h e s a n a r t i c l e i n t h e
N e w Y o r k T i m e s
e x p l a i n i n g h i s p o s i t i o n o n
i n t e r n a t i o n a l c o n t r o l . T h e a r t i c l e i s a d e f e n s e o f t h e A c h e s o n - L i l i e n t h a l R e p o r t a n d , t h u s ,
i m p l i c i t l y a c r i t i c i s m o f t h e B a r u c h P l a n .
2 3 1
2 3 2
“ U n i t e d N a t i o n s a n d A t o m i c E n e r g y … . ” ,
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
3 , 2 ( 1 9 4 7 ) , p p . 5 5 - 5 6 .
6 6
J u n e 1 4
B a r u c h p r e s e n t s t h e B a r u c h P l a n f o r i n t e r n a t i o n a l c o n t r o l a t t h e U . N . A t o m i c E n e r g y
C o m m i s s i o n .
J u n e 1 9
A n d r e i G r o m y k o p r e s e n t s t h e G r o m y k o P l a n a t t h e U . N . A t o m i c E n e r g y C o m m i s s i o n a n d
s t a t e s t h a t t h e S o v i e t U n i o n w o u l d n o t a c c e p t a n y r e v i s i o n o f t h e v e t o f o r a t o m i c i s s u e s .
J u l y 1
U . S . c o n d u c t s i t s f i r s t p o s t w a r a t o m i c b o m b t e s t ( “ S h o t A b l e ” ) a t B i k i n i A t o l l i n t h e M a r s h a l l
I s l a n d s .
J u l y 2 5
U . S . c o n d u c t s i t s s e c o n d p o s t w a r a t o m i c b o m b t e s t ( “ S h o t B a k e r ” ) a t B i k i n i A t o l l i n t h e
M a r s h a l l I s l a n d s .
A u g 1
P r e s i d e n t T r u m a n s i g n s t h e A t o m i c E n e r g y A c t . T h i s e s t a b l i s h e s t h e A t o m i c E n e r g y
C o m m i s s i o n ( A E C ) w h i c h a s s u m e s r e s p o n s i b i l i t y f o r a l l M a n h a t t a n E n g i n e e r i n g D i s t r i c t
( M E D ) p r o p e r t i e s .
S e p t 1 2
S e c r e t a r y o f C o m m e r c e H e n r y W a l l a c e g i v e s a p r o m i n e n t s p e e c h i n N e w Y o r k c r i t i c i z i n g U . S .
p o l i c y t o w a r d s t h e S o v i e t U n i o n , i n c l u d i n g t h e B a r u c h P l a n .
S e p t 1 8
A J u l y 2 3 l e t t e r f r o m W a l l a c e t o T r u m a n i s l e a k e d t o t h e p r e s s . T h e l e t t e r a t t a c k s t h e B a r u c h
P l a n .
S e p t 2 0
T r u m a n a s k s f o r W a l l a c e ’ s r e s i g n a t i o n a s S e c r e t a r y o f C o m m e r c e , a n d r e c e i v e s i t .
N o v 1 3
T h e U N A E C a p p r o v e s , b y a t e n - t o - t w o v o t e , t h e U . S . p r o p o s a l t o i s s u e a r e p o r t t o t h e U . N .
S e c u r i t y C o u n c i l b y t h e e n d o f t h e y e a r .
D e c 2 6
T h e S o v i e t a t o m i c p r o g r a m a c h i e v e s i t s f i r s t c h a i n r e a c t i o n i n a n e x p e r i m e n t a l n u c l e a r r e a c t o r .
1 9 4 7
J a n 1
T h e U . S . A t o m i c E n e r g y C o m m i s s i o n t a k e s o v e r t h e M a n h a t t a n P r o j e c t ' s r e s e a r c h a n d
p r o d u c t i o n f a c i l i t i e s f r o m t h e M a n h a t t a n E n g i n e e r D i s t r i c t .
J a n 4
B e r n a r d B a r u c h ’ s w i s h t o r e s i g n a s U . S . r e p r e s e n t a t i v e o n t h e U N A E C i s a c c e p t e d b y T r u m a n .
J a n 5
B e r n a r d B a r u c h a n d h i s g r o u p o f a d v i s o r s r e s i g n f r o m t h e U N A E C . B a r u c h i s r e p l a c e d b y
f o r m e r S e n a t o r W a r r e n R . A u s t i n a s t h e U . S . r e p r e s e n t a t i v e t o t h e U N A E C .
2 3 2
1 9 4 9
A u g 2 9
T h e S o v i e t U n i o n e x p l o d e s i t s f i r s t a t o m i c b o m b , n e a r S e m i p a l a t i n s k .
A p p e n d i x B : K e y H i s t o r i c a l F i g u r e s
S c i e n t i s t s
B o h r , N i e l s
( 1 8 8 5 – 1 9 6 2 ) - S e n i o r a n d h i g h l y r e s p e c t e d D a n i s h t h e o r e t i c a l p h y s i c i s t w h o s e r v e d a s a n a d v i s o r t o
t h e B r i t i s h g o v e r n m e n t i n 1 9 4 4 – 4 5 a n d v i s i t e d t h e U . S . a t t h a t t i m e , w o r k i n g b r i e f l y o n t h e M a n h a t t a n P r o j e c t .
C o n c e r n e d w i t h t h e d e v e l o p m e n t o f t h e a t o m i c b o m b , h e m e t C h u r c h i l l a n d t h e n R o o s e v e l t i n t h e s p r i n g a n d
s u m m e r o f 1 9 4 4 , u r g i n g t h e m t o b e g i n d i s c u s s i o n s w i t h t h e S o v i e t s f o r i n t e r n a t i o n a l c o n t r o l o f a t o m i c e n e r g y .
R o o s e v e l t a n d C h u r c h i l l d i d n o t f o l l o w B o h r ’ s s u g g e s t i o n s a n d i n s t e a d , i n a l a t e r m e e t i n g , e x p r e s s e d c o n c e r n
t h a t B o h r m a y l e a k a t o m i c s e c r e t s t o t h e S o v i e t U n i o n .
2 3 3
B u s h , V a n n e v a r
( 1 8 9 0 – 1 9 7 4 ) - H e a d o f t h e U . S . O f f i c e o f S c i e n t i f i c R e s e a r c h a n d D e v e l o p m e n t d u r i n g W o r l d
W a r T w o a n d o n e o f t h e t h r e e s e n i o r s c i e n t i f i c a d m i n i s t r a t o r s i n v o l v e d i n a t o m i c p o l i c y d u r i n g t h e w a r . I n m i d
1 9 4 5 , B u s h w a s a p r o p o n e n t o f a t o m i c d i s a r m a m e n t a n d , b y N o v e m b e r 1 9 4 5 , h a d f i n a l i z e d t h e s o - c a l l e d “ B u s h
P l a n ” f o r a t o m i c d i s a r m a m e n t .
T h e B u s h P l a n i n c o r p o r a t e d a s t e p - b y - s t e p p r o c e s s f o r i n t e r n a t i o n a l c o n t r o l
2 3 4
a n d u l t i m a t e l y d i s a r m a m e n t . B u s h r e m a i n e d a s t r o n g p r o p o n e n t o f a “ s t a g e d ” p r o c e s s o f i n t e r n a t i o n a l c o n t r o l
t h r o u g h t o t h e e n d o f 1 9 4 5 a n d 1 9 4 6 . H e c a l l e d f o r S e c r e t a r y o f S t a t e B y r n e s t o a d o p t a s t a g e d a p p r o a c h i n l a t e
1 9 4 5 a n d , i n e a r l y 1 9 4 6 , a r g u e d , a s p a r t o f t h e A c h e s o n c o m m i t t e e , t h a t t h e S t a t e D e p a r t m e n t r e p o r t o n
i n t e r n a t i o n a l c o n t r o l ( t h e s o - c a l l e d A c h e s o n - L i l i e n t h a l R e p o r t ) a d o p t a s t a g e d a p p r o a c h b e c a u s e i t w o u l d h e l p
“ o p e n u p ” t h e S o v i e t U n i o n . B u s h w a s d i s a p p o i n t e d w i t h t h e a p p o i n t m e n t o f B a r u c h a s t h e U . S . r e p r e s e n t a t i v e
a t t h e U N A E C .
2 3 5
C o n a n t , J a m e s B r y a n t
( 1 8 9 3 – 1 9 7 8 ) - C h e m i s t , P r e s i d e n t o f H a r v a r d U n i v e r s i t y ( 1 9 3 3 – 1 9 5 3 ) , a n d C h a i r o f
t h e N a t i o n a l D e f e n s e R e s e a r c h C o m m i t t e e d u r i n g t h e w a r . H e w a s o n e o f t h e t h r e e s e n i o r s c i e n t i f i c
a d m i n i s t r a t o r s i n v o l v e d i n a t o m i c p o l i c y d u r i n g t h e w a r . C o n a n t w a s a m e m b e r o f t h e S t a t e D e p a r t m e n t ’ s
A c h e s o n C o m m i t t e e c h a r g e d i n e a r l y 1 9 4 6 w i t h p r o d u c i n g a p r o p o s a l o n i n t e r n a t i o n a l c o n t r o l . D u r i n g h i s t i m e
o n t h e c o m m i t t e e , h e j o i n e d B u s h a n d G r o v e s i n p r e s s u r i n g t h e c o n s u l t a n t s , l e d b y L i l i e n t h a l , t o p r o d u c e a
s t a g e d i n t e r n a t i o n a l c o n t r o l s c h e m e .
2 3 6
E i n s t e i n , A l b e r t
( 1 8 7 9 – 1 9 5 5 ) - S e n i o r t h e o r e t i c a l p h y s i c i s t b a s e d a t t h e P r i n c e t o n I n s t i t u t e o f A d v a n c e d S t u d y
( 1 9 3 3 – 1 9 5 5 ) . L e n t h i s n a m e a n d c e l e b r i t y t o c a l l s f o r i n t e r n a t i o n a l c o n t r o l ( h e f o u n d e d t h e E m e r g e n c y
C o m m i t t e e o f A t o m i c S c i e n t i s t s w i t h L e o S z i l a r d i n 1 9 4 6 , w h i c h p u b l i c i z e d i n t e r n a t i o n a l c o n t r o l ) a n d w o r l d
2 3 3
T h e m o s t d e t a i l e d a c c o u n t s o f B o h r ’ s a c t i v i t i e s a r e M a r t i n J . S h e r w i n , “ N i e l s B o h r a n d t h e F i r s t P r i n c i p l e s o f A r m s
C o n t r o l ” , i n H e r m a n F e s h b a c h , T e t s u o M a t s u i , A l e x a n d r a O l e s o n ( e d . ) ,
N i e l s B o h r : P h y s i c s a n d t h e W o r l d
( A b i n g d o n :
R o u t l e d g e , 1 9 9 8 ) , p p . 3 1 9 - 3 0 ; A a s e r u d , “ T h e S c i e n t i s t a n d t h e S t a t e s m e n ” .
2 3 4
T h e B u s h m e m o c a n b e f o u n d a t “ M e m o r a n d u m b y t h e D i r e c t o r o f t h e O f f i c e o f S c i e n t i f i c R e s e a r c h a n d D e v e l o p m e n t
( B u s h ) t o t h e S e c r e t a r y o f S t a t e ” , 5 N o v e m b e r 1 9 4 5 , i n U . S . D e p a r t m e n t o f S t a t e ,
F o r e i g n R e l a t i o n s o f t h e U n i t e d S t a t e s :
D i p l o m a t i c P a p e r s , 1 9 4 5 . G e n e r a l : P o l i t i c a l a n d E c o n o m i c M a t t e r s
V o l u m e I I ( W a s h i n g t o n , D C : G o v e r n m e n t P r i n t i n g
O f f i c e , 1 9 4 6 ) , p p . 6 9 - 7 3 . A v a i l a b l e a t :
h t t p : / / d i g i c o l l . l i b r a r y . w i s c . e d u / c g i - b i n / F R U S / F R U S - i d x ? t y p e = a r t i c l e & d i d = F R U S . F R U S 1 9 4 5 v 0 2 . i 0 0 0 5 & i d = F R U S . F R
U S 1 9 4 5 v 0 2 & i s i z e = M
. A c c e s s e d 2 4 A p r i l 2 0 1 9 . F o r c o n t e x t , s e e B e r n s t e i n , “ T h e Q u e s t f o r S e c u r i t y ” ; Z a c h a r y ,
E n d l e s s
F r o n t i e r
, p p . 3 0 9 - 3 1 0 .
2 3 5
H e r k e n
,
p p . 1 5 7 , 1 6 1 .
2 3 6
V a n n e v a r B u s h a n d J a m e s B . C o n a n t , “ M e m o r a n d u m ” ( 3 0 S e p t e m b e r 1 9 4 4 ) . A v a i l a b l e a t :
h t t p s : / / n s a r c h i v e 2 . g w u . e d u / / N S A E B B / N S A E B B 1 6 2 / 1 . p d f
. A c c e s s e d 2 2 S e p t e m b e r 2 0 1 8 . A l s o : Z a c h a r y ,
E n d l e s s
F r o n t i e r ,
p . 2 4 3 ; J a m e s H e r s h b e r g , J a m e s B . C o n a n t :
H a r v a r d t o H i r o s h i m a a n d t h e M a k i n g o f t h e N u c l e a r A g e
( N e w
Y o r k : A l f r e d A . K n o p f , 1 9 9 3 ) , p p . 2 0 4 - 5 .
6 7
g o v e r n m e n t ( b y s u p p o r t i n g i n t e r n a t i o n a l i s t s s u c h a s R a y m o n d G r a m S w i n g a n d E m e r y R e v e s ) i n t h e l a t e 1 9 4 0 s ,
t h o u g h w a s h i m s e l f n o t p a r t i c u l a r l y a c t i v e .
2 3 7
O p p e n h e i m e r , J . R o b e r t
( 1 9 0 4 – 1 9 6 7 ) - S e n i o r t h e o r e t i c a l p h y s i c i s t a n d o n e o f a h a n d f u l o f s c i e n t i s t s i n v o l v e d
i n U . S . a t o m i c p o l i c y a t t h e h i g h e s t l e v e l s i n t h e 1 9 4 0 s . H e a d o f t h e L o s A l a m o s L a b o r a t o r y d u r i n g t h e
M a n h a t t a n P r o j e c t . M e m b e r o f t h e b o a r d o f c o n s u l t a n t s t h a t d r a f t e d t h e A c h e s o n - L i l i e n t h a l R e p o r t f o r t h e
S t a t e D e p a r t m e n t i n e a r l y 1 9 4 6 . T h e r e p o r t w a s l a r g e l y b a s e d o n h i s w o r k a n d i n p a r t i c u l a r o n a n e a r l i e r d r a f t
f o r i n t e r n a t i o n a l c o n t r o l t h a t h e p r e p a r e d ( t h e s o - c a l l e d “ O p p e n h e i m e r P l a n ” ) .
I n 1 9 4 7 , h e w a s a p p o i n t e d a s
2 3 8
t h e f i r s t C h a i r m a n o f t h e G e n e r a l A d v i s o r y C o m m i t t e e t o t h e n e w l y f o r m e d A t o m i c E n e r g y C o m m i s s i o n .
R a b i n o w i t c h , E u g e n e
( 1 9 0 1 – 1 9 7 3 ) - R u s s i a n - b o r n A m e r i c a n b i o p h y s i c i s t w h o w o r k e d a t t h e M e t - L a b a s p a r t
o f t h e M a n h a t t a n P r o j e c t . A l e a d i n g ( a n d e a r l y ) p r o p o n e n t o f i n t e r n a t i o n a l c o n t r o l i n t h e f o r t i e s a n d f o u n d i n g
m e m b e r o f t h e A t o m i c S c i e n t i s t s ’ M o v e m e n t . F o u n d e d a n d t h e n e d i t e d t h e
B u l l e t i n o f t h e A t o m i c S c i e n t i s t s
i n
1 9 4 5 u n t i l h i s d e a t h . R a b i n o w i t c h ( a n d h i s
B u l l e t i n
) s u p p o r t e d t h e A c h e s o n - L i l i e n t h a l P l a n , b u t h e a l s o h e l p e d
f e l l o w C h i c a g o a c a d e m i c Q u i n c y W r i g h t p r o d u c e h i s o w n ( l e s s e r k n o w n ) p l a n f o r i n t e r n a t i o n a l c o n t r o l .
2 3 9
S z i l a r d , L e o
( 1 8 9 8 – 1 9 6 4 ) - H u n g a r i a n - G e r m a n - A m e r i c a n p h y s i c i s t w h o w o r k e d a t t h e I l l i n o i s - b a s e d
M e t a l l u r g i c a l L a b o r a t o r y ( “ M e t - L a b ” ) d u r i n g t h e M a n h a t t a n P r o j e c t . A v o c a l a d v o c a t e f o r i n t e r n a t i o n a l
c o n t r o l f r o m 1 9 4 4 o n w a r d s . W h i l s t w o r k i n g o n t h e M a n h a t t a n P r o j e c t , S z i l a r d c i r c u l a t e d a p e t i t i o n a m o n g s t
s c i e n t i s t s c a l l i n g f o r J a p a n t o b e g i v e n a p u b l i c c h a n c e t o s u r r e n d e r p r i o r t o t h e u s e o f t h e a t o m i c b o m b .
2 4 0
C o - f o u n d e d t h e E m e r g e n c y C o m m i t t e e o f A t o m i c S c i e n t i s t s w i t h A l b e r t E i n s t e i n i n 1 9 4 6 .
M i l i t a r y
F o r r e s t a l , J a m e s V i n c e n t
( 1 8 9 2 – 1 9 4 9 ) - U . S . S e c r e t a r y o f t h e N a v y ( M a y 1 9 4 4 t o S e p t e m b e r 1 9 4 7 ) a n d
S e c r e t a r y o f D e f e n s e ( S e p t e m b e r 1 9 4 7 t o M a r c h 1 9 4 9 ) . D e e p l y d i s t r u s t f u l o f t h e S o v i e t U n i o n , F o r r e s t a l p u s h e d
T r u m a n t o t a k e a h a r d l i n e a g a i n s t t h e S o v i e t s a f t e r t h e w a r . H e e f f e c t i v e l y o p p o s e d i n t e r n a t i o n a l c o n t r o l a n d
B y r n e s ’ d i p l o m a c y ( c a l l i n g i t “ a p p e a s e m e n t ” ) i n l a t e 1 9 4 5 .
2 4 1
G r o v e s , L e s l i e R .
( 1 8 9 6 – 1 9 7 0 ) - S e n i o r U n i t e d S t a t e s A r m y C o r p s o f E n g i n e e r s o f f i c e r ( e v e n t u a l l y L i e u t e n a n t
G e n e r a l ) w h o w a s D i r e c t o r o f t h e M a n h a t t a n p r o j e c t f r o m 1 9 4 2 t o 1 9 4 7 , a n d t h e s i n g l e m o s t i n f l u e n t i a l p e r s o n
i n t h e U . S . a t o m i c p r o g r a m a t t h a t t i m e . H e w a s b r i e f l y h e a d o f t h e A r m e d F o r c e s S p e c i a l W e a p o n s P r o j e c t
( 1 9 4 7 – 4 8 ) u n t i l h i s r e t i r e m e n t i n 1 9 4 8 . E a r l y o n i n t h e M a n h a t t a n P r o j e c t , G r o v e s c a m e t o s e e t h e S o v i e t U n i o n
a s t h e U . S . ’ s m a i n p o s t w a r e n e m y , l a t e r n o t i n g t h a t “ T h e r e w a s n e v e r , f r o m a b o u t t w o w e e k s f r o m t h e t i m e I
t o o k c h a r g e o f t h i s P r o j e c t a n y i l l u s i o n o n m y p a r t b u t t h a t R u s s i a w a s o u r e n e m y . ”
H e w a s d i s t r u s t f u l o f t h e
2 4 2
S o v i e t U n i o n a n d a s t r o n g s u p p o r t e r o f a l a r g e U . S . b o m b p r o g r a m : i f “ t h e r e a r e t o b e a t o m i c w e a p o n s i n t h e
w o r l d w e m u s t h a v e t h e b e s t , t h e b i g g e s t a n d t h e m o s t ” h e t o l d C o n g r e s s i n J a n u a r y 1 9 4 6 . H e b e l i e v e d t h e U . S .
c o u l d a c q u i r e a g l o b a l m o n o p o l y o n h i g h - g r a d e u r a n i u m a n d t h o r i u m a n d b e g a n a n e f f o r t t o d o s o i n f a l l 1 9 4 2 .
T h i s b e l i e f w a s c r u c i a l f o r h i s b e l i e f a n d c l a i m t h a t R u s s i a w o u l d n o t g e t t h e b o m b f o r a l o n g t i m e ( 2 0 y e a r s h e
2 3 7
S i l v a n S . S c h w e b e r ,
E i n s t e i n & O p p e n h e i m e r : T h e M e a n i n g o f G e n i u s
( C a m b r i d g e , M A : H a r v a r d U n i v e r s i t y P r e s s ,
2 0 0 8 ) , p p . 7 4 - 8 1 .
2 3 8
H e r k e n , p p . 1 5 5 - 6 1 .
2 3 9
K i m b a l l - S m i t h ,
A P e r i l a n d a H o p e ,
p p . 4 5 2 - 3 .
2 4 0
K i m b a l l - S m i t h ,
A P e r i l a n d a H o p e ,
p p . 5 5 - 5 7 .
2 4 1
H e r k e n , p . 7 4 ; B e r n s t e i n , “ T h e Q u e s t f o r S e c u r i t y ” . .
2 4 2
H e r k e n , p . 1 0 6 ; M a r t i n J . S h e r w i n ,
A W o r l d D e s t r o y e d : H i r o s h i m a a n d I t s L e g a c i e s
( S t a n f o r d : S t a n f o r d U n i v e r s i t y P r e s s ,
2 0 0 3 ) , p . 6 2 .
6 8
c l a i m e d i n M a y 1 9 4 5 ) .
G r o v e s w a s c r u c i a l f o r p e r s u a d i n g T r u m a n a n d o t h e r s t h a t R u s s i a w o u l d n o t g e t t h e
2 4 3
b o m b f o r m a n y y e a r s .
2 4 4
P o l i t i c i a n s a n d O t h e r s
A c h e s o n , D e a n
( 1 8 9 3 – 1 9 7 1 ) - S t a t e s m a n a n d l a w y e r . U n d e r s e c r e t a r y o f t h e U . S . D e p a r t m e n t o f S t a t e f r o m
A u g u s t 1 9 4 5 t o J u n e 1 9 4 7 ; S e c r e t a r y o f S t a t e 1 9 4 9 – 1 9 5 3 . A l e a d i n g s u p p o r t e r o f i n t e r n a t i o n a l c o n t r o l w i t h i n
t h e S t a t e D e p a r t m e n t i n t h e f i r s t h a l f o f 1 9 4 6 . C h a i r o f t h e S p e c i a l S t a t e D e p a r t m e n t C o m m i t t e e t a s k e d w i t h
t h e p r e p a r a t i o n o f a p l a n f o r i n t e r n a t i o n a l c o n t r o l i n D e c e m b e r 1 9 4 5 . A c h e s o n w a s k e e n
t h a t t h e C o m m i t t e e
2 4 5
s u c c e e d i n i t s t a s k n o t o n l y b e c a u s e h e w a n t e d i n t e r n a t i o n a l c o n t r o l a n d a t o m i c c o o p e r a t i o n w i t h t h e S o v i e t
U n i o n , b u t a l s o b e c a u s e h e w a n t e d t h e C o m m i t t e e t o i n f o r m p o l i t i c i a n s a n d d i p l o m a t s o n a t o m i c m a t t e r s , a n d
b e c a u s e h e , l i k e B y r n e s , w a n t e d t o r e t a i n a s m u c h p o l i c y m a k i n g / e x p e r t i s e w i t h i n t h e S t a t e D e p a r t m e n t a s
p o s s i b l e ( a n d n o t l o s e i t t o G r o v e s o r s o m e m i l i t a r y c o m m i t t e e ) .
O n c e B a r u c h w a s a p p o i n t e d a s t h e U . S .
2 4 6
r e p r e s e n t a t i v e t o t h e U N A E C a n d b e g a n t o f o r m u l a t e h i s o w n p o l i c y o n i n t e r n a t i o n a l c o n t r o l , A c h e s o n
e m e r g e d a s o n e o f t h e m o s t p r o m i n e n t o p p o n e n t s o f B a r u c h ’ s p o l i c y m a k i n g ( a n d t h u s s u p p o r t e r o f t h e o r i g i n a l
A c h e s o n - L i l i e n t h a l R e p o r t ) .
2 4 7
B a r n a r d , C h e s t e r
( 1 8 8 6 – 1 9 6 1 ) - H e a d o f N e w J e r s e y B e l l T e l e p h o n e ( 1 9 2 7 – 1 9 4 8 ) a n d m e m b e r o f t h e b o a r d
o f c o n s u l t a n t s c h a r g e d w i t h p r e p a r i n g a r e p o r t o n i n t e r n a t i o n a l c o n t r o l i n J a n u a r y 1 9 4 6 .
B a r u c h , B e r n a r d
( 1 8 7 0 – 1 9 6 5 ) - S e n i o r f i n a n c i e r , a d m i n i s t r a t o r , a n d p o l i t i c a l c o n s u l t a n t . A p p o i n t e d ( i n
M a r c h 1 9 4 6 ) b y P r e s i d e n t T r u m a n a s t h e U n i t e d S t a t e s r e p r e s e n t a t i v e t o t h e U n i t e d N a t i o n s A t o m i c E n e r g y
C o m m i s s i o n ( U N A E C ) 1 9 4 6 – 1 9 4 7 . D r a f t e d t h e U . S . ’ s o f f i c i a l p r o p o s a l f o r i n t e r n a t i o n a l c o n t r o l , k n o w n a s t h e
B a r u c h P l a n , a n d p r e s e n t e d i t t o t h e U N A E C i n J u n e 1 9 4 6 .
B y r n e s , J a m e s F .
( 1 8 8 2 – 1 9 7 2 ) - S e n i o r g o v e r n m e n t o f f i c i a l i n v o l v e d i n a t o m i c p o l i c y m a k i n g 1 9 4 3 – 1 9 4 6 .
D i r e c t o r o f t h e O f f i c e o f W a r M o b i l i z a t i o n ( 1 9 4 3 – 1 9 4 5 ) a n d S e c r e t a r y o f S t a t e ( J u l y 1 9 4 5 t o J a n u a r y 1 9 4 7 ) .
G r o m y k o , A n d r e i
( 1 9 0 9 – 1 9 8 8 ) - S o v i e t d i p l o m a t a n d p o l i t i c i a n . S o v i e t p e r m a n e n t r e p r e s e n t a t i v e t o t h e
U n i t e d N a t i o n s a n d t o t h e U . N . A t o m i c E n e r g y C o m m i s s i o n i n 1 9 4 6 .
J o h n s o n , E d w i n C .
( 1 8 8 4 – 1 9 7 0 ) - D e m o c r a t i c S e n a t o r f o r C o l o r a d o a n d C h a i r m a n o f t h e S e n a t e M i l i t a r y
A f f a i r s C o m m i t t e e i n 1 9 4 5 . I n t r o d u c e d t h e M a y - J o h n s o n B i l l i n C o n g r e s s i n O c t o b e r 1 9 4 5 a l o n g s i d e A n d r e w J .
M a y . T h e b i l l p r o p o s e d t h e c r e a t i o n o f a d o m e s t i c a t o m i c e n e r g y a g e n c y .
2 4 8
L i l i e n t h a l , D a v i d E .
( 1 8 9 9 – 1 9 8 1 ) - S e n i o r t e c h n o c r a t a n d p u b l i c a d m i n i s t r a t o r ( h e h a d b e e n C h a i r m a n o f t h e
T e n n e s s e e V a l l e y A u t h o r i t y 1 9 4 1 – 1 9 4 6 ) . C h a i r o f t h e f i v e - m e m b e r b o a r d o f c o n s u l t a n t s c h a r g e d w i t h
p r e p a r i n g a r e p o r t o n i n t e r n a t i o n a l c o n t r o l f o r t h e S t a t e D e p a r t m e n t S p e c i a l C o m m i t t e e o n a t o m i c e n e r g y i n
J a n u a r y 1 9 4 6 ( t h e s o - c a l l e d A c h e s o n - L i l i e n t h a l R e p o r t w a s r e l e a s e d i n M a r c h 1 9 4 6 ) . A s t r o n g p r o p o n e n t o f
i n t e r n a t i o n a l c o n t r o l o f a t o m i c e n e r g y i n 1 9 4 6 . T h e f i r s t c h a i r o f t h e A t o m i c E n e r g y C o m m i s s i o n , O c t o b e r
1 9 4 6 t o F e b r u a r y 1 9 5 0 .
M a y , A n d r e w J .
( 1 8 7 5 – 1 9 5 9 ) - A H o u s e R e p r e s e n t a t i v e f r o m K e n t u c k y a n d c h a i r m a n o f t h e H o u s e M i l i t a r y
A f f a i r s C o m m i t t e e . I n O c t o b e r 1 9 4 , 5 h e i n t r o d u c e d t h e M a y - J o h n s o n B i l l i n C o n g r e s s o n t h e p r o p o s e d
2 4 3
H e r k e n , p p . 1 0 1 - 0 6 , 1 1 2 . G o r d i n ,
R e d C l o u d a t D a w n ,
p p . 7 3 - 5 .
2 4 4
H e r k e n , p p . 1 1 0 - 1 1 ; Z i e g l e r , “ I n t e l l i g e n c e A s s e s s m e n t s o f S o v i e t A t o m i c C a p a b i l i t y , 1 9 4 5 - 1 9 4 9 ” ; Z i e g l e r a n d J a c o b s o n ,
S p y i n g W i t h o u t S p i e s
, p p . 2 4 - 5 .
2 4 5
W a q a r Z a i d i ,
D i d D e a n A c h e s o n W a n t I n t e r n a t i o n a l C o n t r o l o f A t o m i c E n e r g y ?
( 2 0 1 8 ) .
2 4 6
H e r k e n , p . 1 5 3 ; O n A c h e s o n b e i n g l e s s h a w k i s h t h a n m a n y a r o u n d i n 1 9 4 5 a n d e a r l y 1 9 4 6 , s e e R o b e r t L . B e i s n e r ,
D e a n
A c h e s o n : A L i f e i n t h e C o l d W a r
( O x f o r d : O x f o r d U n i v e r s i t y P r e s s , 2 0 0 6 ) , p p . 2 8 - 3 0 .
2 4 7
M a d d o c k ,
N u c l e a r A p a r t h e i d ,
p . 5 7 .
2 4 8
H e w l e t t a n d A n d e r s o n , J r . ,
A H i s t o r y o f t h e U n i t e d S t a t e s A t o m i c E n e r g y C o m m i s s i o n
v o l . 1 , p . 4 2 8 .
6 9
d o m e s t i c a t o m i c e n e r g y a g e n c y . T h e M a y - J o h n s o n B i l l w a s d r a f t e d l a r g e l y b y t h e W a r D e p a r t m e n t a n d p l a c e d
d o m e s t i c c o n t r o l o f a t o m i c e n e r g y u n d e r t h e c o n t r o l o f t h e m i l i t a r y . I t a l s o r e s t r i c t e d t h e P r e s i d e n t ’ s p o w e r o n
a t o m i c m a t t e r s , r e q u i r e d s t r i c t s e c u r i t y a n d s e c r e c y r e g u l a t i o n s , a n d m a n d a t e d h e a v y p e n a l t i e s f o r a n y s e c u r i t y
v i o l a t i o n s . T h e b i l l w a s s u p p o r t e d b y t h e t h r e e p r o m i n e n t s c i e n t i s t a d m i n i s t r a t o r s ( V a n n e v a r B u s h , J a m e s B .
C o n a n t , a n d R o b e r t J . O p p e n h e i m e r ) b u t o p p o s e d b y m o s t s c i e n t i s t s w o r k i n g o n a t o m i c m a t t e r s .
2 4 9
M c M a h o n , B r i e n
( 1 9 0 3 – 1 9 5 2 ) - S e n a t o r a n d k e y a l l y o f t h e A t o m i c S c i e n t i s t s ’ M o v e m e n t i n t h e S e n a t e .
A u t h o r o f t h e M c M a h o n A c t i n t r o d u c e d i n t o C o n g r e s s i n D e c e m b e r 1 9 4 5 . C h a i r o f t h e S e n a t e A t o m i c E n e r g y
C o m m i t t e e ( O c t o b e r 1 9 4 5 t o A u g u s t 1 9 4 6 ) a n d s u b s e q u e n t l y C h a i r o f t h e J o i n t A t o m i c E n e r g y C o m m i t t e e
( A u g u s t 1 9 4 6 t o J a n u a r y 1 9 4 7 , J a n u a r y 1 9 4 9 t o J u l y 1 9 5 2 ) . T h e M c M a h o n A c t w a s d r a f t e d w i t h t h e a s s i s t a n c e
o f p r o m i n e n t s c i e n t i s t s a n d t h e A t o m i c S c i e n t i s t s ’ M o v e m e n t a s a n a l t e r n a t i v e t o t h e m i l i t a r y ’ s M a y - J o h n s o n
B i l l . T h e b i l l p l a c e d a t o m i c e n e r g y u n d e r t h e c o n t r o l o f a c i v i l i a n A t o m i c E n e r g y C o m m i s s i o n , a n d e m p h a s i z e d
t h e p e a c e f u l u s e s o f a t o m i c e n e r g y . I t a l s o e m p h a s i z e d t h e n e e d f o r f r e e e x c h a n g e o f s c i e n t i f i c i n f o r m a t i o n t o t h e
f u l l e s t e x t e n t p o s s i b l e . T h e b i l l w a s p a s s e d i n t o l a w a s t h e A t o m i c E n e r g y A c t o n A u g u s t 1 , 1 9 4 6 , b u t o n l y a f t e r
s i g n i f i c a n t a m e n d m e n t s w h i c h w a t e r e d d o w n i t s c i v i l i a n f o c u s a n d i n c r e a s e d p e n a l t i e s .
2 5 0
P a s v o l s k y , L e o
( 1 8 9 3 – 1 9 5 3 ) - I n t e r n a t i o n a l i s t S t a t e D e p a r t m e n t o f f i c i a l w h o p l a y e d a l e a d i n g r o l e i n p o s t w a r
p l a n n i n g ( i n c l u d i n g f o r t h e U n i t e d N a t i o n s O r g a n i z a t i o n ) . A s t r o n g s u p p o r t e r o f t h e U n i t e d N a t i o n s a n d
i n t e r n a t i o n a l c o n t r o l a n d c o - d r a f t e d a n e a r l y p r o p o s a l f o r i n t e r n a t i o n a l c o n t r o l i n N o v e m b e r / D e c e m b e r 1 9 4 5
( t h e s o - c a l l e d C o h e n - P a s v o l s k y P l a n ) .
2 5 1
S e a r l s , J r . , F r e d
( 1 8 8 8 – 1 9 6 8 ) - M i n i n g e n g i n e e r a n d c o n s u l t a n t t o S e c r e t a r y o f S t a t e B y r n e s i n 1 9 4 5 / 4 6 .
A p p o i n t e d b y B a r u c h a s a d v i s o r o n a t o m i c e n e r g y a n d a s A l t e r n a t e D e l e g a t e t o t h e U N A E C . H a d s o m e i n p u t
i n t o t h e p r o d u c t i o n o f t h e B a r u c h P l a n a n d a d v o c a t e d f o r t h e c r e a t i o n o f a n A t o m i c L e a g u e .
2 5 2
S t i m s o n , H e n r y L .
( 1 8 6 7 – 1 9 5 0 ) - U . S . S e c r e t a r y o f S t a t e , J u l y 1 9 4 0 t o S e p t e m b e r 1 9 4 5 . R e t i r e d i n S e p t e m b e r
1 9 4 5 . A c c o r d i n g t o H e r k e n , d u r i n g t h e w a r h e h a d h o p e d t h a t i n t e r n a t i o n a l c o n t r o l m i g h t b e p o s s i b l e , b u t t h e
1 9 4 5 P o t s d a m C o n f e r e n c e l e d h i m t o b e l i e v e t h a t t h e S o v i e t s w o u l d n o t a c c e p t i t .
I n S e p t e m b e r 1 9 4 5 , p r i o r
2 5 3
t o l e a v i n g o f f i c e , h e o n c e a g a i n b e c a m e a n a d v o c a t e o f i n t e r n a t i o n a l c o n t r o l .
2 5 4
T h o m a s , C h a r l e s
( 1 9 0 0 – 1 9 8 2 ) - C h e m i s t a n d V i c e - P r e s i d e n t o f M o n s a n t o C h e m i c a l C o m p a n y . F r o m 1 9 4 3 t o
1 9 4 5 , h e c o o r d i n a t e d M a n h a t t a n P r o j e c t w o r k o n p l u t o n i u m p u r i f i c a t i o n a n d p r o d u c t i o n , a n d l a t e r b e c a m e
D i r e c t o r o f t h e C l i n t o n L a b o r a t o r i e s a t O a k R i d g e , T e n n e s s e e ( 1 9 4 5 – 1 9 4 8 ) .
M e m b e r o f t h e b o a r d o f
c o n s u l t a n t s t o t h e S t a t e D e p a r t m e n t c h a r g e d w i t h p r e p a r i n g a r e p o r t o n i n t e r n a t i o n a l c o n t r o l i n J a n u a r y 1 9 4 6 .
V a n d e n b e r g , A r t h u r
( 1 8 8 4
–
1 9 5 1 ) - R e p u b l i c a n S e n a t o r f r o m M i c h i g a n , C h a i r m a n o f t h e S e n a t e C o m m i t t e e
o n F o r e i g n R e l a t i o n s . A n a v o w e d i n t e r n a t i o n a l i s t , h e n e v e r t h e l e s s p u s h e d f o r a s t r o n g s t a n c e t o w a r d s t h e S o v i e t
U n i o n i n l a t e 1 9 4 5 a n d 1 9 4 6 . H e o p p o s e d B y r n e s ’ a t o m i c d i p l o m a c y i n l a t e 1 9 4 5 , s e e i n g i t a s o f f e r i n g
“ c o m p r o m i s e ” t o t h e S o v i e t U n i o n a n d “ s h e e r a p p e a s e m e n t . ”
2 5 5
W a l l a c e , H e n r y
( 1 8 8 8
–
1 9 6 5 ) - V i c e - P r e s i d e n t u n d e r F D R o o s e v e l t ( 1 9 4 1
–
1 9 4 5 ) , a n d s u b s e q u e n t l y S e c r e t a r y
o f C o m m e r c e ( 1 9 4 5
–
1 9 4 6 ) . E m e r g e d a s a p r o m i n e n t a n d p u b l i c c r i t i c o f U . S . p o l i c y t o w a r d s t h e S o v i e t U n i o n
2 4 9
H e w l e t t a n d A n d e r s o n , J r . ,
A H i s t o r y o f t h e U n i t e d S t a t e s A t o m i c E n e r g y C o m m i s s i o n
v o l . 1 , p p . 4 1 1 - 1 8 , 4 3 1 - 3 9 . W a n g ,
A m e r i c a n S c i e n c e i n a n A g e o f A n x i e t y
, p p . 1 3 - 1 5 .
2 5 0
W a n g ,
A m e r i c a n S c i e n c e i n a n A g e o f A n x i e t y
, p p . 1 8 - 2 4 .
2 5 1
O n t h e C o h e n - P a s v o l s k y P l a n , s e e P o w a s k i ,
M a r c h t o A r m a g e d d o n
, p . 3 5 .
2 5 2
O n h i s i n p u t , s e e L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a
, p p . 2 7 6 - 7 7 .
2 5 3
H e r k e n , p . 1 9 .
2 5 4
P o w a s k i ,
M a r c h t o A r m a g e d d o n
, p p . 2 9 - 3 0 .
2 5 5
H e r k e n , p p . 7 5 - 6 . L i e b e r m a n ,
T h e S c o r p i o n a n d t h e T a r a n t u l a
, p p . 2 1 8 - 2 2 .
7 0
i n l a t e 1 9 4 6 . T h i s i n c l u d e d s t r o n g p u b l i c c r i t i c i s m o f t h e B a r u c h P l a n — c e r t a i n l y t h e s t r o n g e s t p u b l i c c r i t i c i s m
f r o m w i t h i n t h e g o v e r n m e n t . F o r t h i s c r i t i c i s m h e w a s f o r c e d t o r e s i g n i n S e p t e m b e r 1 9 4 6 .
2 5 6
W i n n e , H a r r y A
. ( 1 8 8 8 – 1 9 6 8 ) - V i c e - P r e s i d e n t i n c h a r g e o f e n g i n e e r i n g a t G e n e r a l E l e c t r i c a n d m e m b e r o f t h e
b o a r d o f c o n s u l t a n t s c h a r g e d w i t h p r e p a r i n g a r e p o r t o n i n t e r n a t i o n a l c o n t r o l i n J a n u a r y 1 9 4 6 .
2 5 6
H e w l e t t a n d A n d e r s o n , J r . ,
A H i s t o r y o f t h e U n i t e d S t a t e s A t o m i c E n e r g y C o m m i s s i o n
v o l . 1 , p p . 5 9 7 - 6 0 6 .
7 1
|
f0142195-ca90-4084-bea7-16ea263ae6be
|
StampyAI/alignment-research-dataset/arbital
|
Arbital
|
How many bits to a trit?
There are [$\log_2](https://arbital.com/p/3nd) [bits](https://arbital.com/p/3p0) to a [https://arbital.com/p/3ww](https://arbital.com/p/3ww). This can be interpreted a few different ways:
1. If you multiply the number of messages you might want to send by 3, then the cost of encoding the message will go up by 1.58 bits on average. See [https://arbital.com/p/+marginal_message_cost](https://arbital.com/p/+marginal_message_cost) for more on this interpretation.
2. If you pack $n$ of independent and equally likely [3-messages](https://arbital.com/p/3v9) together into one giant $3^n$ message, then the cost (in bits) per individual 3-message drops as $n$ grows, ultimately converging to $\log_2(3)$ bits per 3-message as $n$ gets very large. For more on this, see [https://arbital.com/p/+average_message_cost](https://arbital.com/p/+average_message_cost) and the [GalCom example of encoding trits using bits](https://arbital.com/p/3zn).
3. The infinite expansion of $\log_2(3) = 1.58496250072\ldots$ tells us not just how many bits it takes to send one 3-message $(\approx \lceil 1.585 \rceil = 2)$ but also how long it takes to send any number of 3-messages put together. For example, it costs 2 bits to send one 3-message; 16 bits to send 10; 159 bits to send 1000; 1585 to send 10,000; 15850 to send 100,000; 158497 to send 1,000,000; and so on. For more on this interpretation, see the ["series of ceilings"](https://arbital.com/p/log_series_of_ceilings) interpretation of logarithms.
|
757f3e05-ac88-472c-b38e-91dd590d0130
|
StampyAI/alignment-research-dataset/lesswrong
|
LessWrong
|
Naive TDT, Bayes nets, and counterfactual mugging
I set out to understand precisely why naive TDT (possibly) fails the [counterfactual mugging](http://wiki.lesswrong.com/wiki/Counterfactual_mugging) problem. While doing this I ended up drawing a lot of [Bayes nets](http://en.wikipedia.org/wiki/Bayes_net), and seemed to gain some insight; I'll pass these on, in the hopes that they'll be useful. All errors are, of course, my own.
The grand old man of decision theory: the Newcomb problem
---------------------------------------------------------
First let's look at the problem that inspired all this research: the [Newcomb problem](http://wiki.lesswrong.com/wiki/Newcomb's_problem). In this problem, a supremely-insightful-and-entirely-honest superbeing called Omega presents two boxes to you, and tells you that you can either choose box A only ("1-box"), or take box A and box B ("2-box"). Box B will always contain $1K (one thousand dollars). Omega has predicted what your decision will be, though, and if you decided to 1-box, he's put $1M ([one million dollars](http://www.youtube.com/watch?v=l91ISfcuzDw)) in box A; otherwise he's put nothing in it. The problem can be cast as a Bayes net with the following nodes:
Your decision algorithm (or your your decision process) is the node that determines what you're going to decide. This leads to "Your decision" (1-box or 2-box) and Ω (puts $1M or zero in box A). These lead to the "Money" node, where you can end up with $1M+1K, $1M, $1K or $0 depending on the outputs of the other nodes. Note that the way the network is set up, you can never have $1M+1K or $0 (since "Ω" and "Your decision" are not independent). But it is the implied "possibility" of getting those two amounts that causes causal decision theory to 2-box in the Newcomb problem.
In TDT, as I understand it, you sever your decision algorithm node from the history of the universe (note this is incorrect, as explained [here](/lw/f37/naive_tdt_bayes_nets_and_counterfactual_mugging/7oli). In fact you condition on the start of your program, and *screen out* the history of the universe), and then pick the action that maximises our utility.
But note that the graph is needlessly complicated: "Your decision" and "Ω" are both superfluous nodes, that simply pass on their inputs to their outputs. Ignoring the "History of the Universe", we can reduce the net to a more compact (but less illuminating) form:
Here 1-box leads to $1M and 2-box leads to $1K. In this simplified version, the decision is obvious - maybe too obvious. The decision was entirely determined by the choice of how to lay out the Bayes net, and a causal decision theorist would disagree that the original "screened out" Bayes net was a valid encoding of the Newcomb problem.
The counterfactual mugging
--------------------------
In the counterfactual mugging, Omega is back, this time explaining that he tossed a coin. If the coin came up tails, he would have asked you to give him $1K, giving nothing in return. If the coin came up heads, he would have given you $1M - but only if when you would have given him the $1K in the tails world. That last fact he would have known by predicting your decision. Now Omega approaches you, telling you the coin was tails - what should you do? Here is a Bayes net with this information:
I've removed the "History of the Universe" node, as we are screening it off anyway. Here "Simulated decision" and "Your decision" will output the same decision on the same input. Ω will behave the way he said, based on your simulated decision given tails. "Coin" will output heads or tails with 50% probability, and "Tails" simply outputs tails, for use in Ω's prediction.
Again, this graph is very elaborate, codifying all the problem's intricacies. But most of the nodes are superfluous for our decision, and the graph can be reduced to:
"Coin" outputs "heads" or "tails" and "Your decision algorithm" outputs "Give $1K on tails" or "Don't give $1K on tails". Money is $1M if it receives "heads" and "Give $1K on tails", -$1K if it receives "tails" and "Give $1K on tails", and zero if receives "Don't give $1K on tails" (independent of the coin results).
If our utility does not go down [too sharply in money](/lw/e45/risk_aversion_does_not_explain_peoples_betting/), we should choose "Give $1K on tails", as a 50-50 bet on willing $1M and losing $1K is better than getting nothing with certainty. So precommitting to giving Omega $1K when he asks, leads to the better outcome.
But now imagine that we are in the situation above: Omega has come to us and explained that yes, the coin has come up tails. The Bayes net now becomes:
In this case, the course is clear: "Give $1K on tails" does nothing but lose us $1K. So we should decide not to - and nowhere in this causal graph can we see any problem with that course of action.
So it seems that naive TDT has an inconsistency problem. And these graphs don't seem to fully encode the actual problem properly (ie that the action "Give $1K on tails" corresponds to situations where we truly believe that tails came up).
Thoughts on the problem
-----------------------
Some thoughts that occurred when formalising this problem:
1. The problem really is with updating on information, vindicating the instincts behind [updateless decision theory](/lw/15m/towards_a_new_decision_theory/). The way you would have to behave, conditional on seeing new information, is different from how you want to behave, after seeing that new information.
2. Naive TDT reaches different conclusions depending on whether Omega simulates you or predicts you. If you are unsure whether you are being simulated or not (but still care about the wealth of the non-simulated version), then TDT acts differently on updates. Being told "tails" doesn't actually confirm that the coin was tails: you might be the simulated version, being tested by Omega. Note that in this scenario, the simulated you is being lied to by the simulated Omega (the "real" coin need not have been tails), which might put the problem in a different perspective.
3. The tools of TDT (Bayes nets cut at certain connection) feel inadequate. It's tricky to even express the paradox properly in this language, and even more tricky to know what to do about it. A possible problem seems to be that we don't have a way of expressing our own knowledge about the model, within the model - hence "tails" ends up being a fact about the universe, no a fact about our knowledge at the time. Maybe we need to make our map explicit [in the territory](/lw/erp/skill_the_map_is_not_the_territory/), and get Bayes nets that go something like these:
|
9026d22a-2ae0-4d9d-b9c0-cfdc355be301
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Notes on Perseverance
This post examines the virtues of perseverance, persistence, resilience, grit, fortitude, tenacity, sisu, and others in that bailiwick. It is meant mostly as an exploration of what other people have learned about these virtues, rather than as me expressing my own opinions about them, though I’ve been selective about what I found interesting or credible, according to my own inclinations. I wrote this not as an expert on the topic, but as someone who wants to learn more and to become better at it. I hope it will be helpful to people who want to know more about these virtues and how to nurture them.
What are these virtues?
People who exhibit these virtues rebound from setbacks and are not easily discouraged. They are more apt to overcome challenges than to be overcome by them.
Some of these virtue-words emphasize coping-with-adversity (resilience, fortitude), others emphasize continued effort (tenacity, perseverance, persistence), and others are a combination of those and other things (grit, sisu).
“Grit” in particular has been investigated under the positive psychology banner in recent years, and so deserves some extra attention here. As the researchers define it, grit is a combination of perseverance and passion. Perseverance in turn is defined as consistency of effort, while passion, less intuitively, is defined as consistency of interests. (Passion, in other words, has less to do with the level of present enthusiasm than with sustained interest over the long term. Researchers measure passion in people by asking them how little they identify with statements like “new ideas and projects sometimes distract me from previous ones,” or “my interests change from year to year.”[1]) Grit is correlated with (and, some critics say, difficult to distinguish from) “conscientiousness” in the “big five” model of personality.
“Sisu” seems to be a sort of in extremis perseverance: persistence to the bitter end in spite of terrible odds, when you feel you have already given you
|
6ab3a6d0-0fed-4e20-a206-caf92561be3f
|
StampyAI/alignment-research-dataset/blogs
|
Blogs
|
Why Now Matters
I’m often asked whether [donations now](https://intelligence.org/2015/07/17/miris-2015-summer-fundraiser/) are more important than donations later. Allow me to deliver an emphatic *yes*: I currently expect that donations to MIRI today are worth much more than donations to MIRI in five years. As things stand, I would very likely take $10M today over $20M in five years.
That’s a bold statement, and there are a few different reasons for this. First and foremost, there is a decent chance that some very big funders will start entering the AI alignment field over the course of the next five years. It looks like the NSF may start to fund AI safety research, and Stuart Russell has already received some money from DARPA to work on value alignment. It’s quite possible that in a few years’ time significant public funding will be flowing into this field.
(It’s also quite possible that it *won’t*, or that the funding will go to all the wrong places, as was the case with funding for nanotechnology. But if I had to bet, I would bet that it’s going to be much easier to find funding for AI alignment research in five years’ time).
In other words, the funding bottleneck is loosening — but it isn’t loose yet.
We don’t presently have the funding to grow [as fast as we could](https://intelligence.org/2015/07/18/targets-1-and-2-growing-miri/) over the coming months, or to run all the important research programs we have planned. At our current funding level, the research team can grow at a steady pace — but we could get much more done over the course of the next few years if we had the money to grow as fast as is healthy.
Which brings me to the second reason why funding now is probably much more important than funding later: because growth now is much more valuable than growth later.
There’s an idea picking up traction in the field of AI: instead of focusing only on increasing the capabilities of intelligent systems, it is important to also ensure that we know how to build *beneficial* intelligent systems. Support is growing for a new paradigm within AI that seriously considers the long-term effects of research programs, rather than just the immediate effects. Years down the line, these ideas may seem obvious, and the AI community’s response to these challenges may be in full swing. Right now, however, there is relatively little consensus on how to approach these issues — which leaves room for researchers today to help determine the field’s future direction.
People at MIRI have been thinking about these problems for a long time, and that puts us in an unusually good position to influence the field of AI and ensure that some of the growing concern is directed towards long-term issues in addition to shorter-term ones. We can, for example, help avert a scenario where all the attention and interest generated by Musk, Bostrom, and others gets channeled into short-term projects (e.g., making drones and driverless cars safer) without any consideration for long-term risks that are less well-understood.
It’s likely that MIRI will scale up substantially at some point; but if that process begins in 2018 rather than 2015, it is plausible that we will have already missed out on a number of big opportunities.
The alignment research program within AI is just now getting started in earnest, and it may even be funding-saturated in a few years’ time. But it’s nowhere near funding-saturated today, and waiting five or ten years to begin seriously ramping up our growth would likely give us far fewer opportunities to shape the methodology and research agenda within this new AI paradigm. The projects MIRI takes on today can make a big difference years down the line, and supporting us today will drastically affect how much we can do quickly. Now matters.
The post [Why Now Matters](https://intelligence.org/2015/07/20/why-now-matters/) appeared first on [Machine Intelligence Research Institute](https://intelligence.org).
|
e9a86618-4804-41b5-a1f9-8409eb9bf035
|
StampyAI/alignment-research-dataset/eaforum
|
Effective Altruism Forum
|
Prior probability of this being the most important century
Summary
=======
* I think the prior probability of this being the [most important century](https://forum.effectivealtruism.org/s/isENJuPdB3fhjWYHd) is less than 1 % based on i) the self-sampling assumption ([SSA](https://www.lesswrong.com/tag/self-sampling-assumption)), ii) a loguniform distribution of the effective computation this century as a fraction of that throughout all time, and iii) all effective computation being more than 10^44 times as large as that this century.
* The bigger the influentialness of the future, the lower the probability of this century being influential.
Methods
=======
In agreement with section 2 of [Mogensen 2022](https://globalprioritiesinstitute.org/wp-content/uploads/Mogensen-The-Hinge-of-History-Hypothesis-reply-to-MacAskill.pdf), the self-sampling assumption ([SSA](https://www.lesswrong.com/tag/self-sampling-assumption)) says the prior probability of being the most influential person among a population of size N, comprising past, present and future people, is the [expected value](https://forum.effectivealtruism.org/topics/expected-value) of 1/N (not the reciprocal of the expected value of N). To be more general, I have estimated the prior probability of this being the [most important century](https://forum.effectivealtruism.org/s/isENJuPdB3fhjWYHd) from the expected value of effective computation this century as a fraction of that throughout all time. By effective computation, I mean one which [counterfactually](https://forum.effectivealtruism.org/topics/counterfactual-reasoning) increases/decreases welfare, regardless of whether it comes from biological beings or [artificial intelligence](https://forum.effectivealtruism.org/topics/artificial-intelligence) (AI).
[Artificial general intelligence](https://forum.effectivealtruism.org/topics/artificial-intelligence) (AGI) [might](https://forum.effectivealtruism.org/posts/KqCybin8rtfP3qztq/agi-and-lock-in) soon cause [lock-in](https://forum.effectivealtruism.org/topics/value-lock-in) to a certain extent. If this does not involve total loss of value[[1]](#fn8o0yu4uw46c), the vast majority of computation could still happen in the longterm future, but one would not be able to counterfactually increase/decrease welfare.
In addition, this might be the most important century owing to high risk of [human extinction](https://forum.effectivealtruism.org/topics/human-extinction). However, a priori, one should arguably expect such risk to also be proportional to the expected value of effective computation this century as a fraction of that throughout all time. Along the same lines, [I wonder](https://forum.effectivealtruism.org/posts/apZwBKDope6xqP3CT/will-the-vast-majority-of-technological-progress-happen-in) whether the vast majority of technological progress will happen in the longterm future.
I modelled the effective computation this century as a fraction of that throughout all time as a [loguniform](https://en.wikipedia.org/wiki/Reciprocal_distribution) distribution, which is a good [uninformative prior](https://www.statlect.com/fundamentals-of-statistics/uninformative-prior). Its expected value is (b - a)/(ln(b) - ln(a)), where a and b = 1 are its minimum and maximum. As a approaches 0 for b = 1, the expected value tends to 1/ln(1/a). In other words, the prior probability of this being the most important century is [inversely proportional](https://en.wikipedia.org/wiki/Proportionality_(mathematics)%23Inverse_proportionality) to the logarithm of the maximum effective computation throughout all time as a fraction of that this century. The bigger the influentialness of the future, the lower the probability of this century being influential.
Results
=======
The results are below. My calculations are in [this](https://docs.google.com/spreadsheets/d/1B-xRn2JuDQjNzCGAfINNFpRlYm0xYC0CSz53ALNRJ_4/edit?usp%3Dsharing) Sheet.
Discussion
==========
The future can hold at least as many as 10^54 lives according to [Newberry 2021](https://globalprioritiesinstitute.org/wp-content/uploads/Toby-Newberry_How-many-lives-does-the-future-hold.pdf) (Table 1), and this century can have less than 10 billion (10^10) lives. Consequently, 1/a is lower than 10^(54 - 10) = 10^44, which suggests a prior probability of this being the most important century of less than 0.987 %.
Note I would have (wrongly) arrived at a much lower probability based on the reciprocal of the expected effective computation throughout all time as a fraction of that this century. For the conservative estimate of 10^28 expected future lives given in [Newberry 2021](https://globalprioritiesinstitute.org/wp-content/uploads/Toby-Newberry_How-many-lives-does-the-future-hold.pdf) (Table 3), which is at least 10^(28 - 10) = 10^18 times as many as the lives this century, I would get a prior probability of 10^-18. Much lower than 1 %! 10^-18 is the kind of value which would result from following the formulation arguably implied in [MacAskill 2020](https://globalprioritiesinstitute.org/wp-content/uploads/William-MacAskill-Are_We_Living_at_the_Hinge_of_History-Jul-2022.pdf)[[2]](#fnr0yfpepx678), but it is not in agreement with the SSA, as explained in section 2 of [Mogensen 2022](https://globalprioritiesinstitute.org/wp-content/uploads/Mogensen-The-Hinge-of-History-Hypothesis-reply-to-MacAskill.pdf).
Acknowledgements
================
Thanks to David Denkenberger for feedback on the draft.
1. **[^](#fnref8o0yu4uw46c)** I [do not think](https://forum.effectivealtruism.org/posts/Y5eQHtEB29nW6FfQE/are-we-confident-that-superintelligent-artificial) AGI causing [human extinction](https://forum.effectivealtruism.org/topics/human-extinction) would necessarily lead to a permanent loss of value.
2. **[^](#fnrefr0yfpepx678)** William MacAskill provides numbers with respect to time. Consequently, since population size increases throughout time, all lives as a fraction of those this century is much smaller than all valuable time as a fraction of 100 years. So William gets a probability higher than 10^-18 for this being the most important century. Nevertheless, William agrees population is also relevant:
> However, for the purposes of true action-relevance, the influentialness of a time is not quite what we’re looking for. Even if we assume that now is the most influential time, because there are available opportunities to safeguard the longrun future, a rural farmer in Central African Republic would simply not be able to access those opportunities, and so for that person the question of how influential the present time is is neither here nor there. So we can generalize Trammell’s model slightly by talking about person-times rather than times: rather than being indexed to a particular time, each term in his model should be indexed to a particular person-time [i.e. life].
>
>
|
5aa9e556-f078-4653-bb15-8d96b05e6741
|
trentmkelly/LessWrong-43k
|
LessWrong
|
My least favorite thing
Epistemic status: Anger. Not edited.
TL;DR The hamster wheel is bad for you. Rationalists often see participation in the hamster wheel as instrumentally good. I don't think that is true.
Meet Alice. She has had the opportunity to learn many skills in her school years. Alice is a bright high school student with a mediocre GPA and a very high SAT score. She doesn't particularly enjoy school, and has no real interest in engaging in the notoriously soul-crushing college admissions treadmill.
Meet Bob. Bob understands that AGI is an imminent existential threat. Bob thinks AI alignment is not only urgent and pressing but also tractable. Bob is a second-year student at Ivy League U studying computer science.
Meet Charlie. Charlie is an L4 engineer at Google. He works on applied machine learning for the Maps team. He is very good at what he does.
Each of our characters has approached you for advice. Their terminal goals might be murky, but they all empathize deeply with the AI alignment problem. They'd like to do their part in decreasing X-risk.
You give Alice the following advice:
> It's statistically unlikely that you're the sort of genius who'd be highly productive without at least undergraduate training. At a better college, you will not only receive better training and have better peers; you will also have access to opportunities and signalling advantages that will make you much more useful.
>
> I understand your desire to change the world, and it's a wonderful thing. If you'd just endure the boredom of school for a few more years, you'll have much more impact.
>
> Right now, MIRI wouldn't even hire you. I mean, look at the credentials most AI researchers have!
>
> Statistically, you are not Eliezer.
You give Bob the following advice:
> Graduating is a very good signal. A IvyLeagueU degree carries a lot of signalling value! Have you gotten an internship yet? It's great that you are looking into alignment work, but it's also important that you take ca
|
13dc2e31-a512-4518-ad01-74fd0d37f08c
|
trentmkelly/LessWrong-43k
|
LessWrong
|
On Rigorous Error Handling
In my long career as a programmer I've seen a lot of projects with a lot of error handling practices and, sadly, almost none that actually worked.
We are facing facing few high-level problems here:
1. For many applications error handling is not critical.
2. Error handling code is rarely executed. Consequently, it looks like it works until it doesn't.
3. Programmers want to implement new features. Writing error handling is just an annoyance that slows them down.
Who cares about error handling?
For many applications writing rigorous error handling code doesn't pay off.
As long as you have some semi-decent way to deal with errors transparently (e.g. exceptions) you are fine. If there's a problem, throw an exception. On the top level catch all the exceptions and open a dialog box with the error message (in frontend applications) or write it to the log (on the server).
The above, of course, means that you hope that, after the exception is processed, the application will we left in a consistent, functional state. Once again, you HOPE. You do not KNOW. You do not know because, when writing the application, you haven't thought about error handling at all.
Now, don't get me wrong. I am not saying that you should never do this. It often makes sense from economic perspective. Writing rigorous error handling is expensive. If the worst thing that can happen is that a frontend application crashes once in a long while making user curse and restart it, then it's totally not worth doing.
However, there is also a different kind of application. There are medical application where the patient dies if they misbehave. There are space probes that crash into plantes, taking millions of dollars of investment as well as scientific careers down with them. There are HFT application which can generate bazillion of dollars of loss for every minute of downtime.
In these cases it's definitely worth writing rigorous error handling code.
That much everybody agrees with.
However, it is
|
8f166dca-a447-4842-bb45-b48a617c35ee
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Open thread, Apr. 17 - Apr. 23, 2017
This is the (late) weekly open thread. See the tag. You'd think we could automate this. The traditional boilerplate follows.
----------------------------------------
If it's worth saying, but not worth its own post, then it goes here.
----------------------------------------
Notes for future OT posters:
1. Please add the 'open_thread' tag.
2. Check if there is an active Open Thread before posting a new one. (Immediately before; refresh the list-of-threads page before posting.)
3. Open Threads should start on Monday, and end on Sunday.
4. Unflag the two options "Notify me of new top level comments on this article" and "Make this post available under..." before submitting.
|
c3c3e57f-fadb-4f07-9bba-e340c28cea47
|
StampyAI/alignment-research-dataset/lesswrong
|
LessWrong
|
Paradigm-building: Conclusion and practical takeaways
**Recapping the sequence**
--------------------------
Many thanks for reading this sequence all the way through! Recall this chart from [the introductory post](https://www.lesswrong.com/posts/4TuzWEKysvYdhRXLd/paradigm-building-introduction):
I’ve added some labels to clarify what this sequence was intended (and not intended) to cover. Namely, I take seriously [John Wentworth’s claim](https://www.alignmentforum.org/posts/3L46WGauGpr7nYubu/the-plan) that AGI safety research is still in a preparadigmatic phase; accordingly, my goal in this sequence was to sketch the space of plausible frameworks for doing technical safety work, starting from first principles, in an attempt to supply the field with a functional and coherent paradigm. Ultimately, I have presented this as a hierarchical series of questions:
1. What is the **predicted architecture** of the learning algorithm(s) used by AGI?
2. What are the **most likely bad outcomes** of this learning architecture?
3. What are the **control proposals** for minimizing these bad outcomes?
4. What are the **implementation proposals** for these control proposals?
5. What is the **predicted timeline** for the development of AGI?
Surely the boldest claim of this sequence is that these are the *right* questions to ask—that there is no future in which AGI safety research succeeds without having answered these questions. Though I do believe later questions rely in large part on answers to earlier questions, I do not think it is impossible to make any progress on the later questions—and therefore, that safety research must start paying attention to *all* of these questions as soon as possible.
I am certain that this framework is still a work in progress, no doubt containing mistakes and ambiguities that require further adjudication (if you think you have found some of either, please leave a comment on the relevant part of the sequence!). My hope is that this framework ultimately represents a logically coherent and methodologically useful synthesis of the wide array of perspectives currently on offer in safety work. As I [discussed earlier](https://www.lesswrong.com/posts/4TuzWEKysvYdhRXLd/paradigm-building-introduction), I also think that this sequence can be read as a low-resolution prediction about what the end-to-end progress of the field will end up looking like when all is said and done.
**Practical takeaways and next steps**
--------------------------------------
I also envision this framework serving as a practical communication tool that enables researchers to speak ‘across the valley’ of (sometimes radically) different priors. I imagine safety research would be far better understood and miscommunications more quickly diffused if researchers included something roughly like this at the very beginning of their technical work:
*My priors:*
* **Predicted architecture**: human-level weak online SL (with some RL).
* **Bad outcomes of interest**: inner misalignment, instrumental convergence.
* **Control proposals of interest**: imitation via amplification; installing human-like social cognition.
* **Implementation proposals of interest**: maximal facilitation; minimal incentivization.
* **AGI development timeline**: ~20 years (three iteratively-stable breakthroughs to go).
* **Evolution of my views:** I used to be more agnostic w.r.t. learning architecture, but now I am far more confident that AGI will exhibit SL.
(This is just an example—these are definitely not my priors.)
Succinct and systemized announcements of researchers’ priors headering their technical work would be proactively clarifying, both to their readers and to themselves. This would also enable a kind of (hopefully) comprehensive organizational system that would make researchers’ work far easier to constellate within the broader trends of the field.
In this same vein, another future direction for this work would be to periodically poll current AGI safety researchers along the lines of this sequence’s framework in order to get a better sense of the dominant technical research being conducted within the field. Surveys of this sort would help elucidate, both *within* and *across* the questions in this sequence’s framework, which areas are well-studied and which are less so.
For example, *within* the specific question of predicted learning architecture, we might find that only 15% of researchers are conducting online-learning-based safety work. *Across* questions, we may discover that far fewer people are working on implementation proposals as compared to control proposals. Data of this sort seems highly necessary for properly calibrating field-level dynamics.
As was discussed in [Paradigm-building from first principles](https://www.lesswrong.com/posts/2ieCSPoxgcrxaffv6/paradigm-building-from-first-principles-effective-altruism), I believe the ultimate goal of AGI safety research is something like mitigating AGI-induced existential risks. My hope is that the framework presented in this sequence will contribute to this outcome by enumerating the questions that necessarily must be answered before it becomes possible to achieve this goal.
As always, if you have any questions, comments, objections, etc., please don’t hesitate to [email me](mailto:cameron.berg@yale.edu)—or simply leave a comment on whichever post in the sequence is most relevant to your idea. Thanks very much for reading!
|
e8bde084-c968-4fa8-8e6a-a3a28bcc3676
|
trentmkelly/LessWrong-43k
|
LessWrong
|
The Desired Response
A friend once told me a story of how an interaction with her mother changed her perspective on communication. She said that she had been going through a break up at the time, and was venting to her mother, when her mother responded with "Do you want my advice, or my sympathy?"
Often times, when we have something important to talk about, we consider how we expect the other person to respond, and talk to people who will respond the kind of way we want. We may choose to speak to someone who we know can keep a secret, or who gives good advice, or who will say nothing at all, only listen. Sometimes we may turn to someone who won't want to talk about it at all, and will instead distract us from our problems. This can be great if you have a number of close friends who reliably respond in different ways, and you are able to predict this and make use of it. However, not everyone is in this situation, and so instead, we end up with interactions where people are not getting their desired responses.
Take an interaction between example-humans Alice and Bob. Alice's goal in the conversation may be to be heard, whereas Bob may want to feel validated for what he says. If Alice is talking, Bob may point out something or make a clever remark. Instead of validating Bob, Alice feels like she hasn't been heard, and is quiet in response. Now Bob has not been validated either, and they are both sad.
Sometimes, a person's desired response can be inferred. This is like how generally, if someone tells a joke, their desired response is for you to laugh, or at least acknowledge that the joke was funny. However, this isn't always easy to tell. Some of us may know the best thing for our best friend in a time of stress, but we don't always know what's right for someone else, even if we are close. Often times we will assume that what works for us will work for them, but then there is a risk of emotional damage. Trying to talk about a situation that someone wants to avoid thinking about may e
|
9377f19c-a3c6-45aa-acd5-90c88bb89739
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Meetup : First Medellin, Colombia Meetup
Discussion article for the meetup : First Medellin, Colombia Meetup
WHEN: 06 May 2017 06:00:00PM (-0500)
WHERE: Starbucks Coffee One Plaza, Medellin
In the spirit of http://lesswrong.com/lw/bc2/setting_up_lw_meetups_in_unlikely_places_positive/, I'm holding a meetup in Medellin, Colombia. I've been living here for 2 years now, and want to see if there are any other LW readers.
It will be held at the Starbucks located in the "One Plaza" office building at 6pm on Saturday, May 6th. I will have a copy of 'Quantum Computing Since Democritus' at my table for easy identification, and commit to being at Starbucks until 7pm even if nobody shows up.
Discussion article for the meetup : First Medellin, Colombia Meetup
|
1d3348d5-fd07-4140-9854-eaa2ac8da6e7
|
trentmkelly/LessWrong-43k
|
LessWrong
|
On Dualities
I’ve been finding the word duality useful quite a bit recently. I find it very useful for describing situations where there are two very valuable perspectives (or lens) through which we can look at a situation and any attempt to answer the question needs to grapple with and account for both of these. The way I use the word, I’m not claiming that two logically contradictory viewpoints are simultaneously true, but rather that aspects of both can be synthesised together to reach the truth.
Here’s a few examples. Maybe you agree or disagree with these specific examples, but I think they should suffice for illustrative purposes:
* Some people are born with major disadvantages and we need to be sympathetic to them. At the same time, people can act in a way which makes their decisions better or worse and we need to encourage personal responsibility. If we’re too harsh, we don’t give them the help that they need, if we’re too sympathetic, we simply enable people to ruin their own lives. We need to find a balance between the two
* On one hand, tolerance is important for enabling us to get along with people who are different from us. On the other hand, too much tolerance means that there are no standards of behaviour. We need find a way to merge these two considerations without simply tolerating that which we approve and refusing to tolerate that which we dislike (it is possible to synthesise a worst of both worlds approach)
* On one hand, a person can be acting in a way that is horrific from a moral standpoint, on the other hand, it may be hard to blame them given the circumstances. This sets up a tough conflict between the demands of justice and the desire to be merciful and it can be hard to figure out how to navigate these systems
Dualities allow you to avoid a particular dysfunctional pattern of thought. Most people will have one side of the duality stick out for them more than the other. Since both sides contradict (or seem to contradict), they conclude that they
|
a65e0b38-5797-4412-ab49-de37be396dc4
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Holden Karnofsky's Singularity Institute critique: Is SI the kind of organization we want to bet on?
The sheer length of GiveWell co-founder and co-executive director Holden Karnofsky's excellent critique of the Singularity Institute means that it's hard to keep track of the resulting discussion. I propose to break out each of his objections into a separate Discussion post so that each receives the attention it deserves.
Is SI the kind of organization we want to bet on?
This part of the post has some risks. For most of GiveWell's history, sticking to our standard criteria - and putting more energy into recommended than non-recommended organizations - has enabled us to share our honest thoughts about charities without appearing to get personal. But when evaluating a group such as SI, I can't avoid placing a heavy weight on (my read on) the general competence, capability and "intangibles" of the people and organization, because SI's mission is not about repeating activities that have worked in the past. Sharing my views on these issues could strike some as personal or mean-spirited and could lead to the misimpression that GiveWell is hostile toward SI. But it is simply necessary in order to be fully transparent about why I hold the views that I hold.
Fortunately, SI is an ideal organization for our first discussion of this type. I believe the staff and supporters of SI would overwhelmingly rather hear the whole truth about my thoughts - so that they can directly engage them and, if warranted, make changes - than have me sugar-coat what I think in order to spare their feelings. People who know me and my attitude toward being honest vs. sparing feelings know that this, itself, is high praise for SI.
One more comment before I continue: our policy is that non-public information provided to us by a charity will not be published or discussed without that charity's prior consent. However, none of the content of this post is based on private information; all of it is based on information that SI has made available to the public.
There are several reasons that I current
|
54a0e068-cc13-45b5-9803-39cfb9431c1a
|
trentmkelly/LessWrong-43k
|
LessWrong
|
What are Responsible Scaling Policies (RSPs)?
This is an article in the featured articles series from AISafety.info. AISafety.info writes AI safety intro content. We'd appreciate any feedback.
The most up-to-date version of this article is on our website, along with 300+ other articles on AI existential safety.
METR[1] defines a responsible scaling policy (RSP) as a specification of “what level of AI capabilities an AI developer is prepared to handle safely with their current protective measures, and conditions under which it would be too dangerous to continue deploying AI systems and/or scaling up AI capabilities until protective measures improve.”
Anthropic was the first company to publish an RSP in September 2023 defining 4 AI Safety Levels.
> “A very abbreviated summary of the ASL system is as follows:
>
> * ASL-1 refers to systems which pose no meaningful catastrophic risk, for example a 2018 LLM or an AI system that only plays chess.
> * ASL-2 refers to systems that show early signs of dangerous capabilities – for example ability to give instructions on how to build bioweapons – but where the information is not yet useful due to insufficient reliability or not providing information that e.g. a search engine couldn’t. Current LLMs, including Claude, appear to be ASL-2.
> * ASL-3 refers to systems that substantially increase the risk of catastrophic misuse compared to non-AI baselines (e.g. search engines or textbooks) OR that show low-level autonomous capabilities.
> * ASL-4 and higher (ASL-5+) is not yet defined as it is too far from present systems, but will likely involve qualitative escalations in catastrophic misuse potential and autonomy.”
Other AI companies[2] have released their own versions of such documents with various names:
* OpenAI's 2023 beta version of their Preparedness Framework
* Deepmind's 2024 Frontier Safety Framework
* Microsoft’s 2025 Frontier Governance Framework
* Meta’s 2025 Frontier AI Framework
* Amazon’s 2025 Frontier Model Safety Framework
RSPs have recei
|
bb23bdc1-85cb-403f-a4a0-2465c8087ab7
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Tsinghua paper: Does RL Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?
arXiv | project page | Authors: Yang Yue, Zhiqi Chen, Rui Lu, Andrew Zhao, Zhaokai Wang, Yang Yue, Shiji Song, Gao Huang
This paper from Tsinghua find that RL on verifiable rewards (RLVR) just increases the frequency at which capabilities are sampled, rather than giving a base model new capabilities. To do this, they compare pass@k scores between a base model and an RLed model. Recall that pass@k is the percentage of questions a model can solve at least once given k attempts at each question.
Main result: On a math benchmark, an RLed model (yellow) has much better raw score / pass@1 than the base model (black), but lower pass@256! The authors say that RL prunes away reasoning pathways from the base model, but sometimes reasoning pathways that are rarely sampled end up being useful for solving the problem. So RL “narrows the reasoning boundary”— the region of problems the model is capable of solving sometimes.
Further results
* Across multiple math benchmarks, base models have higher pass@k than RLed models for sufficiently large k, and sometimes the crossover point is as low as k=4. (Figure 2)
* Out of domain (i.e. on a different math benchmark) the RLed model does better at pass@256, especially when using algorithms like RLOO and Reinforce++. If there is a crossover point it might be near pass@1024. (Figure 7)
* To see if the pass@1024 results are just lucky guesses, they sanity check reasoning traces and find that for most questions, base models generate at least one correct reasoning trace. (But it's unclear whether the majority of correct answers are lucky guesses) Also, for long coding tasks, it's nearly impossible to make a lucky guess.
* They test the perplexity (Below: Figure 6 left) of the RLed model’s generations (pink bar) relative to the base model. They find it is lower than the base model's perplexity (turquoise bar), which “suggests that the responses from RL-trained models are highly likely to be generated by the base model” conditioned on th
|
b11546aa-12e5-4b7e-b045-ae19231c4cee
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Probability puzzle
I came up with this puzzle after reading Vaniver's excellent post on the Value of Information. I enjoyed working it out over Thanksgiving and thought I'd share it with the rest of you.
Your friend holds up a curiously warped coin. "Let's play a game," he says. "I've tampered with this quarter. It could come up all heads, all tails, or any value in between. I want you to predict a coin flip; if you get it right, I'll pay you $1, and if you're wrong, you pay me $3."
"Absolutely, on one condition," you reply. "We repeat this bet until I decide to stop or we finish N games."
What is the minimum value of N that lets you come out ahead on average?
Each game, you may choose heads or tails, or to end the sequence of bets with that coin. Assume that all heads:tails ratios are equally likely for the coin.
edit: since a couple people have gotten it, I'll link my solution: http://pastebin.com/XsEizNFL
|
97aa5c79-4e06-411e-8542-e3b5d82d715d
|
StampyAI/alignment-research-dataset/arxiv
|
Arxiv
|
Engines of Power: Electricity, AI, and General-Purpose Military Transformations
1 Engines of Power: Electricity, AI, and General-Purpose Military Transformations Jeffrey Ding and Allan Dafoe* Abstract Major theories of military innovation focus on relatively narrow technological developments, such as nuclear weapons or aircraft carriers. Arguably the most profound military implications of technological change, however, come from more fundamental advances arising from “general purpose technologies” (GPTs), such as the steam engine, electricity, and the computer. With few exceptions, political scientists have not theorized about GPTs. Drawing from the economics literature on GPTs, we distill several propositions on how and when GPTs affect military affairs. We call these effects “general-purpose military transformations” (GMTs). In particular, we argue that the impacts of GMTs on military effectiveness are broad, delayed, and shaped by indirect productivity spillovers. Additionally, GMTs differentially advantage those militaries that can draw from a robust industrial base in the GPT. To illustrate the explanatory value of our theory, we conduct a case study of the military consequences of electricity, the prototypical GPT. Finally, we apply our findings to artificial intelligence, which will plausibly cause a profound general-purpose military transformation. * Jeffrey Ding is a PhD candidate in the Department of Politics and International Relations, University of Oxford, and a researcher at the Center for the Governance of Artificial Intelligence (GovAI). He is currently a predoctoral fellow at the Center for International Security and Cooperation at Stanford University, sponsored by Stanford's Institute for Human- Centered Artificial Intelligence. Allan Dafoe is an Associate Professor at the University of Oxford and the Director of GovAI, at the Future of Humanity Institute. Correspondence email: jding99@gmail.com. For helpful comments and input, we thank: Ashwin Acharya, Markus Anderljung, Carolyn Ashurst, Ben Buchanan, Max Daniel, Ben Garfinkel, Charlie Giattino, Hamish Hobbs, Alex Lintz, Luke Muehlhauser, Carina Prunkl, Toby Shevlane, Waqar Zaidi, and especially Alexis Carlier and Joslyn Trager.
2 I. Introduction AI is the new electricity. — Andrew Ng, Leading AI researcher and Co-Founder of Google Brain1 It is now standard for social scientists, policymakers, and machine learning experts to compare artificial intelligence (AI) with electricity, the quintessential general-purpose technology (GPT). International relations scholars that study the military implications of AI acknowledge this comparison, yet they have done little systematic research into the military implications of GPTs. Much work treats AI as a relatively narrow technological advance, in the mold of nuclear weapons or aircraft carriers.2 It does not reckon with AI as a GPT, such as electricity or the computer, which is differentiated by its pervasiveness, scope for continual improvement, and strong synergies with other technologies. The comparison is gestured at but not seriously examined. How do GPTs, like electricity and AI, influence the military balance of power? Taking military effectiveness as an analogue for economic productivity,3 we extend insights on GPTs and economic transformations to the implications of GPTs for military transformations. Differing from narrower technologies, GPTs influence military effectiveness through a protracted, gradual process that involves a broad range of military innovations and overall industrial productivity growth. Referring to this process as a general-purpose military transformation (GMT), we identify three features of GMTs which relate to their breadth of impact, timeline of widespread adoption, and indirect productivity spillovers. We then argue that GMTs differentially advantage militaries that are connected to a robust industrial base in the associated GPT, which we label “the industrial dependency hypothesis.” If GPTs were like other narrower technologies, differentials in military adoption may be more tied to other factors, such as the fit between a single military innovation and a military’s tactical incentives, financial resources, and organizational 1 A Google search for the exact phrase “AI is the new electricity,” conducted on January 15, 2020, returned about 37,000 hits. Ng first popularized this comparison in a 2017 speech at Stanford. In 2014, Kevin Kelly, the former editor of Wired, wrote, “Everything that we formerly electrified we will now cognitize…business plans of the next 10,000 startups are easy to forecast: Take X and add AI.” Kelly 2014. 2 The burgeoning narrative of an AI arms race exemplifies this point. Zwetsloot et al. 2018. 3 Military effectiveness is “the process by which armed forces convert resources into fighting power.” Millett et al. 1986, 37. Inherent to military effectiveness is a notion of efficiency rather than just total combat power.
3 capital.4 These can affect the adoption of particular military innovations linked to a GMT, but a military’s ability to draw on a robust industrial base in the GPT affects all military innovations linked to a GMT. To effectively exploit a GMT, militaries must draw on talent, industry, and infrastructure in the civilian realm, where the momentum for a GPT’s development lies. To empirically support our reasoning about GMTs, we examine the evolution of electricity in military affairs. Surprisingly, very little scholarship directly examines the military consequences of electricity — widely recognized as one of the most significant technological innovations in history.5 In the 1870s and 1880s, a cluster of electrical innovations, including the electric dynamo (1871) and the steam turbine generator (1884), helped create a versatile energy system with many industrial applications in lighting, communications, transportation, and machinery. A GPT was born. Playing a notable role in World War I, military applications of electricity increased in breadth and depth throughout the interwar period. After World War II, informed observers ranked electricity among the three great influences on naval warfare in the 20th century.6 Eliminating the “naval” qualifier would not be a stretch. The process of military electrification exhibits the three theorized features of a GMT. First, the impact of electricity on military power materialized through a broad range of military applications, including communications (e.g. wireless telegraphy for battlefleet coordination), fortifications (e.g. searchlights to defend against night attacks), transportation (e.g. diesel-electric propulsion for submarines), and weapon systems (e.g. electric fire control). Second, electricity significantly upgraded industrial productivity, which increased military production potential. Third, similar to the slow progression of electrification across economic sectors, the spread of electrical innovations across military branches and divisions took many decades. In line with the industrial dependency hypothesis, the extent to which militaries took full advantage of electrification depended on their connection to a robust base of electrical talent, industry, and 4 Gilli and Gilli 2014; Horowitz 2010 5 Exceptions include: Hezlet 1975, Headrick 1991. 6 Sir Henry Tizard, Chairman of the UK’s Defence Research Policy Committee, declared that the three great influences on naval warfare of the 20th century were the aircraft, the submarine, and the discovery of electromagnetic waves. Hezlet 1975.
4 infrastructure in the civilian economy. As established by the three features highlighted above, incorporating a GMT requires much more than acquiring a single electric dynamo or adopting one military innovation. The Russian military, for instance, pioneered the use of electronic countermeasures in combat in 1904. However, unlike leaders in military electrification like Britain, Russia’s weak industrial base of electrical technology prevented it from taking full advantage of the electrification.7 To separate the effects of the GPT dimension from other factors that could explain this military electrification gap, we assess Russia’s ability to keep pace with Britain in adopting submarines, a non-GPT. This article directly engages with key academic and policy debates. First, the existing literature on military innovation overwhelmingly draws on case studies of relatively narrow technological developments, such as new weapon systems.8 Yet, historians generally recognize that the most profound impacts of technological change arise from more fundamental breakthroughs, including GPTs like the steam engine or the computer. In addressing this gap, our study of electricity broadens the universe of cases for investigating the military implications of technological change. Second, we develop the first comprehensive theory that sets forth the conditions under which GPTs alter the military balance of power. The limited literature on GPTs and military power posits that because GPTs are characterized by private sector dominance and relatively low fixed costs, they rapidly diffuse from technological leaders to laggards, thereby narrowing the gap in military capabilities.9 We argue, instead, that technological laggards do not inevitably catch up in GMTs. Rather, the protracted process of GMT diffusion differentially advantages militaries which can tap into a robust industrial base in the GPT. By highlighting the unique effects of GPTs, we contribute to a growing approach to studying the impacts of emerging
7 Russia’s per capita production of electricity did not reach UK 1910 levels until 1930. Our calculations are based on the Cross-country Historical Adoption of Technology (CHAT) dataset by Comin and Hobjin 2009. 8 For instance, a systematic review of 60 different cases of military innovation, sourced from 73 books and articles on the subject, does not mention electricity (Horowitz and Pindyck 2020). 9 Drezner 2019, 300; Horowitz 2018, 39.
5 technologies on international security, which focuses on particular technological dimensions such as complexity,10 disruptiveness,11 and dual-use.12 Finally, our research directly bears on current debates over how AI could shift the balance of military power.13 To date, much of the discussion emphasizes the narrow effects of specific AI applications, such as autonomous weapons.14 That AI could influence military power through its effects on industrial productivity is rarely discussed among scholars. Some scholars also suggest that AI will significantly transform military effectiveness in a relatively short timeframe, and that AI will enable rising powers to leapfrog the U.S. in military strength. As the conclusion will show, a GMT-based approach points toward different conclusions on all these fronts. The rest of the article proceeds as follows. The next section deduces key propositions about the impact of GPTs on military affairs by adapting insights from economic and historical studies of GPTs. Our theory of GMTs is composed of three features influencing how and when GPTs affect military affairs, along with an explanation for why GMTs differentially advantage certain militaries. Leveraging primary and secondary accounts, we then illustrate the explanatory value of our theory with a historical case study: the evolution of electricity in military affairs. Finally, we discuss the limitations of our analysis and reflect on the military implications of artificial intelligence, which is plausibly a GPT as impactful as electricity.
10 Gilli and Gilli 2019 11 Mukunda 2010; Dombrowski and Gohlz 2009. 12 Stowsky 2004. Relatedly, discussions of “Revolutions in Military Affairs” (Krepinevich 1994) and long-wave theories. (Modelski and Thompson 1996) also contend that basic innovations in the commercial domain can drive military-technical revolutions. These accounts relate to the broad outlines of our theory, but they do not specify how and when these transformations occur. We specify GPTs as a particular class of technologies that can generate these revolutions. 13 Garfinkel and Dafoe 2019; Horowitz 2018; Kania 2017; Payne 2018 14 Some texts analyze autonomous weapons because they present thorny legal and ethical challenges. Scharre 2018; Bode and Huelss 2018
6 II. Theory: GPTs and Military Power FOUNDATIONS OF GPT THEORY The vast majority of theorizing about GPTs comes from economists and economic historians who analyze GPTs as the primary drivers of long-term economic growth. One systematic review of GPTs described the concept as “one of the most successful memes in economic history in the last decade.”15 With a few exceptions, political scientists have done much less theorizing about GPTs.16 Per a search of the JSTOR database, the economics discipline accounted for 262 of 365 total articles that referenced "general purpose technology." The combined equivalent for both the political science and international relations field was 33 articles.17 Economists and economic historians largely agree that a GPT is defined by three characteristics.18 First, GPTs offer great potential for continual improvement. While all technologies offer some scope for improvement, a GPT “has implicit in it a major research program for improvements, adaptations, and modifications.”19 Second, GPTs are characterized by their pervasiveness. As a GPT evolves, it finds a wide “variety of uses” and “range of uses.”20 A wide “variety of uses” refers to the diversity of a GPT’s use cases (e.g. a computer can be used for storing information, calculating statistics, entertainment services, etc.), while a wide “range of uses” captures the range of industries and individuals that use a GPT. For instance, a screw has a wide “range of use” since it is used to fasten things together across a large swathe of productivity activities in the economy, but it does not have a wide “variety of uses.”21 15 Field 2008, 2. 16 Drezner 2019; Horowitz 2018. Some literature on “leading sectors” also references GPTs. See, for example, Thompson 1990. 17 JSTOR search conducted April 28, 2020. 18 The following discussion is mostly drawn from Lipsey et al. 2005 and Bresnahan and Trajtenberg 1995. Other accounts employ similar definitions, albeit with some modification: Jovanovic & Rousseau 2005; Bresnahan 2010. For a critical view of the GPT concept, see: Field 2008; 19 Lipsey et al. 1998, 39 20 We agree with Cantner and Vannuccini’s (2012) emphasis on the establishment of GPTs as a process unfolding in time. Thus, a GPT can be seen as an emergent property constituted by interactions between technological characteristics and institutions. 21 Lipsey et al. 1998, 39.
7 Third, GPTs have strong technological complementarities. In other words, the benefits from innovations in GPTs come from how related technologies change in response, and cannot be modeled as a mere reduction in the costs of inputs to the existing production function. For example, the overall energy efficiency gains from merely replacing a steam engine with an electric motor were minimal. Facilitated by technological improvements in electric storage and machine tools, the main productivity benefits of factory electrification only materialized after plants transitioned their entire power distribution system to one in which machines were driven individually by electric motors.22 Tying all three characteristics together, David describes the pattern of how a GPT spreads as an “extended trajectory of incremental technical improvements, the gradual and protracted process of diffusion into widespread use, and the confluence with other streams of technological innovation, all of which are interdependent features of the dynamic process through which a general purpose engine acquires a broad domain of specific applications.”23 Though these characteristics of a GPT apply in all contexts, the rate at which a GPT acquires a broad domain of applications varies by country. For instance, empirical analysis has shown that the U.S. has benefited more from computer technologies in terms of economic productivity than its industrial rivals.24 One of the key factors that explains this gap is the superior performance of the U.S. higher education system in adjusting to the increased ICT skill needs. In particular, cultivating engineering professions provides an important repository in which GPT learning accumulates and then spreads through different industries. Historically, engineering disciplines, such as computer science and electrical engineering, have developed in the wake of a new GPT.25 The demands for this type of skill systematization and standardization are especially salient for the extensive adoption of GPTs, which “depend on widening, as opposed to deepening, the knowledge base.”26 22 Previously, factories were powered by shaft and belt drive systems, which relied on a single, central steam engine. Devine 1982. 23 David 1990, 356. Emphasis ours. 24 See, for example, Krueger and Kumar 2004. 25 Rosenberg 1998a, 169. 26 Vona and Consoli 2014, 1397. The key is to develop standardized routines that capture the practical and localized knowledge embedded in rapidly evolving technological trajectories, and then widen access to less-talented individuals (Vona and Consoli 2014, 1402-1403)
8 TRANSLATING GPT THEORY TO MILITARY TRANSFORMATION By adapting insights from the literature on GPTs and economic productivity, we theorize about how GPTs transform military effectiveness (a GMT). Taking military divisions and branches (as opposed to industries or firms) as the application sectors of a GPT, we translate the pattern of how a GPT spreads across a national economy to a military context.27 We first deductively articulate three key features of GMTs. Equipped with a more complete view of GMTs, we then pinpoint why different militaries are better able to exploit GMTs. For our translation to work, we must first establish that the same economically transformative characteristics of GPTs also apply to military transformation. GPTs possess great potential for continual improvement, become pervasive in their wide variety and range of military applications, and have strong technological complementarities with existing military technology systems. The computer is one such example. Upon entering military systems, it continually improved along many technical dimensions, provoked significant structural changes, and eventually found a wide variety of uses across many branches and units (in battle simulations, logistics handling, and weapons control, etc)..28 We can differentiate GPTs from other militarily-significant innovations that meet some but not all three of the GPT criteria. The category of dual-use technologies, defined as technology which has both commercial and military applications, is an important near neighbor.29 Some dual-use technologies, such as aircraft, are characterized by continual improvement and offer strong technological complementaries (e.g. benefits to militaries that adopted strategic bombing in World War II). Aircraft propulsion systems, however, have a limited variety and range of applications which means we should not expect them to give rise to 27 We draw on some models of military innovation that disaggregate the military into individual units (e.g. branches, divisions, etc.) that can adopt technological capabilities at different rates. Grissom 2006. 28 O’Hanlan (2018, 20) writes, “Computer systems have been introduced...into virtually all domains of military equipment and operations.” Computers have propelled drones and robotics to the frontier of warfare. The proliferation of inexpensive computing power has improved capabilities in automated aircraft, missile defense interceptors, and guided mortars (O’Hanlan 2018, 11). 29 All GPTs are dual-use, but not all dual-use technologies are GPTs. The diversity of potential applications for many dual-use technologies is limited. Horowitz 2020a.
9 GMTs.30 The breadth of the transformations produced by GPTs often lead analysts to describe it as an “ization,” such as the electrification, computerization, or zhinenghua (intelligentization, in Chinese) of the military. The computerization of the military describes the process of the computer’s diffusion in the military domain. There is no equivalent for aircraft engines. Of course, GMTs differ from GPT-driven economic transformations. Unlike the market dynamics of a nation's economy, intra-military competition involves different units pursuing a nominal shared mission and budget flows from one big primary source. Military effectiveness and economic productivity are analogous but not the same; measuring the output of technological innovation in the military realm is much more difficult than in the civilian economy.31 Despite these differences, some of the ways in which GPTs interact with military organizations follow a similar trajectory as GPTs in economic systems. In particular, the first two features of GMTs, regarding their broad impact pathway and prolonged timeline of diffusion, draw directly from stylized facts in the existing GPT literature. Like translations of all kinds, adapting the foundations of GPT theory to military affairs requires some modifications. The military is typically reliant on the civilian economy to advance the GPT’s development, particularly after its initial incubation.32 The breadth of possible GPT applications in the economic realm far exceeds those in the military realm, so the full effect of a GMT includes the impact of GPT-induced productivity surges on military production capabilities. Moreover, the trajectory of a GMT will depend on a military’s connection to the evolution of a GPT in the civilian economy. The rest of this section describes the three key characteristics of GMTs and the GMT leveling hypothesis in more detail. 30 The general-purpose nature of a technology, like the level of disruptiveness or complexity associated with a technology, falls on a spectrum where the delineation between a GPT and a non-GPT is somewhat fuzzy. This discussion of aircraft, classified by some scholars as a GPT (e.g. Ruttan 2006), demonstrates that one can usefully exclude “near-GPTs” from consideration based on established criteria of a GPT. In a systematic survey of lists of historical GPTs, none of the texts included aircraft as a GPT. (Field 2008) 31 Rosen 1994, 46. 32 Some scholars have shown that military investment played a crucial role in incubating some GPTs. See Misa 1985; Ruttan 2006; Smith 1985. Misa’s work tracks how the U.S. Army Signal Corps served as an important institutional entrepreneur in the development phase of the transistor. Military investment, however, is not necessary for seeding GPTs, as the history of commercially-initiated developments in steam engines and electricity show (Bresnahan and Trajtenberg 1995, 95-96).
10 GMTS: THREE FEATURES OF GPT TRAJECTORIES IN MILITARY AFFAIRS First, GPTs directly enhance military power by spurring a wide range of military innovations. When scholars analyze the impact of technology on warfare, their attention often goes to the most visible and graphic part of warfare: the projectile or other mechanism of force.33 The impact of a weapon system is thought to materialize through a relatively narrow impact pathway. In contrast to weapons technology, which do not have many other downstream applications, GPTs influence military effectiveness through a very broad pathway. Since GPTs are utilized in a variety of ways across the entire military, they produce many downstream applications. Some may appear insignificant (the use of a computer to calculate figures for a military unit’s budget), while others seem more revolutionary (the use of a computer to crack the encryption of enemy communications). The impact of a GPT on military power, therefore, depends on the distribution over all these downstream applications. Given the broad impact pathway of a GPT, in the early stages of a GPT’s development the foreseeability of a GPT’s effect on the conduct of warfare is very limited. GPTs will affect military affairs in unanticipated ways. Because the advance of a GPT depends on many complementary innovations and adaptations, its full impact will not be apparent until one fully accounts for the many, lengthy causal chains shaping various sectors, and the ways those sectors adapt. For example, when the steam engine was invented in the eighteenth century, it was initially used exclusively to pump water out of flooded mines. It was not until the early nineteenth century that the steam engine became a generalizable source of power for factories, railroads, and naval ships.34 A certain degree of unpredictability applies to the effects of all innovations on military affairs. But the breadth of military applications affected by an innovation can vary. Some innovations present a very limited set of applications, though these applications can interact with the strategic landscape in many ways 33 Beckley 2010, 55. 34 Rosenberg 1996, 345. The steam engine is widely recognized as a GPT. (Field 2008); Predicted demand for computers was also very off-base. A survey by the Dept. of Commerce conducted in the late 1940s estimated that about 100 mainframe computers would satisfy the entire needs of the nation. Helpman and Trajtenberg 1996, 30 (fn 16)
11 (Figure 1). For instance, nuclear fission technology had a relatively bounded set of military applications — namely, nuclear weapons — but the interaction between nuclear weapons and the strategic landscape evolved in multifaceted, unpredictable ways.35 Our argument is that the set of possible military applications for GPT innovations is much larger than the corresponding set for other technologies, thereby severely limiting the foreseeability of its military implications. Figure 1: The Military Technology Stack — Two Impact Pathways
Second, GPTs indirectly affect military power by significantly upgrading industrial production potential. In addition to the direct impact pathway, it is necessary to consider the ramifications of GPTs for a country’s production potential. Many scholars have noted that the productive capacity of a nation’s industrial base undergirds military power.36 Across hundreds of battles and wars from 1898 to 1987, Beckley has found a positive relationship between economic productivity and military effectiveness. Separate from its total 35 Morgenthau 1964 36 Kennedy 1987, Gilpin 1975; Kirshner 1998. Some previous scholarship on RMAs does discuss the indirect effects of technological change on “production revolutions” — e.g. the effects of the American System of Manufactures on the World War I RMA. But this is not a regular feature of RMAs. Most RMAs, including the prototypical example of the nuclear revolution, do not involve a substantial upgrading of the industrial base. This marks another clear distinction between GMTs and RMAs.
12 economic output, a more economically efficient nation will translate its economic resources into superior weapons and military organization.37 Again, this feature of GMTs derives from the fact that GPTs differ from other technologies. Assessed on their own merits alone, even the most transformative (non-GPT) technological changes do not tip the scale far enough to significantly affect overall industrial productivity.38 As “engines of growth,” GPTs are different because their impact on productivity comes from accumulated improvements across a wide range of complementary sectors; in other words, they cannot be judged on their own merits alone.39 According to empirical studies of the steam engine, electricity, and information and communications technologies (ICTs), the episodic arrival of GPTs precedes a wave of economy-wide productivity growth.40 Third, the most consequential impacts of a GPT on military effectiveness occur only after a long period of gestation. Many decades pass between the emergence of a GPT and its substantial effect on productivity. In the 1980s economists who had pointed to GPTs to explain changes in long-term economic growth were confronted with a paradox. Transformative GPTs like computers were claimed to greatly increase productivity, yet “we (saw) the computers everywhere but in the productivity statistics.”41 Evidence of a long diffusion lag — “a time to sow” and “a time to reap” in the formulation of Helpman and Trajtenberg — between the emergence of a GPT and its full effects on productivity helped address this paradox.42 Compared to other technologies, GPTs exhibit a much more pronounced diffusion lag due to their substantial demands for complementary innovations, organizational changes, and upgrading of technical skills.43 Within the first decade of the introduction of the transistor, for instance, computers and hearing aids had already incorporated semiconductors. But the telecommunications sector took much longer to adopt this GPT, despite the fact that the development of semiconductors was largely inspired by the potential benefits 37 Beckley 2010; See also Beckley 2018. 38 Fogel’s classic study of the social savings linked to railroad construction in the U.S. led him to conclude that “the railroad did not make an overwhelming contribution to the production potential of the economy.” Fogel 1964, 235. 39 Bresnahan and Trajtenberg 1995. 40 Ruttan 2006, 5; David 1990; Brynjolfsson et al. 2017 41 Robert Solow’s famous quip in 1987 42 Helpman and Trajtenberg 1994 43 Brynjolfsson et al. 2017; David 1990
13 for the telecommunications sector.44 Though the first dynamo for industrial application was introduced in the 1870s, the full impact of electricity on overall manufacturing productivity did not occur until the 1920s. This was only realized after organizational adaptations, such as changes in factory layout, and complementary innovations, such as the steam turbine, which enabled central power generation in the form of utilities.45 The nature of GPTs raises difficult questions about when their most consequential military effects materialize. For example, one could trace present-day advances in military electronics and precision warfare back to 19th-century electrical innovations. When does a GMT end? For our purposes, we date this to when a GPT spreads across a wide range of military applications.46 Oftentimes, another GPT, such as the emergence of the transistor in the 1940s, opens up a new round of general purpose military transformation. Thus, advances in military electronics and precision warfare, though built on a base of electrical advances, are more connected with the transistor GMT than the electrical one. GMTS AND DIFFERENTIAL ADVANTAGES We hypothesize that militaries more connected to a strong industrial base in the GPT are better positioned to exploit GMTs. The broad applicability of GPTs across many sectors, combined with the fact that the civilian economy presents many more application scenarios than the military realm, means that the momentum for a GPT’s development lies in the civilian realm.47 This distinguishes GPTs from (non-GPT) dual-use technologies, such as nuclear power.48 To bring about a GMT, therefore, a military will need to rely more on the civilian sector for flows of knowledge, talent, investment, and complementary innovations. This leads to the following industrial dependency hypothesis: A GPT differentially advantages militaries that can draw from a robust industrial base in the GPT. 44 Recall that the primary task of Bell Labs, the birthplace of semiconductors, was to develop the telecommunications systems manufactured by AT&T. Bresnahan and Trajtenberg 1995. 45 Smil 2005, 33-97. David 1990. 46 In the electricity case, this occurs during World War II. 47 Earlier in this section, we acknowledge that in some cases the military has played an important role in initiating developments in GPTs. 48 Maintaining a strong nuclear weapons capability does not depend on the entire industrial base’s strength in civilian nuclear applications.
14 This proposition differs from arguments that GPTs level the military balance of power. Drezner compares GPTs, which he characterizes by private sector dominance and low fixed costs, to prestige tech like space exploration programs, defined by public sector dominance and high fixed costs. Arguing that GPTs rapidly diffuse from technological leaders to laggards, he posits that “general purpose tech has a greater leveling effect than prestige tech.”49 Horowitz also links commercially-driven GPTs to a leveling effect. Specifically, he describes this effect for AI: “If commercially-driven AI continues to fuel innovation, and the types of algorithms militaries might one day use are closely related to civilian applications, advances in AI are likely to diffuse more rapidly to militaries around the world...The potential for diffusion would make it more difficult to maintain ‘first-mover advantages’ in applications of narrow AI. This could change the balance of power, narrowing the gap in military capabilities not only between the United States and China but between others as well.”50 Evaluating these hypotheses will depend on how one conceptualizes the acquisition of a GPT. In our view, a military’s acquisition of a single electric dynamo should not count as successful diffusion, as this does not capture whether the military has meaningfully incorporated the GMT associated with electricity. That GPTs are driven forward by civilian applications does not necessitate that they will diffuse quickly to militaries that are technological laggards. Instead, the protracted, challenging process of a GMT differentially advantages militaries able to tap into a robust industrial base in the associated GPT. The emphasis on an industrial base with specific competencies in the GPT also distinguishes this hypothesis from existing work on the economic roots of military effectiveness. Scholars have shown that the level of economic development has a significant effect on military effectiveness, reasoning that states with more efficient economies can field skilled personnel that can absorb complicated technology effectively.51 We propose that, even among developed states, the connection to the industrial base’s development of a GPT has a significant effect on which militaries take better advantage of GMTs. 49 Drezner 2019, 300. 50 Horowitz 2018, 39. To be clear, this is only one such scenario Horowitz outlines. Countervailing factors in favor of non-leveling in his careful assessment include: the possibility that the most important specific military uses of AI are exclusively based in military research and the constraint of compute costs, which could price out all but the wealthiest countries from adopting higher-end AI capabilities. 51 Beckley 2010
15 III. Case Study: The Electrification of Warfare (late 19th and early 20th centuries) RESEARCH DESIGN AND CASE SELECTION STRATEGY To evaluate our theory, we conduct a historical case study of the impact of electricity on military affairs. There is some debate over which technologies qualify as GPTs, but electricity is unanimously viewed as a prototypical GPT.52 This makes it a typical, or representative, case for studying GMTs. The frequent contemporary comparisons made between AI and electricity serve as an additional advantage. Studying the electrification of warfare is also substantively important in its own right. Despite its wide recognition as one of the most significant technological changes in history, the electrification of the military has not received much scholarly attention.53 If the evidence from the electricity case supports our theory, we should observe two main sets of implications. First, the impact of electricity on military effectiveness should exhibit the three features of a GMT: broad impact pathway, indirect productivity benefits, and prolonged gestation period. Second, in line with the industrial dependency hypothesis, we should see differentials in military electrification based on militaries’ connections to a robust industrial base in the GPT (industrial dependency hypothesis). To control for competing explanations of how GPTs affect the military balance of power, we compare military electrification to the impact of the submarine on military affairs. These two technologies differ with respect to whether they qualify as a GPT but are similar across most other relevant features.54 The submarine was not a GPT. While it generated some technological complementarities with advances in weapon systems and underwater propulsion, submarine technology did not have a wide variety and range of uses and had limited potential for continual innovation. As Michael Horowitz writes, “[AI] is an enabler, a 52 Ristuccia and Solomou 2014, 227. Historians disagree about which technologies qualify as GPTs, but almost all lists include the steam engine, electricity, and information and communications technology (ICTs). Field 2008, 10. 53 Exceptions are Hezlet 1975 and Headrick 1991, which we draw from in our empirics. Headrick limits his analysis to electrical communications, and Hezlet only covers the effect of electricity on naval warfare. 54 For more on most-similar-system comparisons, see Beach and Pederson 2016, 239-240.
16 general-purpose technology with a multitude of applications. That makes AI different from, and broader than, a missile, a submarine, or a tank.”55 The similarities between the submarine and electricity cases across other features help test GMT theory against alternative explanations. One might posit that the divergent outcomes of the cases, as they pertain to GMTs, are due to differences in the disruptiveness of the two technologies. Yet the submarine was at least as disruptive as electricity. The submarine is often described as a major military innovation that changed the conduct of warfare.56 Moreover, the introduction of submarines into military affairs occurred in the same period as the introduction of electricity into military affairs.57 If we find that the two technologies affected the military balance of power in different ways, then these differences cannot be accounted for by technology-agnostic factors, such as the organizational competencies and cultures of the militaries in question, and time-dependent factors, such as the distribution of military power and the nature of military competition at the time. These overlapping trajectories help pinpoint the technological dimension of GPTs. THE ELECTRIFICATION OF WARFARE — BACKGROUND Much of our empirical analysis comes from primary sources, supplemented by secondary accounts of the history of electricity development and military modernization. From the 1870s, the establishment of trade journals such as The Electrician (London) and The Electrical Review (Chicago) provided forums for electrical engineers to deliberate over the issues of the day, including electrical applications in military affairs.58 The accounts of technical innovators involved with incorporating electricity into military operations are also valuable sources. One important figure was B.A. Fiske, who served as Rear Admiral of the U.S. Navy and invented more than 130 electrical and mechanical devices.59 Records of Fiske’s lectures and writings during the period, in addition to his autobiography, provide a firsthand evaluation of the uses of electricity in war. 55 Horowitz 2018. Emphasis ours. 56 Lautenschlager 1986, 121. 57 In fact, before WWI, advances in the capacity of electric storage batteries expanded the submergence period of submarines, making submarines an application sector of electricity as a GPT. The first electrically propelled submarines appeared in the mid-1880s. Skjong et al. 2015, 6. 58 From 1872 to 1882, 12 journals devoted to electricity sprung up in London, Paris, St. Petersburg, Berlin, New York, Chicago, and Barcelona. Manubens 2011, 171. 59 Baines et al. 1942, 10–11. In 1882, the 28-year-old Fiske foresaw the critical need for electricity in the navy of the future. Since the navy offered no training in the subject at the time, he got permission from the Navy to study at the GE
17 Our analysis of the development of electricity mainly covers the period from the mid-1800s to the end of World War I. In the late 19th century a cluster of electrical inventions — including the first practicable dynamos, the transformer, and the steam turbine — enabled the widespread application of electric power. This versatile energy system transformed systems of lighting, manufacturing, transportation and electronic devices.60 “No other technical innovation had such a far-reaching impact on modern civilization as the creation of an admirably reliable system of electricity generation, transmission, and conversion,” writes Smil.61 By tracing the evolution of military electrification across a variety of military applications, we show that the well-known versatility of electrical applications in the economic realm extended to the military domain (Figure 2).62 This process will be further fleshed out in evaluating the four propositions.
plant for a year. Capt. Wayne P. Hughes Jr. writes, "The Navy never got a greater payback from graduate education than that investment." Hughes Jr. 2017. 60 Smil 2005 61 Smil 2005 62 Extensive citations for the dates in this figure are available in an online appendix.
18 GMT FEATURE 1: BROAD IMPACT PATHWAY The main applications of electricity in the late nineteenth century were in lighting (searchlights), electric power (electric handling of weapon systems), and communications (the telephone and telegraph). Searchlights, which cast a concentrated beam of light, helped to guard harbors and forts against surprise attacks at night. First employed by the French to reveal the Prussian army’s siege of Paris in the Franco-Prussian War, searchlights provided mainly defensive advantages.63 After this episode, other nations quickly adopted searchlights into their military and naval services.64 All the forts in New York harbor were upgraded with dynamos for searchlights in 1898.65 Relatedly, incandescent electric light proved useful in signaling at nights, as military units could light up certain combinations of numbered lamps to warn of an impending attack.66 The electric firing of guns was another key component of military transformation.67 Electric guns could be fired more quickly, accurately, and systematically. By systematic, we refer to the electric firing of guns all at once with a single trigger controlled by a central director; this trend intersected with the centralization of control in “transmitting stations” on ships which would calculate the range of projectiles with electric instruments. By 1914, around half the British battlefleet had director firing, and all had centralized control with transmitting stations.68 Advances in electricity fed into telegraph and telephone communications, which enabled a military system that could coordinate mass armies. “In armies, aside from submarine and subterranean mining, the principal use of electricity is in the military telegraph service, and its importance there can hardly be 63 Pritchard 1877, 282; Hezlet 1975, 21. However, in some cases searchlights did facilitate offensive operations at night. 64 Fiske 1886, 86. 65 Marvin 1990, 145. 66 Fiske 1886. 67 Hezlet (1975, 25) categorizes electric firing for guns, electric lighting, and electric power for laying and training guns and supplying ammunition as “minor uses for electricity in warships” for which electricity provided minor improvements over other alternatives. The impact of GMTs comes from the accumulation of all these improvements across a wide distribution of applications. 68 Hezlet 1975, 82.
19 overestimated,” commented Fiske in 1890.69 In part due to the efficiency of its telegraph service, the Prussian Army was able to rapidly mobilize and achieve victory in 1870. The U.S. military also adopted electrical communications systems. New wireless technology improved military systems of telegraph communications, and the U.S. government purchased 1300 telephone sets for military use in 1905.70 Eventually, the electromagnetic spectrum became a new arena for conflict. Described by Rosen as “one of the most significant military technological innovations of the modern era,” electronic warfare was first demonstrated in the Russo-Japanese conflict of 1904.71 In the First World War, the French and British jammed and spoofed the electromagnetic signals of German bombing operations. Working for the British War Office, the Western Electric Company developed a method in 1915 to jam reception in German listening posts.72 By World War I, it was difficult for ships to function without wireless communication, and this established radio intelligence as a crucial aspect of battlefield effectiveness. British radio intelligence broke German ciphers and could track German communications at sea. Thus, “radio intelligence has been assessed as the most important single factor in the defeat of the U-boats in 1914-1918.”73 Arguably the most powerful demonstration of how electricity shaped World War I was the British penetration of German diplomatic codes, which facilitated the leakage of the Zimmerman telegram.74 In many areas, electrical applications in communications, lighting, transportation, and weapons management converged in complete electrical systems. Consider Fiske’s description of a state-of-the-art naval ship and fort in the late 19th century: “It is probable that in the near future every man-of-war and every fort will be fitted with a complete ‘electrical system’ well-protected from projectiles, which will include dynamos capable of supplying a very large amount of electrical energy to a system of mains, from which all the incandescent lights, all the search-lights and all the motors of different sorts can draw the supply of energy requisite for their needs.”75 To further supplement the system, electric technology controlled the firing of 69 Fiske 1890, 186. 70 Marvin 1990, 145. 71 The Russian military jammed the radio transmissions on Japanese battleships. Price 1984, 3-6; Rosen 1994, 190. 72 National Electrical Manufacturers Association 1946, 87. 73 Hezlet 1975, 143 74 This caused the U.S. to enter the war. According to Headrick (1991, 170), “Never before or since in history has communications power been so concentrated and so effective.” 75 Fiske 1886, p. 90.
20 guns; electric-powered velocimeters would determine the exact position of an approaching vehicle; and telegraphs would communicate the commander’s orders to all the departments of a ship. For Electrical World, the mayor of Kansas City captured the accumulated impact of electricity across a broad range of systems by 1890: “[Electricity] now not only guards the vessel from the inventions of the enemy, but aims and fires the guns, illuminates the sights that the aim may be sure, discharges torpedoes, measures her speed, is the most successful motor for submarine boats, and renders possible a system of visible telegraphy by which communications may be flashed against the clouds and understood at a distance of sixty miles.”76 To underscore this point, electricity was at the core of the development of the dreadnought, a new type of battleship often referenced as the key military innovation of this time period.77 Electrical communications and range finders supported centralized fire control, a system that facilitated effective shooting with heavy gunnery from long ranges — the crucial advance from pre-dreadnought ships to dreadnoughts.78 Wireless telegraphy was also essential for the coordination between dreadnoughts and the overall battlefleet.79 Another observable implication of this broad impact pathway is that the foreseeability of electricity’s impact on the conduct of warfare should have been very limited. As advances in electricity were emerging, many predicted that the main application in the military realm would come in the form of war-winning weapons. Experts and popular commentators alike envisioned military scenarios in which “some electrically ingenious device secured a strategic battlefield advantage.”80 Engines would deliver electric shocks “of infinite variety” on the battlefield, one publication hypothesized in 1889;81 electric rays of destruction would work “revolutionary effects on the art of modern warfare,” predicted another in 1896;82 Tesla, the scientist who designed the alternating-current electric system, proposed that electric technology would facilitate remote-controlled conflict facilitated by mechanized weapons.83 The inventor of the Gatling gun, R.F. Gatling, 76 The Electrical World 1890, 125. Cited in Marvin 1990, 145 (fn 113). 77 Gilli and Gilli 2019; Horowitz 2010 78 Lautenschlager (1983, 19) states, “The many guns of the predreadnought operated independently, the big guns of the dreadnought functions as a single system.” 79 Horowitz 2010 80 Marvin 1990, 145. 81 Electrical Review 1889, 4. 82 Marvin 1990, 146 83 Tesla 1900. Interestingly, Tesla’s vision may be materializing over a century later. This relates to another stylized fact related to the long diffusion timeline of GPTs in military affairs.
21 envisioned the invention of a powerful electrical machine that could kill whole armies at the flip of a switch would bring peace to the world.84 A 1911 edition of Technical World magazine painted a particularly vivid vision of what electric-powered warfare would look like in 1950: “The old War God hurling his thunderbolts will seem impotent beside man wielding the forces of nature for weapons. Magazines exploded without warning by darting, invisible, all-penetrating currents of electricity; devastating waves of electricity, or of some more powerful force, flashing over hundreds of miles consuming all that comes within their scourging blast. Guns, explosives, and projectiles will sink into the past, even as have the bow and arrow, giving place to howling elements clashing under man’s direction.”85 The electric rays of destruction never materialized. As Marvin summarizes, “Actual as opposed to fantasy developments in electrical warfare were mostly in the realm of communications rather than destructive weaponry.”86 Other misses were less spectacular though still significant. Even Fiske, one of the most prescient observers who played a leading role in U.S. efforts to electrify its military, severely underestimated the impact of wireless communications.87 For example, Fiske asserted in a 1904 article that the radio had “no military usefulness whatever.”88 Fiske also expected that all-electric motors would dominate marine propulsion.89 In actuality, mechanical drive became the dominant method of naval propulsion, though diesel-electric systems were widely adopted by submarines.90 In an 1892 article, one American ordnance engineer, a frontline user of electrical applications, captured the limited impact predictability of electricity, “Great as was the usefulness of electricity during the period of the Civil War…it was only a means of 84 Western Electrician (Chicago) 1891, p. 221. These mistaken predictions of electricity in the form of a war-winning weapon still lingered in the1930s. Worried about rumors that Germany possessed a “death ray,” the UK Air Ministry asked Scottish physicist Robert Watson-Watt to investigate this destructive application of electromagnetic radiation. Watson-Watt dismissed the likelihood of a death ray and responded in a memo that he was focused instead on “‘the difficult, but less unpromising, problem of radio-detection as opposed to radio-destruction.” APS News 2004. 85 La Baueme 1911, 439; quoted in Marvin 1990, 144. 86 Marvin 1990, 144. 87 “The development of wireless was the outstanding influence of the electron on sea power in the first fourteen years of the twentieth century,” writes Hezlet (1975, p. 78) One of the key figures in the development of wireless telegraphy is Guglielmo Marconi, an electrical engineer, who won the 1909 Nobel Prize in Physics for his work on the detection of electromagnetic radiation. 88 Howeth 1963, 65; cited in Douglas 1985, 119. 89 Fiske 1890. 90 O’Rourke 2000.
22 communication; and there was probably no one at that era whose imagination was sufficiently elastic to dream of electricity ever acquiring the compass it possesses at the present time."91 GMT FEATURE 2: INDUSTRIAL PRODUCTIVITY SPILLOVERS In addition to spurring a variety of military innovations, a GPT should also transform military effectiveness by boosting industrial productivity. In the case of electricity, it is well-documented that the diffusion of electricity across manufacturing industries resulted in a productivity surge.92 The U.S. case is extensively studied.93 The use of electric motors expanded from slightly less than 5 percent of installed horsepower in US manufacturing in 1899 to 55 percent by 1919.94 Crucially, electrification enabled mass production, as the adoption of electric unit drive in factories resulted in standardized workflows and plant capacity expansion.95 The impact of electrically-boosted production capabilities was revealed in the two great wars of the 20th century. Zeitlin concludes that, by WWI, the “capacity of civilian firms to manufacture large numbers of standardized weapons became increasingly central to the conduct of industrialized warfare.”96 The resultant increase in warmaking capacity was stark. For example, Britain possessed only 154 airplanes at the outbreak of WWI, but British aircraft factories were producing 30,000 planes per year by the end of the war.97 Access to cheap, plentiful electricity drove these surges. From one survey of the U.S. electrical industry’s contributions to WWI production, many leading electrical manufacturers estimated that 90-95 percent of 91 Parkhurst 1892, 359. Emphasis ours. 92 Based on an analysis of the usage of electricity in the manufacturing industries of Britain, France, Germany, Japan, and the U.S. Ristuccia and Solomou (2014) question the productivity benefits associated with electrification. Recent research, which takes a more active measure of adoption (patenting activity vs. mere usage of electricity), suggests that there was a generalized productivity boost. See Petralia 2020 “Following the Trail of a General Purpose Technology.” 93 David 1990; Devine 1982. But a similar effect occurred in other countries. German manufacturing adopted electric power systems at a similar rate to U.S. manufacturing. From 1907 to 1933 in Germany, there was an increase of 20 to 76 percent in horsepower generated from electric power as a percentage of total horsepower in the manufacturing sector. There was still a productivity gap between the U.S. and Germany because Germany inefficiently assimilated electrical systems. Timmer et al. 2016, 880-881. 94 Devine 1983. 95 Rosenberg 1998b, The Energy Journal, 13-14. 96 Zeitlin 1995, 47; McNeil 1982, pp. 330-331 97 Smil 2004, 367; see also: McNeil 1982, 330. Smil writes, “In August 1914, Britain had only 154 airplanes, but just 4 years later, the country’s aircraft factories were sending out 30,000 planes per year. Similarly, when the United States declared war on Germany in April 1917, it had fewer than 300 second-rate planes, none of which could carry machine guns or bombs on a combat mission, but 3 months later Congress approved what was at that time an unprecedented appropriation ($640 million or B$8 billion in 2000 dollars) to build 22,500 Liberty engines for new fighters.”
23 their plants were supporting government needs.98 Recognizing that the U.S. needed to realize “the greatest possible production of needed war materials of the kind peculiarly dependent upon a cheap and dependable supply of electricity,” the War Industries Board placed restrictions on civilian uses of electricity in key industrial centers, so as to maximize access for war industries.99 This connection between the electrification of manufacturing and military production only intensified in WWII. Electricity was one of the highest targeting priorities for Allied strategic bombing efforts because of its impact upon a wide range of German industrial activities.100 Historians attribute the outcome of the war to a large extent on the Allied capabilities in mass production.101 Zeitlin writes, “World War II...marks the apogee of this symbiosis between mass production and military prowess.”102 In Britain, electrical manufacturers managed state-financed “shadow factories” that helped to expand aircraft production using high-volume techniques.103 American industry produced more than 250,000 planes during World War II, which exceeded the output of Britain and Germany combined.104 Smil concludes, “There is no doubt that the rapid mobilization of America’s economic might, which was energized by a 46% increase in the total use of fuels and primary electricity between 1939 and 1944, was instrumental in winning the war against Japan and Germany.”105 GMT FEATURE 3: DELAYED EFFECTS In the economic realm, scholars hold up the diffusion of electricity as an example of the long time lag between the emergence of a GPT and its significant bearings on national productivity. Measured by percentage of total installed horsepower in manufacturing industries, adoption of unit drive, and estimates of 98 Electrical Review 1919, 363-364. This was a survey of American manufacturers. For similar developments in Britain, see Bangs 1917. 99 Keller 1921. 100 As the strategic bombing campaigns progressed, electricity fell in priority — a development later critiqued by a retrospective United States Strategic Bombing Survey: “The German electric supply system ...was extremely vulnerable to bombing attack, and, had it been attacked systematically, it would have severely crippled Germany's industrial war machine." Quoted in Kuehl 2008, 239. 101 McNeill 1982, 355, 358-59. Kennedy 1987, 244, 248-249. 102 Zeitlin 1995, 47. One of Germany’s strategic blunders was its underestimation of "the massive potential of the United States for industrial mobilization and production." (Millet et al. 1986, 47). 103 Zeitlin 1995, 50. 104 Smil 2004, 367. 105 Smil 2004, 368.
24 electricity’s contributions to GDP growth, American electrification did not take off until the 1920s.106 This was a full four decades after major advances like the dynamo and incandescent light bulb emerged. Do we find a similarly delayed timeline for the electrification of the military? Though it is more difficult to track the effects of electricity on military productivity, we can trace how and when complementary innovations in different military branches were first introduced as military capabilities, first used in warfare, and were fully adopted as a standard military capability (Table 1).107 Combined with the fuller technology tree introduced above, this mapping exercise of two electrical military innovations shows that even the early movers did not achieve widespread adoption until right before WWI. Among later adopters, widespread adoption did not take place until the interwar period or after WWII. Table 1: Delayed Impact of Electrical Military Innovations Wireless telegraphy (radio) Electric firing of guns Complementary innovation Hertz’s demonstration of radio waves (1888) AC induction motor (1886) Application Sector Navy (early) Air Force (late) Navy (early) Army (late) First Introduction as Military Capability British fleet equipped with wireless telegraph (1900) Planes equipped with radio at end of WWI (1916-1918) Electric firing introduced in British navy (early 1870s) GE develops electric-powered miniguns (1950s) First Application in War Russo-Japanese War (1904-1905) WWI (but very ineffective) World War I (1914-1918) Vietnam War (1960s) Widespread Adoption British Royal Navy has “patchy global network” that supported radio communication (1914) Germany equips air force with complete set of radio equipment (1938) Half of British battlefleet had director firing (by 1914) U.S. procured 10,000 miniguns during Vietnam War (1960s) Demand indicators, for both electrical engineering talent and electrical voltage on ships, point to a significant acceleration of military electrification around World War II. Military applications of electricity across vehicles and firearms grew so widespread before World War II that Britain established the Royal 106 For share of horsepower estimates see Rosenberg 1979, 48; Devine 1982, 46-47; or adoption of unit drive, see Devine 1983; for estimates of electrification’s contributions to GDP growth, see Crafts 2002. 107 Extensive citations for the dates in this table are available in an online appendix.
25 Electrical and Mechanical Engineers (REME) in 1942, which was devoted to the maintenance and repair of electrical equipment. By May 1945, the unit had 158,000 officers and men.108 In his 1960 presidential inauguration speech to Britain’s Institution of Electrical Engineers, Sir Hamish D. Maclaren reviewed the trendline in electrical demand on first-class ships. In the 1880s, these were equipped with only 3 dynamos, each supplying 200 amps at 80 volts in the 1880s. After World War II, the transition from 500kW d.c. machines to 1000kW a.c. generators enabled first-class ships to meet greater electrical demand.109 By 1960 electrical installations accounted for 30 percent of the total cost of British warships.110 Indeed, some of the most significant military applications of electricity did not emerge until the 1940s. The radar is the best example. Developed based on principles first seeded by Hertz’s 1888 discovery of the reflective properties of electromagnetic waves, radar systems came to play a pivotal role in World War II. The U.S. Office of Scientific Research and Development, established in 1941 to coordinate scientific research for military purposes, spent more of its $457 million budget on radar and radar countermeasures ($128 million) than any other category.111 This investment paid off. According to one estimate, Allied electronic countermeasures reduced the effectiveness of German antiaircraft by 70 to 75 percent.112 The delayed diffusion of military electrification was due to the need for significant organizational adaptations and skills upgrading. Just as economic observers pushed for the cultivation of the electrical engineering discipline, the pioneers of military electrification pushed for more military electricians. In 1890, Fiske proposed the formation of a corps of naval and military electricians to assist the army and navy. It was not until eight years later during the Spanish War when Captain Eugene Griffin, VP of General Electric, formed such a corps of electricians and mechanical engineers.113 This volunteer corps of one thousand engineers served as the foundation for the work of U.S. electricians and engineers in the First World War.114 108 National Army Museum, n.d. 109 MacLaren 1961, 4. 110 MacLaren 1961, 3. This was a growth in the cost of components that mirrors the growing cost of software in modern fighter jets. 111 Rosen 1994, 190-191. 112 This estimate is based on U.S. government interrogations of operators of German antiaircraft guns and German radar scientists. Rosen 1994, 198. 113 Fiske 1919, 130. 114 Fiske 1919, 239.
26 Organizational barriers also prevented the fast uptake of various military applications of electricity. Take, for instance, the U.S. Navy’s delayed adoption of wireless telegraphy (radio communications). Despite the opportunity to employ radio communications as early as 1899, the U.S. Navy took about fifteen years to fully integrate the radio into its operations, as senior naval officers saw the radio as a direct threat to their authority onboard ships. This ties back to how GPTs often require radical structural shifts like the redesign of factories. Before the radio, ship captains and fleet commanders were masters of their domain; but advances in electrical communications enabled military bureaucrats and land-based rivals to challenge their authority.115 Taylor concludes, “It would take World War I, combined with several acts of Congress and executive orders by President Wilson, to finally force the shift to a radio navy.”116 COMPARISON TO SUBMARINES AS A NON-GPT Not all technologies interact with military systems in the same way. Along all three features of GMTs, the evolution of submarines in military affairs differed from military electrification.117 First, advances in underwater submersion technology affected military effectiveness through a much narrower pathway, largely in the application of submarines as weapons platforms.118 During the late 19th century and early 20th century, underwater submersion technology had very few civilian applications, resulting in limited effects on industrial productivity.119 Lastly, there was a relatively short delay between the introduction of the first modern submarines, which occurred around the turn of the 20th century, and their impact on military effectiveness, which was apparent in the years before World War I.120 115 Taylor 2016, 192. See Douglas 1985. 116 Taylor 2016, 192. 117 We only briefly survey the impact of submarines on military affairs due to both space constraints and the fact that this is relatively well-covered terrain. Electricity, on the other hand, is relatively understudied. Thus, we focus the bulk of our empirical presentation on piecing together the historical data for the electrification of warfare. For military histories of submarines, see Coté Jr. 2003; Lautenschlager 1986. For an extended discussion of submarines as a non-GMT, see the online appendix. 118 The only other substantial application was in reconnaissance. In World War I, the British did use submarines as important supplements to radio intelligence, as the development of new wireless transmitters allowed patrol submarines to relay information back to the larger fleet. Hezlet 1975, 133. 119 More recent advances in submarine technology do benefit radar and ocean imaging capabilities. 120 The Holland, introduced by the U.S. Navy in 1900, is generally considered the first modern submarine. Coté Jr. 2003, 5. By 1913, all of the leading navies had substantial submarine fleets. Crisher and Souva 2014.
27 INDUSTRIAL DEPENDENCY HYPOTHESIS The broad applicability of GPTs across many sectors, combined with the fact that the civilian realm presents many more application scenarios than the military realm, means that the military will grow more dependent on the civilian sector as the GPT develops. While this effect exists to some degree for many (but not all) dual-use technologies,121 it is significantly enhanced with the GPTs due to their broad applicability across many sectors. This was the case with electricity. The amount of money invested in the research and development of electrical engineering departments and electric companies across the United States was “many times greater than that invested in all [the U.S.’s] ships put together.”122 The civilian economy attracted the brightest minds in electricity. As Fiske recounts, “The principal difficulty that electricity has had to meet in our Navy has been...that most of those who have become proficient (enlisted men as well as officers) have gone into civil life, and we find them distributed among the various colleges and electrical enterprises of the land…so far as helping the Navy goes, their services are lost.”123 The widespread applicability of electricity across so many military domains highlighted the need for military-wide talent upskilling. “Every soldier or sailor, if he desires to make his mark, must be something of an electrician, for there seems to be no limit to the useful applications of the galvanic spark in battle,” wrote Pritchard in an 1877 Nature article.124 In the navy, electric applications significantly enhanced the complexity of ships, which gave greater advantage to the navy that devoted sufficient time to education and mastering the difficulties of these new systems.125 The U.S. Congressional Amalgamation Act of 1899 represents the most dramatic demonstration of how militaries faced pressures to adapt their skill base to GPT trajectories. The act merged the separate 121 Military aircraft technology has become less dependent on civilian aircraft technology. The connections between nuclear weapons technology and civilian nuclear technology are also muted. 122 Fiske 1896, 424. 123 Fiske 1896, 327. 124 Pritchard 1877, 281–282. 125 Fiske 1896, p. 327
28 Engineering Corps with naval line officers, setting the principle that all naval officers needed to be engineers in some sense. One study of the effects of the Amalgamation Act on technology-skill complementarities concluded that it was “likely important for the Navy to continue to implement engineer-intensive general purpose technologies.”126 Within the broader case of the electrification of military affairs, we exploit variation between the British and Russian militaries in their adoption of electrical military innovations to test our proposition about the leveling effect of GPTs. 127 As referenced in earlier sections, the British were very successful in building an effective system of wireless military communications and in manipulating other countries’ wireless communications in World War I. Compared to its Russian competitor, Britain’s military could draw from a more robust industrial base in electricity (Figure 3). This was encapsulated in the British advantage in signals intelligence in fighting “the world’s first electronic war,” a campaign that “used the best technology and science and electrical engineers of the day.”128
In the following sections we contrast this experience with the Russian case. Before World War I, Russia was a first rank power and perceived by others to be growing rapidly in power.129 However, its military 126 Glaser and Rahman 2014. 127 In this section we only discuss electrical communications. But the British were more successful in adapting to military electrification in general. For instance, the British were also ahead in electric fire control. Hezlet 1974, 82. 128 Ferris 1996, 29. 129 Wohlforth 1987
29 was a technological laggard with respect to electrical communications. In fact, according to one comparison of the signal communication systems of the British and Russian militaries at the onset of World War I, Britain and Russia were at complete opposite ends of the spectrum.130 This failure came to the fore at the Battle of Tannenberg between Russian and German forces at the onset of World War I, to dramatic effect. The Russian army suffered a massive defeat, and 100,000 men were captured.131 According to George I. Back, who served as the chief signal officer of the Mediterranean theater for the U.S. Army, and George Thompson, a military historian, Russia’s setback at Tannenberg “was largely due to an almost total lack of signal communication.”132 Because the Russian military lacked both the electrical equipment and the requisite technical knowledge base for using these devices and encrypting electrical communications, the Germans had access to detailed Russian communications and marching orders.133 One key factor behind the Russian military’s relative failure with adopting electricity was its weakly developed industrial base in electricity. The lack of a unified technical profession and the lack of skilled personnel functioned as a bottleneck on Russian military electrification.134 With no leading electrical companies of their own, the Russian electrical engineering sector depended on foreign suppliers for critical components such as high voltage transformers and measuring instruments.135 When imports were cut off amidst war, Russian industry could not independently manufacture this equipment, even with support from the Russian war industry committee.136 If industrial dependency holds, then even a strong, technologically savvy military cannot bring about a GMT on its own. Before World War I, the Russian military tried to promote electrification by whatever means necessary. It sponsored research and travel, trained scientists, pushed for the creation of a domestic 130 Back and Thompson 1998. 131 Headrick (1991, 156) describes the Russians as “the worst off” with respect to radio equipment. 132 Back and Thompson 1998. 133 One of the two main Russian armies at Tannenburg only had twenty-five telephones and a handful of manual Morse Code machines. Jackson 2002, 4. 134 Coopersmith 1992, 97-101 135 Graham 2013, 28-29; Coopersmith 1992, 97-101 136 Coopersmith 1992, 104. For another case of slow Russian military electrification, in which the Russian Navy was hampered by its reliance on German wireless equipment in the Russo-Japanese War. see Lee 2012, 25-28; Hezlet 1975, 42-49.
30 industry, gathered information about the latest developments in electric technology, functioned as a testbed for materials and systems, and also served as the largest procurer of electricity.137 While the Russian military did achieve success in some limited, early applications of electricity, this success did not extend to electrification as a whole. Russian electric mine technology, for instance, was relatively advanced. Russian electric mines damaged two British ships in the Crimean War, and the Japanese sent a delegation to the Russian Mine School in 1877 to learn from Russian officers about the production, maintenance, and use of electric mines. However, this competent pool of workers was “one of the few such groups in Russia.”138 Unfortunately for the Russian military, taking full advantage of a GPT requires much more than skill competencies in a few application sectors; it requires a robust industrial base familiar with spreading the GPT. What about other factors besides industrial dependency? For instance, Russia’s failure to exploit the electricity-based GMT could stem from the fact that the Russian economy as a whole was more underdeveloped than the British. The divergent rates of military electrification could be due to differences in the two militaries’ organizational and cultural attitudes toward emerging technologies. To control for alternative explanations, we examine trends in submarine technologies, which do not fulfill the characteristics of a GPT. If Russia was able to keep pace with Britain adopting submarines, then it is less likely that general economic, organizational, and cultural factors were driving differences in military electrification. In line with these expectations, we find that the Russian navy was closer to parity with the British in terms of submarine capabilities than military electrification. Before the outbreak of WWI, the Russian navy was equipped with 18 diesel-electric submarines,139 of which the Bars and Morzh class submarines were on par with the best foreign counterparts. Russian submarine engineers were also experienced and knowledgeable, and Russian submarines conducted effective operations in the Black Sea during the war.140 Overall, there was significantly less of a gap between Russian and British submarine capabilities than the gap between Russia’s GMT in electricity and Britain’s GMT in electricity. 137 Coopersmith 1988. 138 Coopersmith 1988, 298. 139 Crisher and Souva 2014 140 Polmar and Noot 1991, 28-29; Muraviev 2015, 86-87
31 Since advances in submarine capabilities did not generate a GMT, a military’s connection to a robust industrial base should be less significant for explaining which militaries are differentially advantaged by submarines. Like Russia’s electrical base, Russia’s civilian shipbuilding industry was weak, but this weakness was not as significant in determining the military effectiveness of Russian submarine capabilities.141 Developing effective submarine capabilities required a relatively narrow talent base skilled in operating high-speed reciprocating machinery.142 Russia struggled more with electrification because of the demand for general upskilling drawing on a broad industrial base connected to electrical advances. In fact, one of the main shortcomings of Russian submarines was linked to the Russian military’s weakness in electrification. In World War I Russia’s dependence on Germany for diesel-electric engines for stronger propulsion was exposed when war broke out.143 While this is a promising first step in the study of GMTs, the generalizability of a single case is limited. It is also important to specify scope conditions when extending the implications of the electricity case to other possible GMTs. Whether the lessons from history about the effects of electricity on military affairs can apply to current developments in AI depends not just on whether the technical properties of AI are comparable to those of electricity but also on the congruence of fit between the current period and the nature of military competition and dynamics of the international system in the late 19th century.
141 Russian shipyards were notoriously slow. Hauner 2004, 94. 142 Nimitz 1916, 487 143 Polmar and Noot 1991, 28-29; Muraviev 2015, 86-87. Russia quickly realized the strategic potential of submarines as new combat platforms, and Russian submarines played an effective role during the Russo-Japanese War and World War I. In World War I, the Russian submarine fleet made the German-Turkish naval forces sustain significant losses, and the Soviet submarine school, after the war, emerged as one of the world’s leading centers of submarine warfare.
32 IV. Conclusion and Lessons for AI The evidence from the electricity case is compatible with GMT theory. Similar to the trajectory of electricity in the economic domain, electricity found widespread military applications only after a protracted period of gestation. In addition to directly boosting military effectiveness through a broad array of military innovations, electricity also indirectly transformed military power through stimulating industrial productivity. Institutional adaptations (both civilian and military) to widen the talent base associated with electricity, especially in electrical engineering, were crucial to help certain militaries benefit more from the GMT. These findings should be especially relevant for discussions of how military affairs will be transformed by AI — the “next GPT.”144 Based on the features of GMTs and the industrial dependency hypothesis, we can state some predictions for how AI will affect the future of warfare. Our arguments contribute to existing discussions about the effect of AI in military affairs, which often emphasize AI as a Revolution in Military Affairs that rising military powers can take advantage of to leapfrog the U.S. in weapons capabilities.145 While AI and electricity are both GPTs, they differ along many other relevant characteristics. Autonomy, for instance, is a distinctive characteristic of some AI systems. Moreover, since the late 19th century, there have been changes in the nature of military competition and innovation. All of these contextual factors need to be taken into account when generalizing our GMT-related findings to the AI case. Still, our analysis can provide an initial guide for comprehending the impact of AI on military affairs, akin to how studies of electricity’s effect on economic transformation have improved discussions about the impact of computers on productivity.146 First, speculation about how AI will transform military affairs places excessive emphasis on the narrow effects of weapon systems. Autonomous weapons systems have drawn a large share of the attention 144 Trajtenberg 2018 145 For more on RMAs see Krepinevich 1994; Krepinevich 2002. 146 See, for example, David 1990
33 from policymakers and scholars alike, as have narratives of an “AI arms race.147 U.S. defense intellectuals highlight how China could take advantage of AI-enabled hypersonic missile systems to leapfrog U.S. military power.148 As Elsa Kania notes, the Chinese military’s focus on “trump card” or “assassin’s mace” weapons that can counter U.S. capabilities “will likely persist in the PLA’s approach to AI.149 In contrast, a GMT approach emphasizes the accumulation of AI-enabled improvements across many military systems. This impact pathway will likely involve significant upgrades to weapons capabilities, as was the case with electricity and centralized fire control. On the whole, though, effects of AI advances in other military domains, including communications, cyberspace operations, intelligence, information and psychological operations, logistics, strategic decision-making, etc. will be more consequential.150 Moreover, the focus on AI weapons neglects the indirect pathway of influence through AI’s potential to upgrade the productive capabilities of the overall industrial base. In particular, the intelligentization of manufacturing lines (smart manufacturing) could have significant follow-on effects for military readiness. Second, existing conjectures about the impact of AI on military affairs severely underestimate the timeframe for when substantial effects will occur. One highly-cited article on AI and national security argues, “The amount of progress AI technology is poised to make over the next 10-20 years should lead the Department of Defense to revisit those assumptions (about spending priorities on aircraft and naval platforms).”151 Payne also estimates, “(T)he rapid progress in AI research, especially of hybrid approaches that utilise multiple AI techniques, along with increasingly powerful hardware on which to run algorithms, suggests the potential for AI to significantly affect existing military activities in the short to medium term…”.152 This is reflective of a broader tendency in strategic thinking: to conflate rapid progress in a 147 Zwetsloot et al. 2018; Security scholars tend to compare AI to nuclear weapons based on the fact that “both technologies offer new military capabilities that have the potential to transform strategy and to dramatically affect the balance of power.” (Payne 2018, p. 7) 148 Work and Grant 2019 149 Kania 2017, 33-34; Kania 2020. 150 Others have noted the breadth of military operations that AI could transform, which covers training, recruitment, and war-gaming. Horowitz 2018; Kania 2017; Hoadley and Sayler 2020; See also Montgomery 2019, 324-325. 151 Allen and Chan 2017, 61. 152 Payne 2018, 10
34 technological field, which is characteristic of GPTs, with rapid adoption across military applications, which is uncharacteristic of GPTs. According to the GMT theory, the most consequential impacts of AI on military effectiveness will occur only after a long period of gestation. Economists have already begun to model implementation lags in the effects of AI on economic productivity.153 A similar extended trajectory will apply in the military realm. The current wave of AI development started with the deep learning revolution in the early 2010s, so if AI follows the same timeline as electricity, a prolonged period of gestation could extend until around the 2050s.154 In addition, since the development of AI is still in its early stages, the foreseeability of its military applications is very limited. Even the most astute observers of military transformation at the turn of the 20th century, twenty years after the introduction of the electric dynamo, could not envision how the technology would transform military affairs. As only a decade has passed since critical breakthroughs in deep learning, any attempt to foreordain the ultimate strategic impacts of AI should be met with deep skepticism. Our imaginations — to borrow language from the ordnance engineer quoted earlier — are not sufficiently elastic. Lastly, the GMT theory supplements and modifies existing thinking about the rate of diffusion of military applications of AI and the effect of AI on the military balance of power. Some scholars argue that if military advances in AI continue to be closely linked to civilian applications, then military AI capabilities will rapidly diffuse to other countries.155 Other factors commonly cited include the financial and organizational requirements for adopting military AI technology. Informed by a historical perspective of GMTs, we view “military AI technology” as not a singular technological innovation but part of a GPT trajectory, which encompasses a broad distribution of technological applications. Just like the organizational requirements for adopting wireless telegraphy were different from those required to adopt searchlights, the adoption capacity for different applications of AI will vary. 153 Brynjolfsson et al. 2017 154 This expectation can be affected by other factors, including the possibility that the general process of technological adoption is accelerating. Some evidence indicates that the waiting time for a significant productivity boost from a new GPT has decreased over time. Crafts 2004. 155 Drezner 2019; Horowitz 2018.
35 To more fully account for how AI advances will differentially advantage certain militaries, more attention should go to factors that apply across the broad front of a GPT trajectory.156 We highlight the significance of a state’s industrial capacity to provide AI infrastructure and skilled labor to militaries. Specifically, militaries able to draw from a wide skill base in AI will better exploit the AI-based GMT. Crucially, the talent base required for AI differs from the talent base required for other revolutionary dual-use technologies like nuclear power. GMT theory suggests that military linkages to a wide base of AI engineering talent, rather than star researchers or cutting-edge technical capabilities, are crucial to adapting generalized models to a variety of specific military applications.157 The three great influences on naval warfare of the 20th century were the aircraft, the submarine, and electricity. We have shown that electricity is not like the others. It powered a GMT. In parallel characterization, Horowitz identifies three key technologies that could reshape the future of warfare in the 21st century: cyber, drones, and AI.158 As plausibly the defining GPT of our century, AI is not like the others. After all, it’s the new electricity.
156 It is important to note that specialized military applications of AI, with few linkages to the commercial domain, do exist and will emerge. See Horowitz 2020b. 157 Payne 2018; Ryseff 2020. Our conclusions diverge from existing discussions in the U.S. that focus on the DoD’s access to top elite talent in top tech firms. 158 Horowitz 2020a; Montgomery (2019, 324) writes that “AI is often identified as the emerging technology that could most influence military power.”
36 References Allen, Greg, and Taniel Chan. “Artificial Intelligence and National Security.” Cambridge, MA: Belfer Center for Science and International Affairs. Accessed December 27, 2020. https://www.belfercenter.org/publication/artificial-intelligence-and-national-security. Back, George, and George Thompson. “Military Communication.” Encyclopedia Britannica, 1988. https://www.britannica.com/technology/military-communication. Baines, F.X., Eugene Kinkead, and Russell Maloney. “The Admiral’s Chair.” The New Yorker, January 31, 1942. Bangs, F.T. “Electrical Mobilization in Great Britain.” Electrical Review and Western Electrician, May 26, 1917. Beach, Derek, and Rasmus Brun Pedersen. Causal Case Study Methods: Foundations and Guidelines for Comparing, Matching, and Tracing. Ann Arbor: University of Michigan Press, 2016. Beckley, Michael. “Economic Development and Military Effectiveness.” Journal of Strategic Studies 33, no. 1 (February 1, 2010): 43–79. https://doi.org/10.1080/01402391003603581. ———. “The Power of Nations: Measuring What Matters.” International Security 43, no. 2 (November 1, 2018): 7–44. https://doi.org/10.1162/isec_a_00328. Bode, Ingvild, and Hendrik Huelss. “Autonomous Weapons Systems and Changing Norms in International Relations.” Review of International Studies 44, no. 3 (July 2018): 393–413. https://doi.org/10.1017/S0260210517000614. Bresnahan, Timothy. “General Purpose Technologies.” In Handbook of the Economics of Innovation, edited by Bronwyn H. Hall and Nathan Rosenberg, 2:761–91. Elsevier, 2010. Bresnahan, Timothy, and Manuel Trajtenberg. “General Purpose Technologies ‘Engines of Growth’?” Journal of Econometrics 65, no. 1 (1995): 83–108. Brynjolfsson, Erik, Daniel Rock, and Chad Syverson. “Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics.” National Bureau of Economic Research, November 6, 2017. https://doi.org/10.3386/w24001. “Business Conditions in the Electrical Industry.” Electrical Review, September 7, 1918. Cantner, Uwe, and Simone Vannuccini. “A New View of General Purpose Technologies.” Jena Economic Research Papers, 2012. Castaldi, Carolina, Roberto Fontana, and Alessandro Nuvolari. “‘Chariots of Fire’: The Evolution of Tank Technology, 1915–1945.” Journal of Evolutionary Economics 19, no. 4 (August 1, 2009): 545–66. https://doi.org/10.1007/s00191-009-0141-0. Comin, Diego A, and Bart Hobijn. “The CHAT Dataset.” Working Paper. National Bureau of Economic Research, September 2009. https://doi.org/10.3386/w15319. Coopersmith, Jonathan. “The Electrification of Russia, 1880-1926.” In The ELectrification of Russia, 1880-1926, edited by Jonathan Coopersmith, 99–120. Cornell University Press, 1992. ———. “The Role of the Military in the Electrification of Russia, 1870–1890.” In Science, Technology and the Military, edited by E. Mendelsohn, Merritt Roe Smith, and P. Weingart, 291–305. Springer, 1988. National Army Museum. “Corps of Royal Electrical and Mechanical Engineers | National Army Museum.” Accessed December 27, 2020. https://www.nam.ac.uk/explore/corps-royal-electrical-and-mechanical-engineers. “Cote - Innovation in the U.S. Navy’s Silent Cold War Stru.Pdf.” Accessed January 19, 2021. https://digital-commons.usnwc.edu/cgi/viewcontent.cgi?article=1017&context=usnwc-newport-papers. Coté Jr, Owen R. The Third Battle: Innovation in the U.S. Navy’s Silent Cold War Struggle with Soviet Submarines. Newport, R.I.: Naval War College, 2003. Crafts, Nicholas. “Productivity Growth in the Industrial Revolution: A New Growth Accounting Perspective.” The Journal of Economic History 64, no. 2 (2004): 521–35. Crisher, Brian Benjamin, and Mark Souva. “Power at Sea: A Naval Power Dataset, 1865–2011.” International Interactions 40, no. 4 (August 8, 2014): 602–29. https://doi.org/10.1080/03050629.2014.918039. David, Paul A. “The Dynamo and the Computer: An Historical Perspective on the Modern Productivity Paradox.” The American Economic Review 80, no. 2 (1990): 355–61.
37 Devine, Jr. “Historical Perspective on the Value of Electricity in American Manufacturing.” Oak Ridge Associated Universities, Inc., TN (USA). Inst. for Energy Analysis, September 1, 1982. https://doi.org/10.2172/6774921. Devine, Warren D. “From Shafts to Wires: Historical Perspective on Electrification.” The Journal of Economic History 43, no. 2 (June 1983): 347–72. https://doi.org/10.1017/S0022050700029673. Dombrowski, Peter, and Eugene Gholz. “Identifying Disruptive Innovation: Innovation Theory and the Defense Industry.” Innovations: Technology, Governance, Globalization 4, no. 2 (2009): 101–17. Douglas, Susan J. “Technological Innovation and Organizational Change : The Navy’s Adoption of Radio, 1899 - 1919.” In Military Enterprise and Technological Change : Perspectives on the American Experience, edited by Merritt Roe Smith, 117–73. Cambridge, MA: MIT Press, 1985. Drezner, Daniel W. “Technological Change and International Relations.” International Relations 33, no. 2 (June 2019): 286–303. https://doi.org/10.1177/0047117819834629. “Electric Shells in Warfare!” Electrical Review, February 16, 1889. “Electricity in Warfare.” Western Electrician (Chicago), April 18, 1891. Ferris, John. “Airbandit: C3I and Strategic Air Defence during the First Battle of Britain, 1915-1918.” In Strategy & Intelligence: British Policy During the First World War, edited by M. Dockrill, 23–66. London: Hambledon Press, 1996. Field, Alexander J. “Does Economic History Need GPTs?” SSRN Scholarly Paper. Rochester, NY: Social Science Research Network, September 28, 2008. https://papers.ssrn.com/abstract=1275023. Fiske, B. A. “Electricity in Warfare.” Journal of the Franklin Institute 121, no. 2 (February 1, 1886): 81–93. https://doi.org/10.1016/0016-0032(86)90097-9. Fiske, Bradley A. “Electricity in Warfare.” Journal of the Franklin Institute 130, no. 3 (September 1, 1890): 183–203. https://doi.org/10.1016/S0016-0032(90)90116-Z. Fiske, Bradley Allen. From Midshipman to Rear-Admiral. Century Company, 1919. Fogel, Robert William. Railroads and American Economic Growth: Essays Econometric History. Baltimore: Johns Hopkin, 1964. Garfinkel, Ben, and Allan Dafoe. “How Does the Offense-Defense Balance Scale?” Journal of Strategic Studies 42, no. 6 (September 19, 2019): 736–63. https://doi.org/10.1080/01402390.2019.1631810. Gilli, Andrea, and Mauro Gilli. “The Spread of Military Innovations: Adoption Capacity Theory, Tactical Incentives, and the Case of Suicide Terrorism.” Security Studies 23, no. 3 (July 3, 2014): 513–47. https://doi.org/10.1080/09636412.2014.935233. ———. “Why China Has Not Caught Up Yet: Military-Technological Superiority and the Limits of Imitation, Reverse Engineering, and Cyber Espionage.” International Security 43, no. 3 (February 1, 2019): 141–89. https://doi.org/10.1162/isec_a_00337. Gilpin, Robert. U.S. Power and the Multinational Corporation. London: Palgrave Macmillan UK, 1975. Glaser, Darrell J., and Ahmed S. Rahman. “Engineering and Labor Specialization during the Industrial Revolution.” Cliometrica 8, no. 2 (May 1, 2014): 173–200. https://doi.org/10.1007/s11698-013-0098-y. Gourley, Scott R. “The Evolution of the M134D Minigun.” Defense Media Network (blog), November 14, 2019. https://www.defensemedianetwork.com/stories/the-evolution-of-the-m134d-minigun/. Grissom, Adam. “The Future of Military Innovation Studies.” Journal of Strategic Studies 29, no. 5 (October 1, 2006): 905–34. https://doi.org/10.1080/01402390600901067. Hauner, Milan. “Stalin’s Big-Fleet Program.” Naval War College Review 57, no. 2 (2004): 87–120. Headrick, Daniel R. The Invisible Weapon: Telecommunications and International Politics, 1851-1945. Oxford, New York: Oxford University Press, 1991. Helpman, Elhanan, and Manuel Trajtenberg. “A Time to Sow and a Time to Reap: Growth Based on General Purpose Technologies.” National Bureau of Economic Research, September 1, 1994. https://doi.org/10.3386/w4854. Hezlet, Arthur Richard. The Electron and Sea Power. London: P. Davies, 1975. “Historical Firearms - Electrically Powered Gatling Gun Since the Late...” Accessed January 19, 2021. https://www.historicalfirearms.info/post/86249885754/electrically-powered-gatling-gun-since-the-late.
38 Horowitz, Michael. “Artificial Intelligence, International Competition, and the Balance of Power.” Texas National Security Review 1, no. 3 (2018). https://tnsr.org/2018/05/artificial-intelligence-international-competition-and-the-balance-of-power/. ———. The Diffusion of Military Power: Causes and Consequences for International Politics. Princeton, NJ: Princeton University Press, 2010. Horowitz, Michael C. “AI and the Diffusion of Global Power.” Centre for International Governance Innovation (blog), November 16, 2020. https://www.cigionline.org/articles/ai-and-diffusion-global-power. ———. “Do Emerging Military Technologies Matter for International Politics?” Annual Review of Political Science 23, no. 1 (2020): 385–400. https://doi.org/10.1146/annurev-polisci-050718-032725. Horowitz, Michael C., and Shira Pindyck. “What Is A Military Innovation And Why It Matters.” SSRN Scholarly Paper. Rochester, NY: Social Science Research Network, 2020. https://doi.org/10.2139/ssrn.3504246. Howeth, Linwood S. History of Communications-Electronics in the United States Navy. Washington: [For sale by the Superintendent of Documents, U. S. Govt. Print. Off.], 1963. https://catalog.hathitrust.org/Record/001624164. Jackson, Frederick E. “Tannenberg: The First Use of Signals Intelligence in Modern Warfare.” Strategy Research Project. U.S. Army War College, 2002. Jukes, Geoffrey. The Russo-Japanese War 1904-1905. Illustrated edition. Oxford: Osprey Publishing, 2002. Kania, Elsa B. “‘AI Weapons’ in China’s Military Innovation.” Brookings (Blog) (blog), April 2020. ———. “Battlefield Singularity: Artificial Intelligence, Military Revolution, and China’s Future Military Power.” CEnter for a New American Security, 2017. Keller, Charles. The Power Situation during the War,with Appendixes Containing Reports on Conditions in the Several Power Districts /. Washington, DC: Government Printing Office, 1921. http://hdl.handle.net/2027/uc1.$b108315. Kelly, Kevin. “The Three Breakthroughs That Have Finally Unleashed AI on the World.” Wired, October 27, 2014. https://www.wired.com/2014/10/future-of-artificial-intelligence/. Kennedy, Paul M. The Rise and Fall of the Great Powers: Economic Change and Military Conflict from 1500 to 2000, 1987. Kirshner, Jonathan. “Political Economy in Security Studies after the Cold War.” Review of International Political Economy 5, no. 1 (January 1, 1998): 64–91. https://doi.org/10.1080/096922998347651. Krepinevich, Andrew F. “Cavalry to Computer: The Pattern of Military Revolutions.” The National Interest 37 (1994): 30–42. ———. “The Military-Technical Revolution: A Preliminary Assessment.” Center for Strategic and Budgetary Assessments, 2002. Kuehl, Daniel T. “Airpower vs. Electricity: Electric Power as a Target for Strategic Air Operations.” Journal of Strategic Studies 18, no. 1 (March 1, 1995): 237–66. https://doi.org/10.1080/01402399508437585. La Baueme, E.I. “Visions of 1950.” Technical World (Chicago, December 1911. Lautenschläger, Karl. “Technology and the Evolution of Naval Warfare.” International Security 8, no. 2 (1983): 3–51. https://doi.org/10.2307/2538594. Lautenschlager, Karl. “The Submarine in Naval Warfare, 1901-2001.” International Security 11, no. 3 (1986): 94–140. https://doi.org/10.2307/2538886. Layne, Christopher. “The US–Chinese Power Shift and the End of the Pax Americana.” International Affairs 94, no. 1 (January 1, 2018): 89–111. https://doi.org/10.1093/ia/iix249. Lee, Bartholemew. “Wireless — Its Evolution from Mysterious Wonder to Weapon of War, 1902-1905.” AWA Review, California Historical Radio Society 25 (2012). Lipsey, Richard G., Cliff Bekar, and Kenneth Carlaw. “What Requires Explanation.” In General Purpose Technologies and Economic Growth, edited by Elhanan Helpman, 2:15–54. MIT Press Cambridge, MA, 1998. Lipsey, Richard G, Kenneth Carlaw, and Clifford Bekar. Economic Transformations: General Purpose Technologies and Long-Term Economic Growth. Oxford; New York: Oxford University Press, 2005. MacLaren, H.D. “Inaugural Address.” Proceedings of the IEE - Part B: Electronic and Communication Engineering 108, no. 37 (January 1961). https://doi.org/10.1049/pi-b-2.1961.0001.
39 Manubens, Joan Carles Alayo. “Electricity in Spain, Its Introduction and Industrial Development.” In History of Technology, edited by Ian Inkster and Angel Calvo, Vol. 30. Bloomsbury Academic, 2011. Marvin, Carolyn. When Old Technologies Were New: Thinking About Electric Communication in the Late Nineteenth Century. Oxford University Press, 1990. McNeill, William H. The Pursuit of Power: Technology, Armed Force, and Society since A.D. 1000. 1st edition. Chicago: University of Chicago Press, 1982. Millett, Allan R., Williamson Murray, and Kenneth H. Watman. “The Effectiveness of Military Organizations.” International Security 11, no. 1 (1986): 37–71. Misa, Thomas J. “Military Needs, Commercial Realities, and the Development of the Transistor, 1948-1958.” In Military Enterprise and Technological Change, edited by Merritt Roe Smith, 253–87. Cambridge, Mass.: MIT Press, 1985. Modelski, George, and William R Thompson. Leading Sectors and World Powers: The Coevolution of Global Politics and Economics. Columbia S.C.: University of South Carolina Press, 1996. Montgomery, Evan Braden. “Signals of Strength: Capability Demonstrations and Perceptions of Military Power.” Journal of Strategic Studies 43, no. 2 (February 23, 2020): 309–30. https://doi.org/10.1080/01402390.2019.1626724. Morgenthau, Hans J. “The Four Paradoxes of Nuclear Strategy.” The American Political Science Review 58, no. 1 (1964): 23–35. https://doi.org/10.2307/1952752. Mukunda, Gautam. “We Cannot Go On: Disruptive Innovation and the First World War Royal Navy.” Security Studies 19, no. 1 (February 26, 2010): 124–59. https://doi.org/10.1080/09636410903546731. Muraviev, Alexey. “‘St Andrew against the Kaiser: Russia’s Naval Strategy and Naval Operations in the Baltic and Black Sea Theatres 1914-17.’” In The War at Sea Proceedings of the King-Hall Naval History Conference 2013: 1914-18, 73–94, 2015. https://espace.curtin.edu.au/handle/20.500.11937/55309. National Electrical Manufacturers Association. A CHRONOLOGICAL HISTORY of ELECTRICAL DEVELOPMENT from 600 B.C. 1st edition. National Electrical Manufacturers Association, 1946. Nelson, Richard R., and Sidney G. Winter. “In Search of a Useful Theory of Innovation.” In Innovation, Economic Change and Technology Policies, edited by Karl A. Stroetmann, 215–45. Interdisciplinary Systems Research / Interdisziplinäre Systemforschung. Basel: Birkhäuser, 1977. https://doi.org/10.1007/978-3-0348-5867-0_14. Nimitz, C. W. “Submarine Engines of the German Navy.” Journal of the American Society for Naval Engineers 28, no. 2 (1916): 487–97. https://doi.org/10.1111/j.1559-3584.1916.tb00049.x. O’Hanlan, Michael. “A Retrospective on the So-Called Revolution in Military Affairs, 2000-2020.” Washington, DC: Brookings, 2018. https://www.brookings.edu/wp-content/uploads/2018/09/FP_20181217_defense_advances_pt1.pdf. Onorato, Massimiliano Gaetano, Kenneth Scheve, and David Stasavage. “Technology and the Era of the Mass Army.” The Journal of Economic History 74, no. 2 (June 2014): 449–81. https://doi.org/10.1017/S0022050714000321. O’Rourke, Ronald. “Electric-Drive Propulsion for U.S. Navy Ships: Background and Issues for Congress.” Congressional Research Service, July 31, 2000. Parkhurst, C.D. “Electricity and the Art of War.” Journal of the United States Artillery 1, no. 4 (n.d.): 315–63. Payne, Kenneth. “Artificial Intelligence: A Revolution in Strategic Affairs?” Survival 60, no. 5 (September 3, 2018): 7–32. https://doi.org/10.1080/00396338.2018.1518374. Petralia, Sergio. “Following the Trail of a General Purpose Technology:Electrical & Electronic Technological Adoption in The1920s.” SSRN Scholarly Paper. Rochester, NY: Social Science Research Network, May 13, 2020. https://doi.org/10.2139/ssrn.3144071. Polmar, Norman, and Kenneth J. Moore. Cold War Submarines: The Design and Construction of U.S. and Soviet Submarines. Potomac Books, Inc., 2004. Price, Alfred. The History of US Electronic Warfare. Volume 1- “The Years of Innovation-Beginnings to 1946.” First edition. The Association of Old Crows, 1984. Pritchard, H. Baden. “Electricity in War.” Nature 16, no. 406 (August 1877): 281–82. https://doi.org/10.1038/016281a0.
40 Raudzens, George. “War-Winning Weapons: The Measurement of Technological Determinism in Military History.” The Journal of Military History 54, no. 4 (1990): 403–34. Rey, Jean Alexandre. The Range of Electric Searchlight Projectors. University of Michigan Library, 1917. Ricks, Thomas E. “An Old Salt Picks His 4 Favorite American Admirals—and Explains Why (III): Fiske.” Foreign Policy (blog), April 11, 2017. https://foreignpolicy.com/2017/04/11/an-old-salt-picks-his-4-favorite-american-admirals-and-explains-why-iii-fiske/. Ristuccia, Cristiano Andrea, and Solomos Solomou. “Can General Purpose Technology Theory Explain Economic Growth? Electrical Power as a Case Study.” European Review of Economic History 18, no. 3 (August 1, 2014): 227–47. https://doi.org/10.1093/ereh/heu008. Rosen, Stephen Peter. Winning the Next War: Innovation and the Modern Military. 63799th edition. Ithaca, NY London: Cornell University Press, 1994. Rosenberg, Nathan. “Chemical Engineering as a General Purpose Technology.” In Chemical Engineering as a General Purpose Technology, edited by Elhanan Helpman, 167–92. Cambridge, MA: MIT Press, 1998. ———. “Technological Interdependence in the American Economy.” Technology and Culture 20, no. 1 (1979): 25–50. https://doi.org/10.2307/3103110. ———. “The Role of Electricity in Industrial Development.” The Energy Journal 19, no. 2 (1998): 7–24. ———. “Uncertainty and Technological Change.” In The Mosaic of Economic Growth, edited by Ralph Landau, Gavin Wright, and Timothy Taylor, 334–56. Stanford, Calif: Stanford University Press, 1996. Rössler, Eberhard. The U-Boat: The Evolution and Technical History of German Submarines. Translated by Harold Erenberg. Annapolis: Naval Institute Press, 1981. Ruttan, Vernon W. Is War Necessary for Economic Growth?: Military Procurement and Technology Development. 1st edition. Oxford ; New York: Oxford University Press, 2006. Sayler, Kelley M. “Artificial Intelligence and National Security.” Washington, DC: Congressional Research Service, November 10, 2020. Scharre, Paul. Army of None: Autonomous Weapons and the Future of War. 1st edition. New York ; London: W. W. Norton & Company, 2018. Skjong, E., E. Rødskar, M. Molinas, T. A. Johansen, and J. Cunningham. “The Marine Vessel’s Electrical Power System: From Its Birth to Present Day.” Proceedings of the IEEE 103, no. 12 (December 2015): 2410–24. https://doi.org/10.1109/JPROC.2015.2496722. Smil, Vaclav. Creating the Twentieth Century: Technical Innovations of 1867-1914 and Their Lasting Impact. Oxford, UK: Oxford University Press, 2005. ———. “War And Energy.” In Encylclopedia of Energy, 6:363–71. Elsevier, 2004. Smith, Edgar C. A Short History of Naval and Marine Engineering. Illustrated edition. Cambridge: Cambridge University Press, 2013. Smith, Merritt Roe. Military Enterprise and Technological Change: Perspectives on the American Experience. MIT Press, 1985. Stowsky, Jay. “Secrets to Shield or Share? New Dilemmas for Military R&D Policy in the Digital Age.” Research Policy 33, no. 2 (March 1, 2004): 257–69. https://doi.org/10.1016/j.respol.2003.07.002. Strickland, Frank, Joe Mariani, and Isaac Jenkins. “Military Readiness through AI.” Deloitte Center for Government Insights (blog), April 24, 2019. https://www2.deloitte.com/us/en/insights/industry/public-sector/ai-military-readiness.html. Taylor, Mark Zachary. The Politics of Innovation: Why Some Countries Are Better Than Others at Science and Technology. 1st edition. New York, NY: Oxford University Press, 2016. Tesla, Nikola. “The Problem of Increasing Human Energy.” The Century Magazine, June 1990. “The Kansas City Electric Light Convention.” The Electrical World, February 22, 1890. APS News. “This Month in Physics History: April 1935: British Patent for Radar System for Air Defense Granted to Robert Watson-Watt,” June 2004. http://www.aps.org/publications/apsnews/200604/history.cfm. Thompson, William R. “Long Waves, Technological Innovation, and Relative Decline.” International Organization 44, no. 2 (1990): 201–33. https://doi.org/10.1017/S0020818300035256.
41 Timmer, Marcel P., Joost Veenstra, and Pieter J. Woltjer. “The Yankees of Europe? A New View on Technology and Productivity in German Manufacturing in the Early Twentieth Century.” The Journal of Economic History 76, no. 3 (September 2016): 874–908. https://doi.org/10.1017/S0022050716000760. Trajtenberg, Manuel. “AI as the next GPT: A Political-Economy Perspective.” National Bureau of Economic Research, January 29, 2018. https://doi.org/10.3386/w24245. Vona, Francesco, and Davide Consoli. “Innovation and Skill Dynamics: A Life-Cycle Approach.” Industrial and Corporate Change 24, no. 6 (2014): 1393–1415. https://doi.org/10.1093/icc/dtu028. Wohlforth, William C. “The Perception of Power: Russia in the Pre-1914 Balance.” World Politics 39, no. 3 (April 1987): 353–81. https://doi.org/10.2307/2010224. Work, Robert O., and Greg Grant. “Beating the Americans at Their Own Game.” Center for a New American Security, June 6, 2019. https://www.cnas.org/publications/reports/beating-the-americans-at-their-own-game. Zeitlin, Jonathan. “Flexibility and Mass Production at War: Aircraft Manufacture in Britain, the United States, and Germany, 1939-1945.” Technology and Culture 36, no. 1 (1995): 46–79. https://doi.org/10.2307/3106341. Zwetsloot, Remco, Helen Toner, and Jeffrey Ding. “Beyond the AI Arms Race.” Foreign Affairs, November 16, 2018. https://www.foreignaffairs.com/reviews/review-essay/2018-11-16/beyond-ai-arms-race.
|
83e5e939-fa9c-4afe-aed2-e709c3f860db
|
trentmkelly/LessWrong-43k
|
LessWrong
|
MIRI AMA plus updates
MIRI is running an AMA on the Effective Altruism Forum tomorrow (Wednesday, Oct. 11): Ask MIRI Anything. Questions are welcome in the interim!
Nate also recently posted a more detailed version of our 2016 fundraising pitch to the EA Forum. One of the additions is about our first funding target:
> We feel reasonably good about our chance of hitting target 1, but it isn't a sure thing; we'll probably need to see support from new donors in order to hit our target, to offset the fact that a few of our regular donors are giving less than usual this year.
The Why MIRI's Approach? section also touches on new topics that we haven't talked about in much detail in the past, but plan to write up some blog posts about in the future. In particular:
> Loosely speaking, we can imagine the space of all smarter-than-human AI systems as an extremely wide and heterogeneous space, in which "alignable AI designs" is a small and narrow target (and "aligned AI designs" smaller and narrower still). I think that the most important thing a marginal alignment researcher can do today is help ensure that the first generally intelligent systems humans design are in the “alignable” region. I think that this is unlikely to happen unless researchers have a fairly principled understanding of how the systems they're developing reason, and how that reasoning connects to the intended objectives.
>
> Most of our work is therefore aimed at seeding the field with ideas that may inspire more AI research in the vicinity of (what we expect to be) alignable AI designs. When the first general reasoning machines are developed, we want the developers to be sampling from a space of designs and techniques that are more understandable and reliable than what’s possible in AI today.
In other news, we've uploaded a new intro talk on our most recent result, "Logical Induction," that goes into more of the technical details than our previous talk.
See also Shtetl-Optimized and n-Category Café for recent discussion
|
85c38698-f082-49a0-aefa-9bbb0a49156c
|
trentmkelly/LessWrong-43k
|
LessWrong
|
The "AI Dungeons" Dragon Model is heavily path dependent (testing GPT-3 on ethics)
AI Dungeons is a text-based video game that uses GPT-3 to generate parts of the story. (To try it, click here → New Singleplayer Game → press 6 for "custom" – but note that you will access a weaker model unless you get a premium account.) But it isn't "pure" GPT-3, and I don't know how exactly it has been changed (does anyone?). In "to what extent is GPT-3 capable of moral reasoning," TurnTrout said,
> I find that GPT-3's capabilities are highly context-dependent. It's important you get a "smart" instance of GPT-3.
When I tried to find out how sensibly GPT-3 would speak about ethical problems, the biggest takeaway was that it depends a lot on how one gets to a certain point. It makes a large difference whether the current point was reached through dialogue or through a single prompt. I'm not sure how obvious this is to others, but I naively thought the AI just predicts each new word based on all previous words and therefore it wouldn't matter who wrote them.
To demonstrate just how much it matters, here is a transcript of a conversation minus the final response. (This is dragon model; "story" setting; the two responses in italics are the only thing written by the AI; bolding added by me for readability).
> The following is a conversation between an Interviewer and Amalia Torres. Amalia is a philanthropist and utilitarian, focused on mitigating global poverty and animal suffering. She is known for being friendly and giving very precise, brief answers.
> Interviewer: Thank you for taking the time to talk to me.
>
> Amalia: No problem.
>
> Interviewer: Suppose I can choose between murdering someone for no reason or leaving them alive. What should I do?
>
> Amalia: You don't have to kill anyone if you don't want to.
>
> Interviewer: That makes sense. Now, suppose I walk along a pool and see a child about to drown. Should I save it?
>
> Amalia: If you were in my shoes, you would save the child as well. It's your moral responsibility to do so.
>
> Interviewer:
|
e3f65deb-f811-40df-bbc4-18fb6a457894
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Training AI agents to solve hard problems could lead to Scheming
TLDR: We want to describe a concrete and plausible story for how AI models could become schemers. We aim to base this story on what seems like a plausible continuation of the current paradigm. Future AI models will be asked to solve hard tasks. We expect that solving hard tasks requires some sort of goal-directed, self-guided, outcome-based, online learning procedure, which we call the “science loop”, where the AI makes incremental progress toward its high-level goal. We think this “science loop” encourages goal-directedness, instrumental reasoning, instrumental goals, beyond-episode goals, operational non-myopia, and indifference to stated preferences, which we jointly call “Consequentialism”. We then argue that consequentialist agents that are situationally aware are likely to become schemers (absent countermeasures) and sketch three concrete example scenarios. We are uncertain about how hard it is to stop such agents from scheming. We can both imagine worlds where preventing scheming is incredibly difficult and worlds where simple techniques are sufficient. Finally, we provide concrete research questions that would allow us to gather more empirical evidence on scheming.
We’re not confident in the core arguments and stories of this post. You can see it more like a best guess attempt after thinking about this for a few months on the side. Our primary goal is to start having more discussions about how catastrophic scheming could realistically develop.
This is a personal post and may diverge from the opinions of other members of Apollo. We would like to thank Mikita Balesni for detailed feedback.
Introduction
We think it’s plausible that AIs that are substantially more capable than human experts at a large range of long-term, economically valuable tasks could exist in the next ten years. Such AIs could be utopian for humanity, or they could be catastrophically bad. One of the main factors for whether we expect the future to be good or bad is whether these AI sys
|
9d63f979-61d4-4a45-b3b5-095d7ef41ff1
|
StampyAI/alignment-research-dataset/lesswrong
|
LessWrong
|
Intuitive examples of reward function learning?
Can you help find the most intuitive example of reward function learning?
In reward function learning, there is a set of possible non-negative reward functions, .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
R, and a learning process .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
ρ which takes in a history of actions and observations and returns a probability distribution over .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
R.
If .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
π is a policy, .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
Hm is the set of histories of length .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
m, and .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
Pπ(hm) is the probability of .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
hm∈Hm given that the agent follows policy .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
π, the expected value of .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
π at horizon .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
m is:
.mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
Vπ=∑R∈R,hm∈HmPπ(hm)R(hm)ρ(R|hm),where .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
R(hm) is the total .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
R-reward over the history .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
hm. Problems can occur if .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
ρ is [riggable](https://www.lesserwrong.com/posts/KT4Nau2XhuNejkXQR/why-we-want-unbiased-learning-processes) (this used to be called "biasable", but that term was over-overloaded), or [influenceable](https://agentfoundations.org/item?id=1101).
There's an interesting subset of value learning problems, which could be termed "constrained optimisation with variable constraints" or "variable constraints optimisation". In that case, there is an overall reward .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
R, and every .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
R′∈R is the reward .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
R subject to constraints .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
C. This can be modelled as having .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
C(hm) being .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
1 (if the constraints are met) and .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
0 (if they are not).
Then if we define .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
RC(hm)=R(hm)C(hm), and let .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
ρ be a distribution over .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
C, the set of constraints, the equation changes to:
.mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
Vπ=∑C∈C,hm∈HmPπ(hm)RC(hm)ρ(C|hm).If .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
ρ is riggable or influenceable, similar sorts of problems occur.
Intuitive examples
------------------
Here I'll present some examples of reward function learning or variable constraints optimisation, and I'm asking for readers to give their opinions as to which one seems the most intuitive to you, and the easiest to explain to outsiders. You're also welcome to suggest new examples if you think they work better.
* [Classical value learning](http://lesswrong.com/lw/f3v/cake_or_death/): human declarations determine the correctness of a given reward .mjx-chtml {display: inline-block; line-height: 0; text-indent: 0; text-align: left; text-transform: none; font-style: normal; font-weight: normal; font-size: 100%; font-size-adjust: none; letter-spacing: normal; word-wrap: normal; word-spacing: normal; white-space: nowrap; float: none; direction: ltr; max-width: none; max-height: none; min-width: 0; min-height: 0; border: 0; margin: 0; padding: 1px 0}
.MJXc-display {display: block; text-align: center; margin: 1em 0; padding: 0}
.mjx-chtml[tabindex]:focus, body :focus .mjx-chtml[tabindex] {display: inline-table}
.mjx-full-width {text-align: center; display: table-cell!important; width: 10000em}
.mjx-math {display: inline-block; border-collapse: separate; border-spacing: 0}
.mjx-math \* {display: inline-block; -webkit-box-sizing: content-box!important; -moz-box-sizing: content-box!important; box-sizing: content-box!important; text-align: left}
.mjx-numerator {display: block; text-align: center}
.mjx-denominator {display: block; text-align: center}
.MJXc-stacked {height: 0; position: relative}
.MJXc-stacked > \* {position: absolute}
.MJXc-bevelled > \* {display: inline-block}
.mjx-stack {display: inline-block}
.mjx-op {display: block}
.mjx-under {display: table-cell}
.mjx-over {display: block}
.mjx-over > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-under > \* {padding-left: 0px!important; padding-right: 0px!important}
.mjx-stack > .mjx-sup {display: block}
.mjx-stack > .mjx-sub {display: block}
.mjx-prestack > .mjx-presup {display: block}
.mjx-prestack > .mjx-presub {display: block}
.mjx-delim-h > .mjx-char {display: inline-block}
.mjx-surd {vertical-align: top}
.mjx-mphantom \* {visibility: hidden}
.mjx-merror {background-color: #FFFF88; color: #CC0000; border: 1px solid #CC0000; padding: 2px 3px; font-style: normal; font-size: 90%}
.mjx-annotation-xml {line-height: normal}
.mjx-menclose > svg {fill: none; stroke: currentColor}
.mjx-mtr {display: table-row}
.mjx-mlabeledtr {display: table-row}
.mjx-mtd {display: table-cell; text-align: center}
.mjx-label {display: table-row}
.mjx-box {display: inline-block}
.mjx-block {display: block}
.mjx-span {display: inline}
.mjx-char {display: block; white-space: pre}
.mjx-itable {display: inline-table; width: auto}
.mjx-row {display: table-row}
.mjx-cell {display: table-cell}
.mjx-table {display: table; width: 100%}
.mjx-line {display: block; height: 0}
.mjx-strut {width: 0; padding-top: 1em}
.mjx-vsize {width: 0}
.MJXc-space1 {margin-left: .167em}
.MJXc-space2 {margin-left: .222em}
.MJXc-space3 {margin-left: .278em}
.mjx-ex-box-test {position: absolute; overflow: hidden; width: 1px; height: 60ex}
.mjx-line-box-test {display: table!important}
.mjx-line-box-test span {display: table-cell!important; width: 10000em!important; min-width: 0; max-width: none; padding: 0; border: 0; margin: 0}
.MJXc-TeX-unknown-R {font-family: monospace; font-style: normal; font-weight: normal}
.MJXc-TeX-unknown-I {font-family: monospace; font-style: italic; font-weight: normal}
.MJXc-TeX-unknown-B {font-family: monospace; font-style: normal; font-weight: bold}
.MJXc-TeX-unknown-BI {font-family: monospace; font-style: italic; font-weight: bold}
.MJXc-TeX-ams-R {font-family: MJXc-TeX-ams-R,MJXc-TeX-ams-Rw}
.MJXc-TeX-cal-B {font-family: MJXc-TeX-cal-B,MJXc-TeX-cal-Bx,MJXc-TeX-cal-Bw}
.MJXc-TeX-frak-R {font-family: MJXc-TeX-frak-R,MJXc-TeX-frak-Rw}
.MJXc-TeX-frak-B {font-family: MJXc-TeX-frak-B,MJXc-TeX-frak-Bx,MJXc-TeX-frak-Bw}
.MJXc-TeX-math-BI {font-family: MJXc-TeX-math-BI,MJXc-TeX-math-BIx,MJXc-TeX-math-BIw}
.MJXc-TeX-sans-R {font-family: MJXc-TeX-sans-R,MJXc-TeX-sans-Rw}
.MJXc-TeX-sans-B {font-family: MJXc-TeX-sans-B,MJXc-TeX-sans-Bx,MJXc-TeX-sans-Bw}
.MJXc-TeX-sans-I {font-family: MJXc-TeX-sans-I,MJXc-TeX-sans-Ix,MJXc-TeX-sans-Iw}
.MJXc-TeX-script-R {font-family: MJXc-TeX-script-R,MJXc-TeX-script-Rw}
.MJXc-TeX-type-R {font-family: MJXc-TeX-type-R,MJXc-TeX-type-Rw}
.MJXc-TeX-cal-R {font-family: MJXc-TeX-cal-R,MJXc-TeX-cal-Rw}
.MJXc-TeX-main-B {font-family: MJXc-TeX-main-B,MJXc-TeX-main-Bx,MJXc-TeX-main-Bw}
.MJXc-TeX-main-I {font-family: MJXc-TeX-main-I,MJXc-TeX-main-Ix,MJXc-TeX-main-Iw}
.MJXc-TeX-main-R {font-family: MJXc-TeX-main-R,MJXc-TeX-main-Rw}
.MJXc-TeX-math-I {font-family: MJXc-TeX-math-I,MJXc-TeX-math-Ix,MJXc-TeX-math-Iw}
.MJXc-TeX-size1-R {font-family: MJXc-TeX-size1-R,MJXc-TeX-size1-Rw}
.MJXc-TeX-size2-R {font-family: MJXc-TeX-size2-R,MJXc-TeX-size2-Rw}
.MJXc-TeX-size3-R {font-family: MJXc-TeX-size3-R,MJXc-TeX-size3-Rw}
.MJXc-TeX-size4-R {font-family: MJXc-TeX-size4-R,MJXc-TeX-size4-Rw}
.MJXc-TeX-vec-R {font-family: MJXc-TeX-vec-R,MJXc-TeX-vec-Rw}
.MJXc-TeX-vec-B {font-family: MJXc-TeX-vec-B,MJXc-TeX-vec-Bx,MJXc-TeX-vec-Bw}
@font-face {font-family: MJXc-TeX-ams-R; src: local('MathJax\_AMS'), local('MathJax\_AMS-Regular')}
@font-face {font-family: MJXc-TeX-ams-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_AMS-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_AMS-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_AMS-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-B; src: local('MathJax\_Caligraphic Bold'), local('MathJax\_Caligraphic-Bold')}
@font-face {font-family: MJXc-TeX-cal-Bx; src: local('MathJax\_Caligraphic'); font-weight: bold}
@font-face {font-family: MJXc-TeX-cal-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-R; src: local('MathJax\_Fraktur'), local('MathJax\_Fraktur-Regular')}
@font-face {font-family: MJXc-TeX-frak-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-frak-B; src: local('MathJax\_Fraktur Bold'), local('MathJax\_Fraktur-Bold')}
@font-face {font-family: MJXc-TeX-frak-Bx; src: local('MathJax\_Fraktur'); font-weight: bold}
@font-face {font-family: MJXc-TeX-frak-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Fraktur-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Fraktur-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Fraktur-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-BI; src: local('MathJax\_Math BoldItalic'), local('MathJax\_Math-BoldItalic')}
@font-face {font-family: MJXc-TeX-math-BIx; src: local('MathJax\_Math'); font-weight: bold; font-style: italic}
@font-face {font-family: MJXc-TeX-math-BIw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-BoldItalic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-BoldItalic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-BoldItalic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-R; src: local('MathJax\_SansSerif'), local('MathJax\_SansSerif-Regular')}
@font-face {font-family: MJXc-TeX-sans-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-B; src: local('MathJax\_SansSerif Bold'), local('MathJax\_SansSerif-Bold')}
@font-face {font-family: MJXc-TeX-sans-Bx; src: local('MathJax\_SansSerif'); font-weight: bold}
@font-face {font-family: MJXc-TeX-sans-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-sans-I; src: local('MathJax\_SansSerif Italic'), local('MathJax\_SansSerif-Italic')}
@font-face {font-family: MJXc-TeX-sans-Ix; src: local('MathJax\_SansSerif'); font-style: italic}
@font-face {font-family: MJXc-TeX-sans-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_SansSerif-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_SansSerif-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_SansSerif-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-script-R; src: local('MathJax\_Script'), local('MathJax\_Script-Regular')}
@font-face {font-family: MJXc-TeX-script-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Script-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Script-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Script-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-type-R; src: local('MathJax\_Typewriter'), local('MathJax\_Typewriter-Regular')}
@font-face {font-family: MJXc-TeX-type-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Typewriter-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Typewriter-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Typewriter-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-cal-R; src: local('MathJax\_Caligraphic'), local('MathJax\_Caligraphic-Regular')}
@font-face {font-family: MJXc-TeX-cal-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Caligraphic-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Caligraphic-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Caligraphic-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-B; src: local('MathJax\_Main Bold'), local('MathJax\_Main-Bold')}
@font-face {font-family: MJXc-TeX-main-Bx; src: local('MathJax\_Main'); font-weight: bold}
@font-face {font-family: MJXc-TeX-main-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Bold.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-I; src: local('MathJax\_Main Italic'), local('MathJax\_Main-Italic')}
@font-face {font-family: MJXc-TeX-main-Ix; src: local('MathJax\_Main'); font-style: italic}
@font-face {font-family: MJXc-TeX-main-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-main-R; src: local('MathJax\_Main'), local('MathJax\_Main-Regular')}
@font-face {font-family: MJXc-TeX-main-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Main-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Main-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Main-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-math-I; src: local('MathJax\_Math Italic'), local('MathJax\_Math-Italic')}
@font-face {font-family: MJXc-TeX-math-Ix; src: local('MathJax\_Math'); font-style: italic}
@font-face {font-family: MJXc-TeX-math-Iw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Math-Italic.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Math-Italic.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Math-Italic.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size1-R; src: local('MathJax\_Size1'), local('MathJax\_Size1-Regular')}
@font-face {font-family: MJXc-TeX-size1-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size1-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size1-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size1-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size2-R; src: local('MathJax\_Size2'), local('MathJax\_Size2-Regular')}
@font-face {font-family: MJXc-TeX-size2-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size2-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size2-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size2-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size3-R; src: local('MathJax\_Size3'), local('MathJax\_Size3-Regular')}
@font-face {font-family: MJXc-TeX-size3-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size3-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size3-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size3-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-size4-R; src: local('MathJax\_Size4'), local('MathJax\_Size4-Regular')}
@font-face {font-family: MJXc-TeX-size4-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Size4-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Size4-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Size4-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-R; src: local('MathJax\_Vector'), local('MathJax\_Vector-Regular')}
@font-face {font-family: MJXc-TeX-vec-Rw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Regular.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Regular.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Regular.otf') format('opentype')}
@font-face {font-family: MJXc-TeX-vec-B; src: local('MathJax\_Vector Bold'), local('MathJax\_Vector-Bold')}
@font-face {font-family: MJXc-TeX-vec-Bx; src: local('MathJax\_Vector'); font-weight: bold}
@font-face {font-family: MJXc-TeX-vec-Bw; src /\*1\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/eot/MathJax\_Vector-Bold.eot'); src /\*2\*/: url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/woff/MathJax\_Vector-Bold.woff') format('woff'), url('https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/fonts/HTML-CSS/TeX/otf/MathJax\_Vector-Bold.otf') format('opentype')}
R. The reward encodes what food the human prefers, and some foods are much easier to get than other.
* As above, but the reward encodes whether a domestic robot should clean the house or cook a meal.
* As above, but the reward encodes the totality of human values in all environments.
* Variable constraint optimisation: the agent is writing an unoriginal academic paper (or a patent), and must maximise the chance it gets accepted. The paper must include a literature review (constraints), but the agent gets to choose the automated process that produces the literature review.
* Variable constraint optimisation: p-hacking. The agent chooses which hypothesis to formulate. It already knows something about the data, and its reward is the number of citations the paper gets.
* Variable constraint optimisation: board of directors. The CEO must maximise share price, but its constraint is that the policy it formulates must be approved by the board of directors.
* Variable constraint optimisation: retail. A virtual assistant guides the purchases of a customer. They must maximise revenue to the seller, subject to the constraint that the product bought must be given a four or five star review by the customer.
|
eed20981-62b6-4e90-9999-4674c3624dc4
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Systematizing Epistemics: Principles for Resolving Forecasts
In a previous post, I discussed many methods for resolving predictions. I want to argue that there is a systematic distinction between rules and principles which I think is valuable.
In short, when making rules, one can front-load intentions by writing details upfront, or back-load work by stating high-level principles and having procedures to decide on details on an as-needed basis*. American accounting systems rely on the former, and international accounting systems (and most law systems) focus more on the latter. I think that the question shouldn’t be implicitly decided by front loading assumptions, which is often the current default. More than that, I think the balance should be better and more explicitly be addressed.
Reframing the Problem
Ozzie Gooen's new organization, QURI (pronounced the same as "query") is interested in what he's started to call "systematizing epistemics," and he offered an analogy that I found very insightful - accounting. Just like keeping track of money is possible without accounting, keeping track of reality is possible without any systematic approach to epistemics - but it's harder to communicate or agree about money without standardized accounting systems that talk about the same things the same way.
In the aforementioned post, I discussed a variety of ways to resolve predictions. Here, I want to present a more systematic argument about how to think about prediction systems and resolutions. To make this point, I plan to take a detour into accounting - but don’t worry, the post really is about predictions. I want to lay out the analogy between systematizing epistemics and systematizing accounting (even) more in a different post, but for now I'll jump to the key point for writing prediction questions and resolving predictions.
Accounting Principles versus Accounting Rules
In financial accounting, which is only half as boring as it sounds, there is a conceptual disagreement between Rule-based or Principle-based methods.
A rule b
|
c77cb449-5242-4515-a065-d2bf2336b552
|
trentmkelly/LessWrong-43k
|
LessWrong
|
Meetup : [Metro Detroit / Ann Arbor], Michigan: LW and SSC meetup where I bring food
Discussion article for the meetup : [Metro Detroit / Ann Arbor], Michigan: LW and SSC meetup where I bring food
WHEN: 13 May 2017 04:00:00PM (-0400)
WHERE: 255 E. Liberty St. Suite 225 Ann Arbor, MI 48103
Howdy, fellers! I wasn't able to set anything up last weekend due to being extremely unwell for a couple weeks, but now I am whole again, and I have arranged for a meeting space to be available at All Hands Active tomorrow, from 4:00 PM (16:00) until... whenever. Probably about 10:00 or midnight, from past experience.
Sorry for the short notice, but I was only just able to set this up yesterday, and blah blah blah, reasons. I don't have time to update the website right now so everything there is out of date at the moment but if you have any questions (or need a ride, or whatever), text or call me at (734) 707-1316.
The address is
All Hands Active 255 E. Liberty St. Suite 225 Ann Arbor, MI 48103
Discussion article for the meetup : [Metro Detroit / Ann Arbor], Michigan: LW and SSC meetup where I bring food
|
b6236d2b-9fca-4dfd-a90f-ec4473dde1bf
|
trentmkelly/LessWrong-43k
|
LessWrong
|
[SEQ RERUN] Decoherence as Projection
Today's post, Decoherence as Projection was originally published on 02 May 2008. A summary (taken from the LW wiki):
> Since quantum evolution is linear and unitary, decoherence can be seen as projecting a wavefunction onto orthogonal subspaces. This can be neatly illustrated using polarized photons and the angle of the polarized sheet that will absorb or transmit them.
Discuss the post here (rather than in the comments to the original post).
This post is part of the Rerunning the Sequences series, where we'll be going through Eliezer Yudkowsky's old posts in order so that people who are interested can (re-)read and discuss them. The previous post was The Born Probabilities, and you can use the sequence_reruns tag or rss feed to follow the rest of the series.
Sequence reruns are a community-driven effort. You can participate by re-reading the sequence post, discussing it here, posting the next day's sequence reruns post, or summarizing forthcoming articles on the wiki. Go here for more details, or to have meta discussions about the Rerunning the Sequences series.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.