text
stringlengths 1
2.31M
| meta
dict |
|---|---|
Q:
Why can we use matrices to represent elements of polynomial vector space?
I think I understand the concept of the polynomial vector space over a field F (set of polynomial functions whose coefficients are part of F), but when describing a transformation from one polynomial vector space to another —- why are we allowed to use matrices to describe that? Matrices aren’t elements of the polynomial vector space, but it is its own vector space right? For example, when describing the transformation matrix for the linear transformation that is the derivative of a polynomial, we have a matrix of coefficients and then multiply that by whatever input column coefficient vector we have that represents a polynomial, but that column vector itself is not part of the polynomial vector space so how do we know it describes a polynomial in the first place?
A:
Consider the vector space $P_{n - 1}(k)$ of polynomials of degree $\leq n - 1$ with coefficients in the field $k.$ Observe that the standard basis of $P_{n - 1}(k)$ is $\mathscr B = \{1, x, x^2, \dots, x^{n - 1}\}.$ (Use the derivative to prove that these are linearly independent.) Consider a linear operator $T : P_{n - 1}(k) \to P_{n - 1}(k).$ Given any polynomial $p(x)$ in $P_{n - 1}(k),$ we can write $T(p(x))$ in terms of the basis $\mathscr B.$ Particularly, $$\begin{align*} T(1) &= a_{11} + a_{21} x + \cdots + a_{n1} x^{n - 1}, \\ \\
T(x) &= a_{12} + a_{22} x + \cdots + a_{n2} x^{n-1}, \\ &\phantom{\,\,} \vdots \\ T(x^{n-1}) &= a_{1n} + a_{2n} x + \cdots + a_{nn} x^{n - 1}, \end{align*}$$ from which we obtain an $n \times n$ matrix $A = [a_{ij}]_{1 \leq i, j \leq n}$ with coefficients in $k.$ (Explicitly, $A$ is the matrix whose $j$th column is $T(x^{j - 1}).$) We refer to $A$ as the matrix of $T$ with respect to $\mathscr B.$
Example. Consider the vector space $P_2(\mathbb R)$ of polynomials of degree $\leq 2$ with coefficients in $\mathbb R$ (the real numbers). Recall that the derivative satisfies $\frac d {dx} [C \cdot p(x)] = C \cdot p'(x)$ for all real numbers $C$ and $\frac d {dx} [p(x) + q(x)] = p'(x) + q'(x)$ for all polynomials $p(x)$ and $q(x),$ hence differentiation is a linear transformation. By the Power Rule, we have that $\deg p'(x) = \deg p(x) - 1,$ hence the map $D : P_2(\mathbb R) \to P_2(\mathbb R)$ defined by $D(p(x)) = p'(x)$ is a linear operator. Considering that $$\begin{align*} D(1) &= 0, \\ \\ D(x) &= 1, \text{ and} \\ \\ D(x^2) &= 2x, \end{align*}$$ we find that the matrix of $D$ with respect to $\mathscr B = \{1, x, x^2\}$ is given by $A = \begin{pmatrix} 0 & 1 & 0 \\ 0 & 0 & 2 \\ 0 & 0 & 0 \end{pmatrix}.$
|
{
"pile_set_name": "StackExchange"
}
|
Blackboard needs to apply a security patch that will cause Biola's Blackboard to be unavailable for about an hour sometime between 9:01 p.m. and 3:00 a.m. on Saturday night / Sunday morning.
Here's the note from Blackboard:
Managed Hosting Engineers will perform a system maintenance on your hosted site between the hours of 12:01 am to 06:00 am US Eastern on Sunday, March 31st, 2013.
Please take note that this is a service impacting maintenance occurring on a Saturday when we deploy the a Security Patch for Blackboard Learn Release 9.1 Service Pack 8. It comprises of critical product patches which enhance the performance and stability of your hosted environment.
Managed Hosting Engineers will perform system maintenance on your hosted site between the hours of 12:01 am to 06:00 am US Eastern on Sunday, Feb 24th, 2013.
Please take note that this is a service impacting maintenance occurring on a Sunday when we deploy a Security Patch. It comprises of critical product patches which enhance the security and stability of your hosted environment.
Support Bulletin you received recently detailing the Patch Set can be found at:
Blackboard needs to restart our servers. Blackboard will be unavailable for "less than one hour" (probably more like 10 minutes) sometime between 11:00 p.m. Friday, July 6, 2012 and 3:00 a.m. Saturday, July 7, 2012. Here's the official notice:
Date: Saturday, July 7th, 2012
Time: 02:00 am to 06:00 am US Eastern
Work Type: Building Block Upgrade and Server Reboot of Learn Production Environment
Services: Clients at Northern Virginia Data Center VA2 which are on Blackboard Learn Versions 8.x and 9.x Only
On Saturday, July 7th, 2012 between the hours of 02:00 am to 06:00 am US Eastern, Managed Hosting Engineers will be doing a service restart of your hosted Learn Blackboard site.
Building Block upgrade for Safe Assign has a mandatory after action which requires the restart of Blackboard systems running Learn Versions 8.x and 9.x. The Support Bulletin detailing that work can be found at
When an instructor creates a new announcement on a Blackboard course (by going to Announcements then clicking "Create Announcement"), students will see that announcement next time they log in to Blackboard. When you create an announcement you have the option to check a box that says: "Send a copy of this announcement immediately."
However, there is currently a bug in the Blackboard system which prevents that from happening. If read carefully, when you submit the announcement, you'll see this notification:
Notice that although it says "Success," it also informs you that the email was not sent.
The work-around for this problem: is just to create the announcement then also copy & paste it into an email. After you create the announcement, send an email by going to Control Panel / Course Tools / Send Email.
More details: this bug in Blackboard is caused by the colon character (":") appearing in the Course Name (probably because a colon is used in the subject line of the email). So this bug only happens to those courses. Unfortunately, every class at Biola has a colon in the name (like "Spring 2012: Intro to Psychology"). This is Blackboard Known Issue LRN-33329, AS-163487 and LRN-43623, and will be fixed in the next Service Pack (hopefully to be installed at Biola in the next few weeks.
For the Spring 2012 semester, we are trying to keep your Blackboard rosters closely matched with the official Registrar's roster, which you can find on my.Biola. The list on my.Biola is the only official roster.
When a student adds a class, typically that student will appear on Blackboard the next business morning.
When a student drops a class, it may take a few days before that is reflected on Blackboard. During the first week of the semester, we completely deleted the student from the roster on Blackboard. At this point, however, we will now just mark that dropped student as "unavailable" on Blackboard. This means the student cannot access that class and they will not get emails sent to the class, but it preserves any work or grades for that student may have completed on Blackboard.
Instructors can edit their Blackboard roster by following these instructions: Managing Users. Typically, instructors don't need to edit the rosters. Here are some reasons an instructor may want to:
Give a TA or department secretary access to the class
Give a student immediate access (rather than waiting for the automatic process)
Delete a student who dropped (if there are no grades or past work you need to preserve)
Blackboard sent us the following notice, describing their preparations for Hurricane Irene. The servers that run Biola's Blackboard system are in Virginia.
Subject: Emergency Preparedness for Hurricane Irene
Dear Managed Hosting Customers,
Please be advised that the Managed Hosting Northern Virginia datacenter team, in working closely with our datacenter vendors, have taken additional precautionary measures to safe guard against any impact Hurricane Irene may bring to this region. The National Oceanic and Atmospheric Administration (NOAA) earlier in the week advised that Hurricane Irene is projected to impact much of the East Coast of the United States this weekend.
In preparation of the anticipated severe weather, the following actions have been taken,
1. Datacenter generators have been tested, fuel levels verified, and fuel supplier is on notice in the event of extended power interruption.
2. Roof drains and scuppers have been cleared of debris.
3. All equipment on the roof and equipment yards has been secured to prevent potential damage by wind.
4. Dumpsters and open top containers will be emptied prior to anticipated severe weather.
5. Emergency supplies for potential roof leaks have been placed in strategic locations in each building.
6. All vendors are on notice and have confirmed for emergency deliveries/service. These include from the following groups at a minimum; Generators, UPS System, Chillers, Fuel System Delivery and Switchgear.
Special arrangements have been made to make sure DC / Virginia staff are available on site as necessary to maintain operability standards plus our teams in The Netherlands, Australia and India are on standby.
We will continue to monitor all of our sites and client systems, and take any emergency actions as weather conditions require.
We have added some classes to Blackboard for the Fall 2011 semester. However, we only add the classes specifically requested by each individual professor. So the "My Courses" list here on Blackboard does not contain official enrollment data, only my.Biola does.
Students: This is not your official schedule. Not all professors use Blackboard, so you may be missing some classes in your "My Courses" list. You may also have extra classes in this list: if you drop a class, it may take a while before it disappears from your list of "My Courses." When you add a class, it takes one business day before it will appear on Blackboard (or your professor can add you immediately). Do not contact the Registrar's Office about your class list on "My Biola." Use my.Biola to see your official class schedule.
Instructors: If you don't see your Fall 2011 class(es) here, you need to request them. See these instructions.
|
{
"pile_set_name": "Pile-CC"
}
|
I don't mind being called an Obamabot. I mean, I've written a few columns about the guy that were brutal, toughing than anything Dowd's written, especially at the time of the debt ceiling fiasco. But I understand the game, and it doesn't bother me.
I have something I wish to make crystal clear, however. If it seems to you (I mean you, pumpkinface!) that I'm always excusing Obama, you're misreading me. I am instead seeking to cast blame where it properly belongs. And that is almost always the Republican Party. I've said all this a jillion times before, but it is simply not a mainstream political party in the traditional American sense. It is a radical oppositionalist faction, way beyond the normal American parameters both in terms of ideology and tactics. And that needs to be pointed out, unfortunately, again and again and again.
Just today, Pat Toomey said of the background-check bill:
"In the end it didn’t pass because we're so politicized. There were some on my side who did not want to be seen helping the president do something he wanted to get done, just because the president wanted to do it."
A helpful admission on his part, and a rare piece of Republican candor. But this is the case time after time after time. It's not normal. It's not--and I mean not remotely--"the same thing" the Democrats did under Bush. Today's GOP is a complete historical outlier.
Yes, I'm sure there were many Democrats who didn't want to hand Reagan or either Bush a political victory. But historically, that is one of a handful of legislative considerations, and not even the first. Probably more like the fourth, after votes and money and what's right for the country. But today's GOP has turned it into iron law. It is relentlessly destructive.
On the subject of Gitmo, which I wrote about yesterday: In normal America, when a presidential candidate says he wants to do X once in office and then wins the election by a significant margin, Congress usually does X. The opposition party always attaches strings and conditions and so forth, but they obey the will of the people. Democrats, enough of them, led by Tip O'Neill, put Reagan's programs through. Same thing with Bush's tax cuts. (Republicans did not grant Clinton the same courtesy, but as bad as they were then, they're worse now.)
So in normal America, a deal would have been worked out whereby Gitmo would close. After all, remember, the Republican candidate in 2008 supported closing Gitmo too. It was the GOP's position! And yet, once Obama as president wanted to do it, they killed it cold in 2009.
They have been blocking it ever since. Here's a vote on the question of use of funds to transfer Gitmo detainees from last November, after Obama had been handily reelected. Every Republican present voted no. Every one.
That was on an amendment to to the defense reauthorization. That passed, and Obama signed it. But he issued a statement to accompany the signing explaining that he was dead-set against the provisions I referred to in this morning's post. Under the Constitution, of course, there is no line-item veto; a president either signs or vetoes an entire bill. This was a defense authorization, so he signed. But he made his position crystal clear. Here's the letter for you to see.
I'm sure there's more he could have done or could now be doing. But wouldn't you get a little discouraged? Oh, fucking hell, he thinks to himself at 3 am. Yes, I want to keep this promise I made. But why should I bang my head against that particular wall again? If I'm for it, they're against it. I won't get one Republican vote.
He is, obviously, a flawed human being; aloof, a little superior, not especially warm (so it seems), and no, he doesn't scare anybody. He has all of these flaws and more. Maybe a different human being could get Susan Collins or Rob Portman or Lamar Alexander to vote his way once in a while.
But I don't really think so. Collins and Portman and Alexander and others are, I'm certain, a little ashamed of their party today, and of themselves. But they are afraid of the right-wing agitprop media and their hard-shell base (and of course the threat of a primary from the right). So they don't have the guts to the right thing, and they likely never will.
So it's not that I'm always straining to defend Obama, although I can understand how it ends up looking that way. I am trying to tell as many people as I can that this Republican Party is extreme and wholly against American norms and traditions. And I think any opinion writer who isn't saying this over and over is, in ascending order of likelihood, lying, dense, or deceiving him or herself.
|
{
"pile_set_name": "OpenWebText2"
}
|
Cotton is an important fiber crop in many areas of the world. The methods of biotechnology have been applied to cotton for improvement of the agronomic traits and the quality of the product. The method of introducing transgenes into cotton plants has been demonstrated in U.S. Pat. No. 5,004,863. One such agronomic trait important in cotton production is herbicide tolerance, in particular, tolerance to glyphosate herbicide. This trait has been introduced into cotton plants and is a successful product now used in cotton production. The current commercial Roundup Ready® cotton event (1445) provides excellent tolerance to glyphosate, the active ingredient in Roundup®, through the four-leaf stage (Nida et al., J. Agric. Food Chem. 44:1960-1966, 1996; Nida et al., J. Agric. Food Chem. 44:1967-1974, 1996). However, foliar application beyond the four-leaf stage must be limited due to insufficient tolerance in male reproductive tissues in certain environmental conditions. This lack of male reproductive tolerance appears to be a result of insufficient CP4 EPSPS expression in critical tissues, higher sensitivity of these tissues to glyphosate, and accumulation of high amounts of glyphosate in these strong sink tissues (Pline et al., Weed Sci. 50:438-447, 2002). There is a need for a cotton plant more highly glyphosate tolerant than Roundup Ready® cotton 1445.
It would be advantageous to be able to detect the presence of a particular event in order to determine whether the progeny of a sexual cross contain a transgene of interest. In addition, a method for detecting a particular event would be helpful for complying with regulations requiring pre-market approval or labeling of foods derived from recombinant crop plants, for example. It is possible to detect the presence of a transgene by any well known nucleic acid detection method such as the polymerase chain reaction (PCR) or DNA hybridization using nucleic acid probes. These detection methods generally focus on frequently used genetic elements, such as promoters, 3′ transcription terminators, marker genes, etc. As a result, such methods may not be useful for discriminating between different events, particularly those produced using the same DNA construct unless the sequence of genomic chromosomal DNA adjacent to the inserted DNA (“flanking genomic DNA”) is known. Event-specific DNA detection methods for a glyphosate tolerant cotton event 1445 have been described (US 20020120964, herein incorporated by reference in its entirety).
The present invention relates to a glyphosate tolerant cotton event MON 88913, compositions contained therein, and to the method for the detection of the transgene/genomic insertion region in cotton event MON 88913 and progeny thereof.
|
{
"pile_set_name": "USPTO Backgrounds"
}
|
In 2009 I was experimenting with RFID tags and was also getting into steampunk. The RFID Enabled Globe was my first real steampunk artifact and was a labor of love. In July of 2010 I was able to demonstrate it to many people at the Inaugural Maker Faire in Detroit Michigan.
|
{
"pile_set_name": "Pile-CC"
}
|
Over 3 million people trust Angie's List to help make the right choice
Be informed to avoid costly mistakes
Shop with us to ensure a fair price
Our complaint resolution team will help if a project goes bad
+See Verified Local Reviews
Over 1,408 reviews forHouston Storm Shutter Installers from people just like you.
A
"I am very pleased with the service provided.
and his crew were prompt, courteous and professional. I did get two other bids - the reason" ...More I went with
was his honesty. He did not try to sell me something I did not need.
-Sandra K.
A
"I can't say enough good things about this company from sales to service. Their installation team is the hardest working group I have ever had doing work on" ...More my house. This is a 3 story house so it was not an easy job. They have an attention to detail and quality work that I have not seen in a long time. The shutters look great and work very well! Thanks
No. 1: fire stations make good neighbors. If you’re more than five miles from a fire station or more than 1,000 feet from a fire hydrant, you’ll pay several hundred dollars more per year for your policy.
Hurricane Shutter Installation And Repair reviews in Houston
A
Rating
“
I have an older brick home with windows that have decorative trim and decorative wooden shutters. Because of the trim and the shutters, I did not like the appearance of rolling shutters on my older home, which has very traditional architecture. Retrofitting this older home became a challenge because I didn't want to remove the existing trim on ...More the windows. In addition, the windows on the second floor are almost impossible to board-up during hurricane season because the windows are so wide and tall that plywood boards are heavy and unwieldy and difficult to handle. The windows on the second floor do not have shutters or trim. Those windows -- and other large windows -- became the focus of my inquiries into storm protection, and I'm very glad that I took the time and made the effort to find the best product.
I began a long
and eventually met
at HHS. He showed me the many different kinds of products available and suggested that I might like Tapco screens.
I LOVE the Tapco screens. They look very much like regular screens that one sees on windows, but the strength of the wire makes the screens able to withstand storm winds. It isn't dark inside the house; in fact, there is little difference in the amount of light that comes into the home. Since the windows face west, I do think that we are getting some sun protection, as well.
The screens look really good, and my neighbors like them, too -- a very important point because I didn't want to install anything unsightly or unattractive.
We had many challenges retrofitting this older home, and
was very conscientious about service and installation. I am very pleased with the product and the service I received.
One suggestion that I would make to anyone wanting hurricane protection (shutters) would be to start as early in the season as possible so that your order doesn't have to be placed at the peak of the season. If you place your order at the peak of the season in a city as large as
, Texas, you may find that delivery of your product is delayed due to a supplier's fabrication schedules elsewhere.”
- K T.
A
Rating
“
I went through three companies and estimates before I decided on
. They gave the lowest quotes and was very punctual.
explained the process and how long it takes to get it ordered from the manufacturer. He was also punctual with his and was very flexible in what time they want to come and install it.”
- SAJ T.
A
Rating
“
My home has several large, fixed, low r windows that also have frosting on some. My goal was to have some window protection that, in case I was away, could be installed by my wife. The large window are big enough that it would be work for two to hold up and nail plywood.
After an Angies List and web
I contacted several ...More providers of metal screens. Two did not even return calls, one sent an email telling me the square foot cost and stating they would send out someone to estimate if i wished.
King of
didn't mess around. He and I worked out a date for his estimate and stayed in contact until it was done. I had expressed my concerns about staying within a budget and he worked with me on that, showing me some smaller window that were somewhat sheltered by fences and shrubs. Also, there was one window that i did want alumiminum blinds on but
Knig showed me why it would not be recomended at that window becasue on the nature of the brick facade. He could have sold it to me anyway and I would not been any wiser.( I took that oppertunity to go ahead and cut plywood to fit but made it in sections so it would be lighter)
His price of $1,445 was within my budget and I told him to go ahead. Covered was three 6ft x 5ft picture windows, three 3ft x 5ft and one french door. The blinds are corrogated aluminum that span the windows or door and each section is some 3 ft wide. They are secured on studs that are mounted on the facings. I had planned on uploading some photos but thought I had better finish this review.
When the blinds were ready,
and his assistant, I regret his name has slipped my mind, installed the mounts in about
2 1\2 hours. He then demonstrated installation to my wife and I, left me an ample supply of wing nuts and best of all, a map of the windows labeled to match the labels on the blinds.
I have used many contractors both at home and in my business and I could not have asked for better service.”
- Carl A.
All Storm Shutter Installers in Houston
Companies below are listed in alphabetical order. To view top rated service providers along with reviews and ratings, Join Angie's List Now!
|
{
"pile_set_name": "Pile-CC"
}
|
Boysie Oakes
Boysie Oakes is fictional secret agent created by the British spy novelist John Gardner in 1964 at the height of a period of fictional spy mania.
Character Biography
Oakes is a richly comic character who is inadvertently taken to be a tough, pitiless man of action and is recruited into a British spy agency. He is, in actuality, a devout coward with many other character flaws who wants nothing more than to be left alone.
An example of the endearing points in these novels is the continuing appearance of monogrammed "BO" gifts from his mother - shirts, handkerchiefs, cigarette lighter, and so forth. He would never dream of not using them, even though he resents them.
The cowardly Oakes starred in another seven novels over the next 15 years and eventually, once again by inadvertence, becomes the head of the secret agency that has caused him to be in a constant state of terror for so long.
Film
The first novel in the series, The Liquidator, was made into a feature film of the same name in 1965, starring Rod Taylor as Boysie Oakes.
Boysie Oakes novels
The Liquidator (1964)
Understrike (1965)
Amber Nine (1966)
Madrigal (1967)
Founder Member (1969)
Traitor's Exit (1970)
The Airline Pirates (1970) - published in the U.S. as Air Apparent
A Killer for a Song (1975)
Two Boysie Oakes short stories appear in The Assassination File (1974): A Handful of Rice, and Corkscrew.
Two Boysie Oakes short stories appear in Hideaway (1968): Boysie Oakes and The Explosive Device, Sunset At Paleokastritsa.
External links
Oakes, Boysie
Category:Series of books
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Posts tagged 'college'
At first I thought I would give this higher education fair a miss but as you can see, I changed my mind.
Karnival Jom Masuk U is actually part of the Karnival Pengajian Tinggi Negara 2011 kali ke-7 (7th National Higher Education Carnival) roadshow that kicked off this January in Terengganu. There are a total ... continue reading »
|
{
"pile_set_name": "Pile-CC"
}
|
//
// main.m
// MenuBarApp
//
// Created by Charles Ross on 11/2/13.
// Copyright (c) 2013 Chivalry Software. All rights reserved.
//
int main(int argc, const char *argv[]) {
return NSApplicationMain(argc, argv);
}
|
{
"pile_set_name": "Github"
}
|
---
abstract: 'We discuss and test possible evolutionary connections between Blue Compact Dwarf galaxies (BCDs) and other types of dwarf galaxies. BCDs provide ideal laboratories to study intense star formation episodes in low mass dwarf galaxies, and have sometimes been considered a short-lived evolutionary stage between types of dwarf galaxies. To test these connections, we consider a sample of BCDs as well as a comparison sample of nearby galaxies from the Local Volume Legacy (LVL) survey for context. We fit the multi-wavelength spectral energy distributions (SED, far-ultra-violet to far-infrared) of each galaxy with a grid of theoretical models to determine their stellar masses and star formation properties. We compare our results for BCDs with the LVL galaxies to put BCDs in the context of normal galaxy evolution. The SED fits demonstrate that the star formation events currently underway in BCDs are at the extreme of the continuum of normal dwarf galaxies, both in terms of the relative mass involved and in the relative increase over previous star formation rates. Today’s BCDs are distinctive objects in a state of extreme star formation which is rapidly transforming them. This study also suggests ways to identify former BCDs whose star formation episodes have since faded.'
author:
- 'Steven Janowiecki, John J. Salzer, Liese van Zee, Jessica L. Rosenberg, Evan Skillman'
title: Constraining the Stellar Populations and Star Formation Histories of Blue Compact Dwarf Galaxies with SED fits
---
Introduction
============
The formation and evolutionary pathways of dwarf galaxies are still not well-understood, despite the fact that they are fairly simple systems and are the most abundant type of galaxy in the universe. Many studies have focused on the idea that dwarf galaxy evolution is driven by interactions and mergers, through ram pressure stripping, tidal disruption, or other external transformations. Today, there are many well-established mechanisms to evolve dwarf galaxies in high density group and cluster environments. It has been more difficult to study the ongoing secular dwarf galaxy evolution in isolated environments where mergers and interactions have a smaller impact. In this regime, star formation is the most transformative process a dwarf galaxy can undergo.
The best places to study the transformative effects of star formation are in the dwarf galaxies with the highest star formation rates (SFRs): the Blue Compact Dwarf galaxies (BCDs). Dwarf galaxies are often broadly categorized as either dwarf irregulars (dIs) which lack a regular morphology, possess substantial gas reservoirs, and are often forming stars, or as dwarf spheroidals (dSphs, or dwarf ellipticals, dEs), which have regular isophotes, are gas-poor, and are not forming stars. BCDs are classified in a variety of ways, but are observationally remarkable for their strong emission lines and star formation [@sargentsearle70; @searlesargent72; @izotovthuan04], their underlying old stellar population [@loosethuan86; @aloisi07], their exceptionally low gas-phase metallicity [@terlevich91; @izotov94; @izotovthuan99; @hunter99], and their compact underlying stellar and HI distributions [@papaderos96; @vanzee98; @janowiecki14; @lelli14]. Some groups have hypothesized that a compact mass distribution is the most distinctive characteristic of a BCD [@vanzee98; @lelli14; @mcquinn15c].
Recently, many groups have used observations to study the evolutionary connections between BCDs and other dwarf galaxies. Detailed studies of individual extreme objects [e.g., @guseva01; @ashley14] and systematic surveys of many objects [e.g., @noeske07] have both suggested that BCDs are members of a rapidly evolving class of galaxies, and play an important role in dwarf galaxy evolution. Deep surface photometry studies [@noeske03; @noeske05] in addition to spectroscopic observations [@papaderos08] have highlighted the possible connections between BCDs and dwarf irregulars in terms of their structural parameters and chemical evolution. Other observations suggest that the unusually low metallicity of BCDs is more likely maintained by outflows of enriched winds [@carigi95; @maclow99; @mcquinn15b] rather than pristine gas infall [@matteucci83]. There have also been efforts to find galaxies which may have experienced a BCD-like star formation event at some point in their history, or to predict the future evolutionary state of today’s BCDs [@sanchezalmeida08; @sanchezalmeida09; @amorin12; @lelli12; @koleva13; @meyer14].
One of the big questions about BCDs is the reason for their intense star formation, and whether they have been triggered in some way. There are well-studied examples where mergers or interactions between dIs have externally triggered intense star formation [e.g., II Zw 40, @sargentsearle70; @baldwin82; @terlevich91; @vanzee98; @bordalo09]. Recently, cosmologically-relevant hydrodynamical simulations are beginning to approach these questions from a theoretical perspective [e.g., @valcke08 and references therein]. While dwarf-dwarf galaxy mergers become increasingly rare at lower masses [@deason14], some groups have had success in producing BCD-like galaxies through these interactions [@bekki08]. Others have considered the effects of in-spiraling star-forming clumps in dwarf galaxies [@elmegreen12], or the interaction between a dwarf galaxy and an infalling cloud of gas [@verbeke14], both of which can reproduce some of the observed properties of BCDs. Both the merger and gas infall scenarios frequently result in substantial structural changes to the simulated galaxies, and represent extreme transformations in the life of a dI. However, in this work we focus on the internal secular evolution of isolated BCDs, and avoid discussing mergers and interactions.
Compared with typical dIs, BCDs have been found to be especially compact, in their underlying stellar and HI distributions [@papaderos96; @vanzee98; @janowiecki14; @lelli14]. This compactness may be related to their ability to host such intense starbursts. When parametrizing the strength of a star-forming event, the birthrate parameter ($b$$=$SFR$/$$\langle$SFR$\rangle$, see Section \[bstrength\]) is often used, which compares the current star formation with the lifetime average. In the local universe, intense star-forming galaxies are rare as only $\sim$$1\%$ of star-forming galaxies are considered starbursts with $b\ge3$ [@bergvall15]. When combining this rarity with the stochastic nature of star-formation in dwarf galaxies [e.g., @lee09; @weisz12; @bauer13], it would appear that dwarf galaxies can experience increases and decreases in their star formation rate, and only a small fraction are starbursting at any given time. Perhaps the especially-compact BCDs are able to burst more effectively than typical dIs, and so can reach higher SFR during their periods of starburst.
To compare the star formation properties and evolutionary history of BCDs and dIs from the Local Volume Legacy survey [LVL, @dale09], we consider the wealth of information contained in their stellar populations. An understanding of these stellar populations can constrain the amount and impact of recent and past star-formation activity. We have obtained multi-wavelength photometry in order to fit the spectral energy distributions (SEDs) for the BCDs and LVL galaxies. SED fitting has recently become a widely-used tool to derive physical properties of galaxies, including their stellar masses and star formation rates [cf., @walcher11].
Given that stellar mass is often considered the most fundamental parameter driving a galaxy’s evolutionary path [e.g., @tremonti04; @kewleyellison08 and references therein], star forming galaxies are often plotted on a “main sequence”, analogously to the main sequence of stellar evolution. This correlation between the stellar masses and SFRs of star-forming galaxies has been observed for massive galaxies in the nearby universe [@brinchmann04; @salim07] and at higher redshift [@daddi07], with a scatter of only $\sim$$0.2$ dex. The fact that this relationship exists with such low scatter across a wide range of redshifts seems to imply a universal mode of star formation, from which galaxies rarely deviate [@noeske07]. At smaller masses (below $10^9$[M\_]{}) the scatter becomes larger, as episodic bursts of star formation can affect dwarf galaxies more significantly [e.g., @cook14b; @mcquinn10a; @weisz11].
Galaxy stellar masses are most commonly determined by converting observed luminosities to masses [e.g., @bell01; @McGaughSchombert14]. Often a color (e.g., B$-$V), is used to determine the mass-to-light ratio, but these estimates become less reliable for galaxies which deviate from mean scaling relations (e.g., if they are currently experiencing a starburst). In order to more uniformly determine stellar masses across our sample of BCDs and LVL galaxies, a full multi-wavelength SED-fit is necessary.
Our SED fits can be also used to make crude estimates of the SFR and star formation histories (SFHs) for BCDs and LVL galaxies, which describe the amount of star formation they experienced throughout their lifetimes. Different observational indicators are sensitive to star formation of different ages, and the SED fits incorporate all of the multi-wavelength information into a single best fitting SFH. While the broadband photometry can never produce as accurate and well-constrained a SFH as resolved stellar photometry [e.g., @tolstoy09; @mcquinn10a; @mcquinn15a], it can still be useful in comparing BCDs with typical LVL galaxies. In particular, the stellar masses and SFRs from the SED fits allow us to quantify how extreme the BCDs are compared with typical LVL galaxies, and to constrain the possible evolutionary connections between BCDs and other dwarf galaxies. This paper is organized as follows. In Section \[obs\], we describe the multi-wavelength photometric observations of the BCDs and comparison samples. In Section \[cigale\] we discuss our SED-fitting methods and the consistency checks and verifications of our results. In Section \[sed\_results\] we show the results of our best SED fits. Section \[sed\_discussion\] contains a discussion of these results and their implications, and we briefly summarize our main findings in Section \[sed\_summary\].
[cccccccccccc]{} UM 323 & F/N & & SN/W & W & W & SN & SN & SN/W & SN/W & W &\
UM 408 & F/N & & SN/W & W & W & SN & SN & SN/W & SN/W & W &\
Mk 600 & F/N & & SN/W & W & W & SN & SN & SN/W & SN/W & SN/W & I/M\
II Zw 40 & F/N & SN & SN & SN & SN & SN & SN & SN & SN & SN & I/M\
Mk 5 & F/N & SN & SN/W & SN & W & SN & SN & SN & SN & SN & I/M\
CG 10 & F/N & & W & W & W & & & W & W & W &\
I Zw 18 & F/N & SN & SN & SN & SN & & SN & SN & SN & SN & I/M\
Was 5 & F/N & SN & SN & SN & W & & SN & SN & SN & SN &\
Mk 36 & F/N & SN & SN & SN & W & & SN & SN & SN & SN & I/M\
UM 439 & F/N & SN & SN & SN & W & & SN & SN & SN & SN &\
Mk 750 & & SN & SN & SN & W & & SN & SN & SN & SN &\
UM 461 & N & SN & SN & SN & W & & SN & SN & SN & SN & I/M\
UM 462 & F/N & SN & SN & SN & W & & SN & SN & SN & SN & I/M\
Mk 67 & F/N & SN & SN & SN & SN & SN & SN & SN & SN & SN & I\
Mk 475 & F/N & SN & SN & SN & W & & SN & SN & SN & SN & I/M\
Mk 900 & F/N & & SN/W & W & W & SN & SN & SN & SN & SN & I/M\
Mk 324 & F/N & SN & SN/W & SN & SN/W & SN & SN & SN & SN & SN &\
Mk 328 & F/N & SN & SN/W & SN & SN/W & SN & SN & SN & SN & SN &\
[lrrrrrrrrrrrrrrrrr]{} UM 323 & 0.712 & 0.700 & & 1.445 & 1.540 & 1.670 & 1.864 & 2.638 & 2.338 & 1.900 & & & & & & &\
& & 0.001 & & 0.013 & 0.040 & 0.062 & 0.017 & 0.165 & 0.125 & 0.258 & & & & & & &\
UM 408 & 0.153 & 0.179 & & 0.361 & 0.420 & 0.437 & 0.484 & 0.492 & 0.472 & & & & & & & &\
& 0.003 & 0.002 & & 0.003 & 0.037 & 0.039 & 0.008 & 0.040 & 0.031 & & & & & & & &\
Mk 600 & 0.792 & 0.924 & & 2.524 & 2.788 & 3.051 & 3.296 & 3.978 & 3.985 & 2.647 & 2.083 & 1.879 & 1.680 & 1.440 & 21.568 & 161.2 & 139.6\
& 0.005 & 0.004 & & 0.026 & 0.226 & 0.248 & 0.024 & 0.161 & 0.143 & 0.217 & 0.250 & 0.238 & 0.237 & 0.208 & 1.716 & 9.4 & 20.2\
II Zw 40 & 0.086 & 0.091 & 1.028 & 2.034 & 3.336 & 6.241 & 9.719 & 20.722 & 23.902 & 20.627 & 24.962 & 23.463 & 54.652 & 134.6 & 1616 & 4787 & 1427\
& 0.005 & & 0.044 & 0.058 & 0.092 & 0.161 & 0.242 & 0.534 & 0.815 & 0.494 & 0.903 & 0.869 & 1.391 & 2.1 & 15 & 48 & 109\
Mk 5 & 0.308 & 0.410 & 0.946 & 1.550 & 1.902 & 2.984 & 3.198 & 3.509 & 3.802 & 2.742 & & 6.745 & & 19.041 & 16.118 & 236.8 & 144.7\
& 0.004 & 0.003 & 0.044 & 0.033 & 0.070 & 0.259 & 0.027 & 0.239 & 0.210 & 0.202 & & 0.462 & & 0.790 & 1.459 & 10.0 & 17.0\
CG 10 & 0.141 & 0.153 & & 0.281 & 0.433 & 0.458 & & 0.422 & 0.504 & 0.355 & & & & & & &\
& 0.003 & 0.002 & & 0.014 & 0.035 & 0.038 & & 0.019 & 0.033 & 0.033 & & & & & & &\
I Zw 18 & 1.293 & 1.176 & 0.881 & 0.905 & 0.795 & 0.731 & 0.649 & 0.601 & 0.452 & 0.565 & & & & & 5.827 & 36.058 &\
& 0.010 & 0.006 & 0.043 & 0.022 & 0.018 & 0.017 & 0.019 & 0.037 & 0.041 & 0.085 & & & & & 0.863 & 4.595 &\
Was 5 & 0.247 & 0.299 & 0.322 & 0.501 & 0.685 & 0.814 & & 0.986 & 0.786 & 0.667 & & & & & & &\
& 0.004 & 0.003 & 0.013 & 0.014 & 0.019 & 0.011 & & 0.054 & 0.072 & 0.096 & & & & & & &\
Mk 36 & 1.590 & 1.574 & 1.615 & 1.893 & 1.808 & 2.432 & & 2.643 & 2.312 & 1.937 & & 1.165 & & 1.645 & 26.197 & 368.0 & 51.638\
& 0.012 & 0.007 & 0.062 & 0.052 & 0.048 & 0.031 & & 0.122 & 0.143 & 0.148 & & 0.182 & & 0.224 & 1.950 & 15.9 & 9.159\
UM 439 & 1.482 & 1.779 & 2.188 & 3.251 & 3.449 & 3.628 & & 4.752 & 4.357 & 2.858 & & & & & & &\
& 0.013 & 0.006 & 0.103 & 0.096 & 0.099 & 0.043 & & 0.149 & 0.133 & 0.155 & & & & & & &\
Mk 750 & & & 1.603 & 2.235 & 2.687 & 3.108 & & 3.054 & 2.938 & 2.968 & & & & & & &\
& & & 0.069 & 0.066 & 0.074 & 0.040 & & 0.107 & 0.152 & 0.159 & & & & & & &\
UM 461 & 0.462 & 0.534 & 0.618 & 0.974 & 1.052 & 1.251 & & 1.328 & 1.205 & 0.637 & 0.787 & 0.625 & 0.826 & 1.659 & 34.178 & 272.1 &\
& 0.010 & 0.007 & 0.031 & 0.033 & 0.037 & 0.016 & & 0.067 & 0.087 & 0.118 & 0.148 & 0.130 & 0.160 & 0.225 & 2.164 & 16.3 &\
UM 462 & 2.832 & 3.232 & 3.480 & 4.529 & 4.622 & 5.629 & & 5.572 & 4.840 & 3.163 & 3.452 & 3.262 & 4.943 & 8.185 & 115.9 & 866.6 & 278.1\
& 0.022 & 0.016 & 0.215 & 0.163 & 0.170 & 0.067 & & 0.205 & 0.147 & 0.175 & 0.327 & 0.317 & 0.439 & 0.546 & 4.0 & 19.7 & 21.9\
Mk 67 & 0.237 & 0.371 & 0.767 & 1.064 & 1.293 & 1.565 & 1.800 & 2.268 & 1.988 & 1.247 & & & & & & &\
& 0.004 & 0.003 & 0.035 & 0.027 & 0.032 & 0.039 & 0.051 & 0.094 & 0.099 & 0.113 & & & & & & &\
Mk 475 & 0.318 & 0.436 & 0.687 & 0.992 & 1.292 & 1.452 & & 1.883 & 1.698 & 1.297 & 1.029 & 0.811 & 0.751 & 0.950 & 10.571 & 117.3 &\
& 0.011 & 0.003 & 0.027 & 0.028 & 0.035 & 0.019 & & 0.094 & 0.097 & 0.098 & 0.171 & 0.151 & 0.148 & 0.167 & 1.167 & 7.5 &\
Mk 900 & 0.715 & 1.101 & & 4.613 & 8.318 & 10.745 & 12.394 & 16.384 & 17.475 & 12.707 & 10.060 & 6.775 & 10.556 & 19.652 & & 417.9 & 380.6\
& 0.005 & 0.005 & & 0.081 & 0.261 & 0.485 & 0.069 & 0.498 & 0.515 & 0.421 & 0.567 & 0.464 & 0.586 & 0.798 & & 14.4 & 25.4\
Mk 324 & 0.700 & 0.818 & 1.514 & 2.780 & 3.726 & 5.321 & 5.167 & 6.874 & 7.593 & 5.593 & & & & & & &\
& 0.008 & 0.005 & 0.056 & 0.013 & 0.103 & 0.147 & 0.024 & 0.215 & 0.224 & 0.191 & & & & & & &\
Mk 328 & 0.324 & 0.435 & 1.213 & 2.542 & 3.938 & 6.457 & 6.843 & 10.765 & 12.463 & 8.938 & & & & & & &\
& 0.004 & 0.003 & 0.045 & 0.028 & 0.109 & 0.178 & 0.032 & 0.317 & 0.230 & 0.181 & & & & & & &\
Observational Data {#obs}
==================
Our primary sample of 18 actively star-forming galaxies is the same as that of @janowiecki14, and includes a variety of BCDs and related galaxies. Some are canonical BCDs (e.g., I Zw 18), while some have smooth outer isophotes (similar to a dE) with a strong central starburst (e.g., Mk 900). Some have offset starbursts (e.g., Mk 36, Mk 750), dual starburst regions (UM 461, Mk 600), cometary shapes (Mk 5), or large numbers of active star formation sites (e.g., UM 439, UM 462, UM 323). Our sample is faint ($\langle$M$_\textrm{B} \rangle$=$-16$ mag), blue ($0$$<$B$-$V$<$$0.5$), metal deficient ($\langle 12+$log$($O$/$H$) \rangle
$$\sim$$8$), and less than 50 Mpc away [see Table 1 in @janowiecki14 for more details]. Our BCD sample is not comprehensive, but instead is representative and its members span the range of parameter space that BCDs typically occupy. Samples of BCDs have been defined in various ways, beginning from their identification as extragalactic HII regions [@sargentsearle70], continuing through the definitions of @tm81 and @gildepaz03. All of these samples commonly include dwarf galaxies which are compact (or merely small) and intensely forming stars. However, conclusions about the evolution of BCDs can depend strongly on sample selection [see also Section 5 of @janowiecki14]. Rather than creating or adopting a definition of BCDs, we instead select a sample of BCDs and BCD-like galaxies to study the extremes of dwarf galaxy evolution. Toward this end, we include the galaxies in the Local Volume Legacy (LVL) survey as a comparison sample. LVL is a volume-limited survey of 258 galaxies within $11$ Mpc, which includes flux observations from the ultraviolet to the far-infrared. The LVL sample is a particularly good comparison sample as its volume limited nature means that a majority of its galaxies have stellar masses below $10^9M_\odot$, similar to our BCD sample [@dale09]. In fact, some of the galaxies in the LVL sample are classified as BCDs, as discussed in Section \[bcd\]. In this analysis, we do not consider the environment of LVL galaxies (or BCDs), and treat each galaxy individually; further work is needed to explore the effects of the environment on star formation processes in dwarf galaxies.
Photometric observations of the BCD sample
------------------------------------------
The multi-wavelength photometric observations of our BCD sample come from many sources. Table \[bcdobs\] shows a complete summary of the photometry for each object, and Table \[bcddata\] contains all of the observed fluxes of our BCD sample. We also use gas-phase abundances from spectroscopic observations of the BCDs from sources in the literature [@zhao10; @brinchmann08; @izotov07]. These abundances allow us to consider the chemical evolution of the BCDs and put them in context of galaxy scaling relations.
In the UV, we use far-ultraviolet (FUV) and near-ultraviolet (NUV) observations from the *Galaxy Evolution Explorer* . Most of our galaxies were observed through our Cycle 3 program (GI3-089), but six are from archival images. *GALEX* acquires images in FUV ($1344-1786$Å) with a $4.3''$ FWHM PSF, and in NUV ($1771-2831$Å) with a $5.3''$ FWHM PSF. The images are processed and calibrated with the standard pipeline and were downloaded from the archive. The fluxes are measured in large apertures on the calibrated images. These UV images are especially sensitive to recent star formation, and provide a critical measurement of the presence and impact of young stars in the BCDs.
Some of our optical photometry comes from @sudarsky95 and @norton97, which contain complete details of their reduction and calibration. In brief, the UBVRI optical CCD observations were carried out at the Kitt Peak National Observatory[^1] (KPNO) 0.9m telescope in November/December 1989 and April 1990. The images were calibrated with observations of Landolt standards, and total fluxes were measured. NIR JHK photometry was also obtained at KPNO with IR arrays [@salzerelston92]. Table \[bcdobs\] shows which photometric measurements come from this dataset (labeled “SN”), and the fluxes are given in Table \[bcddata\].
We have expanded this existing optical/NIR photometry with new observations from the WIYN[^2] 3.5-m telescope at KPNO. Complete details of the observations, reductions, and calibrations are given in @janowiecki14. In short, the Minimosaic and OPTIC (Orthogonal Parallel Transfer Imaging Camera) imagers were used to obtain optical observations and the WHIRC [WIYN High Resolution Infrared Camera, @meixner10] imager was used to obtain NIR observations. Observations were taken between November 2008 and April 2010, and were calibrated with catalog measurements from Sloan Digital Sky Survey (SDSS) and Two Micron All Sky Survey (2MASS) of stars in the fields with our targets. Table \[bcdobs\] shows which flux measurements come from our recent WIYN observations (labeled “W”), and our complete set of fluxes and uncertainties are reported in Table \[bcddata\].
We also have calibrated narrow band Ha observations of the BCD sample which are used to determine SFRs from the standard prescription of @kennicutt98. These [H$\alpha$ ]{}fluxes were measured from narrow-band imaging [@salzerelston92], and the calibrated [H$\alpha$ ]{} luminosities are given in Table \[bcdha\], along with metallicity values from the literature for the BCD sample. Distances in Table \[bcdha\] use flow models (Virgo+GA+Shapley) from the NASA Extragalactic Database (NED[^3]), and for I Zw 18 we use the tip of the red giant branch distance from @aloisi07.
[cccc]{} UM 323 & 38.5 & 25.6 & 7.96$^a$\
UM 408 & 35.2 & 47.5 & 7.74$^a$\
Mk 600 & 27.3 & 13.6 & 7.94$^a$\
II Zw 40 & 40.3 & 11.1 & 8.09$^a$\
Mk 5 & 7.7 & 15.3 & 8.06$^a$\
CG 10 &&30.7 &\
I Zw 18 & 24.3 & 18.2 & 7.18$^a$\
Was 5 & 14.3 & 23.1 & 7.85$^b$\
Mk 36 & 6.8 & 8.4 & 7.82$^a$\
UM 439 & 16.2 & 15.9 & 8.08$^a$\
Mk 750 & 2.7 & 5.2 & 8.18$^b$\
UM 461 & 9.1 & 12.7 & 7.81$^c$\
UM 462 & 32.6 & 13.5 & 7.80$^c$\
Mk 67 & 17.4 & 18.7 & 8.08$^a$\
Mk 475 & 6.9 & 11.9 & 7.93$^a$\
Mk 900 & 84.0 & 18.9 & 8.07$^a$\
Mk 324 & 31.3 & 23.2 & 8.18$^d$\
Mk 328 & 23.8 & 20.6 & 8.64$^d$\
Observations of the LVL sample
------------------------------
The multi-wavelength observational data of the 258 galaxies in the LVL sample are nearly identical in quality and wavelength coverage to our observations of the BCD sample. The SEDs for the complete LVL sample are presented in @cook14b. These SEDs include *GALEX* FUV and NUV images [@lee11], ground-based optical UBVRI images [@cook14a], NIR JHK images from 2MASS [Two Micron All Sky Survey @dale09], and Spitzer IR images [@dale09]. This observational dataset is used to determine the multi-wavelength SEDs from $1500$Å$-24\mu$m for galaxies in the LVL sample. Constructing SEDs with self-consistent UV, optical, NIR, and FIR fluxes is a complex process, and care was taken to measure the galaxies within a common aperture across all wavelengths [@cook14b]. In all, fluxes are reported in 14 band-passes between *GALEX* FUV and Spitzer $24\mu$m for all of the 258 LVL galaxies. As noted by @cook14b, there are 47 upper limits on non-detections in the IR observations, 13 upper limits in the UV observations, and 1 upper limit in the optical data. In total, 3551 fluxes are used in the LVL SEDs. @cook14b present the LVL SEDs and derive physical properties from them, including their star formation rates, stellar masses, and internal extinction
![Different symbols show the L-Z relation for all of the LVL galaxies with measured gas-phase metallicities (various methods). The solid blue lines show the best forward-fit and reverse-fit to this relationship, and the red line shows their average. Members of our BCD sample are shown as blue points at their measured abundances. The magenta line shows the best-fit L-Z relation from @tremonti04, which agrees well with our fit. \[MZ\]](MZ-crop.eps){width="7.2cm"}
In addition to the complete set of panchromatic SEDs, gas-phase metallicities have been measured for much of the LVL sample. @cook14b compiles metallicity measurements from @marble10, @berg12, and @moustakas10, which come from different metallicity methods calibrations [see @marble10; @mcgaugh91; @pilyuginthuan05; @kobulnickykewley04]. Of the 258 galaxies in LVL, 155 have measured metallicities (about half “direct” and half “strong-line” methods), which are shown in the luminosity-metallicity relationship (L-Z) in Figure \[MZ\]. Also shown in Figure \[MZ\] is the L-Z relation from @tremonti04 for a sample of $\sim$$53,000$ SDSS galaxies brighter than M$_\textrm{B}$=$-$$16$ mag, which agrees well with the L-Z relation from the LVL and BCD samples.
BCDs in LVL {#bcd}
-----------
Some of the galaxies in LVL have been classified as BCDs – in particular Mk 475 is a member of both our BCD sample and of the LVL sample. For brevity, throughout this work we refer to the “BCD sample” and the “LVL sample”, but there exists a continuum of galaxy properties across both samples. The LVL sample [@dale09] includes both early- and late-type galaxies, and in particular includes 7 objects which identified as BCDs in the Palomar/Las Campanas Atlas of BCDs [@gildepaz03; @kennicutt08].
On subsequent plots and figures we take care to indicate these individual populations (BCD, irregular/spiral, early-type) within the LVL sample. Our goal is not to demarcate BCDs from “normal” dwarf galaxies, but rather to explore the extremes of dwarf galaxy evolution. BCDs are often characterized by their significant star formation rates, but they are unique in other aspects as well. Throughout this work we will discuss the continuum of dwarf galaxy properties, and show the extreme position that BCDs occupy.
[sed\_biggrid-crop.eps]{} (20.91,81.6) [{width="1.1cm"}]{} (20.91,45.0) [{width="1.1cm"}]{} (20.91,26.69) [{width="1.1cm"}]{} (20.91, 8.32) [{width="1.1cm"}]{}
(39.2,81.6) [{width="1.1cm"}]{} (39.2,63.3) [{width="1.1cm"}]{} (39.2,45.0) [{width="1.1cm"}]{} (39.2,26.69) [{width="1.1cm"}]{} (39.2, 8.32) [{width="1.1cm"}]{}
(57.49,81.6) [{width="1.1cm"}]{} (57.49,63.3) [{width="1.1cm"}]{} (57.49,45.0) [{width="1.1cm"}]{} (57.49,26.69) [{width="1.1cm"}]{}
(75.84,63.3) [{width="1.1cm"}]{} (75.84,45.0) [{width="1.1cm"}]{} (75.84,26.69) [{width="1.1cm"}]{} (75.21,81.65) [{width="1.1cm"}]{}
SEDs
----
Figure \[allsed\] shows the complete set of observed SEDs for our sample of BCDs. The SEDs have been corrected for Galactic extinction [via @sfd98 for consistency with the LVL photometry] and are shown in flux units of mJy. Also shown are $100''$$\times$$100''$ color cutout images from SDSS Data Release 12 [DR12, @dr12], where available. The SEDs shown in Figure \[allsed\] demonstrate the amount and quality of the photometric observations of this sample. Most of the BCDs have UV fluxes, and many have complete Spitzer FIR coverage as well. When compared with the average SEDs of the low-mass late-type (non-BCD) LVL galaxies [e.g., Figure 5 of @cook14b, and shown in our Figure \[allsed\]], the BCDs show much flatter (bluer) SEDs at UV/optical wavelengths. This is consistent with the presence of substantial recent star formation activity. The three most massive galaxies in our sample (Mk 324, Mk 328, Mk 475, Mk 900) have noticeably redder optical SEDs, and also have smooth elliptical isophotes in their outskirts. These three galaxies are classified as BCD type “nE” by @gildepaz03, and will be referred to as “BCD/E” on subsequent plots.
Across the sample of BCDs, the UV slope shows significant variations from the steep rise of Mk 67 to the flat slope of Mk 36. The UV slope is sensitive to both the internal absorption from dust as well as the current star formation. The IR observations are necessary to disentangle the effects of dust and star formation.
The detailed shapes of these panchromatic SEDs encode much of the information about the star formation history, stellar populations, and dust in the BCD sample. The SEDs of each LVL galaxy are not reproduced here, but are shown and discussed extensively in @cook14b.
Fitting SEDs {#cigale}
============
We fit our SEDs with CIGALE [Code Investigating Galaxy Emission, @noll09]. CIGALE fits SEDs from UV to far infrared in order to account for dust absorption and re-emission in a self-consistent manner. It creates a grid of synthetic SEDs based on theoretical models to account for all of the relevant line and continuum emission and absorption from stars, gas, and dust. This grid of models is then compared with observed SEDs in order to determine the most likely values and uncertainties for various physical parameters. In the following sub-sections we describe the models that CIGALE uses to generate SEDs, our verifications of the appropriateness of the models, and the consistency checks we employ to understand the reliability of the fit results.
Input Models {#input_models}
------------
CIGALE uses theoretical models to parametrize the flux emitted and absorbed by the stars, gas, and dust in model galaxies, and produces a grid of model spectra which are then converted to SEDs. @noll09 describes the input models in complete detail, and we briefly review each component and its contribution.
The stellar emission is modeled using stellar population synthesis models from @bc03 with a Salpeter IMF, between metallicities of $Z$$=$$0.0001$ and $Z$$=$$0.05$. Two stellar populations are typically used: young and old stellar populations. The light from the stellar populations represents the dominant source of emission in the UV-optical range.
Nebular line and continuum emission are included in the UV-NIR range, taking into account escape fraction and dust absorption. Using the number of Lyman continuum photons (from the stellar continuum) to compute the strength of the H$\beta$ line, a metallicity-dependent template is used to determine the strengths of 119 other nebular lines via an estimate of the number of ionizing photons [@inoue11].
Dust attenuation is handled with the formulas from @calzetti2000 and @leitherer02, based on the method of @ccm89. In general, this requires computing and applying the attenuation curve (A$(\lambda)/$E$($B$-$V$)$) to all of the relevant flux-emitting components considered in the model (both stellar and nebular emission). The @calzetti2000 attenuation law is used as a baseline but we allow its slope to change by multiplying it by $(\lambda/\lambda_\textrm{V})^\delta$ where $\lambda_\textrm{V} = 5500$Å, and $\delta$ ranges between $-1$ and $2.5$. This modification is required based on the variations seen in the (limited number) of observations of the extinction law in other galaxies [@witt00; @inoue06]. The attenuation is combined with the emission models as a “negative” flux, and is calculated separately for each component of the model. The attenuation applied to the light coming from the old stellar population is reduced by a factor of $f_\textrm{att} = 0.5$ from the value of the young population, to account for the dustier nature of star forming regions. This has a small effect. A key advantage of CIGALE’s SED-fitting is its multi-wavelength energy balance, and that consideration drives its treatment of infrared re-emission from heated dust. CIGALE combines the amount of attenuation present in the models with the semi-empirical re-emission templates of @dale14. The templates are generated by considering the contributions from a variety of dust heating intensities and depend on a single heating parameter, $\alpha$. This exponent is the only free parameter in the dust re-emission models as the total energy is constrained to be equivalent to the amount that has been attenuated. CIGALE is also capable of modelling the emission from a dust-enshrouded AGN, but we do not include that option in our fits; BCDs do not host AGN, and LVL contains mostly late-type dwarf galaxies (NGC 855 is the most massive elliptical galaxy in sample, at only $\sim$$10^9$[M\_]{}), so we do not expect any significant AGN emission.
[lccc]{} Old stellar population age & age$_o$ & $1000, 3000, 10000$ & Myr\
Old stellar population $e$-folding time & $\tau_o$ & 100, 1000, 10000 & Myr\
Young stellar population age & age$_y$ & $30, 100, 300, 1000, 3000$ & Myr\
Young stellar population $e$-folding time & $\tau_y$ & $10, 100, 1000$ & Myr\
Young stellar population mass fraction & $f_{b}$ & &\
Stellar metallicity & $Z$ & & $Z/Z_\odot$\
Amount of dust attenuation & E$($B$-$V$)_y$ & & mag\
Power-law slope on extinction law & $\delta$ & &\
Dust heating parameter & $\alpha$ & &\
CIGALE computes and applies the attenuation resulting from the effects of the intergalactic medium (IGM), based on the prescription of @meiksin06. While this attenuation is relatively small for the nearby galaxies, we include it for consistency when comparing with galaxies at different redshifts. The contribution of the IGM to the overall attenuation is most relevant at very short wavelengths.
We can consider a wide variety of SFH scenarios in our models. CIGALE currently includes three standard options: A double decreasing exponential model, a “delayed” model, and a manually created SFH supplied by the user. We primarily consider the double exponential model, which describes the SFH as two exponential functions characterized by $e$-folding time $\tau$. The age and $\tau$ of both populations are free parameters, as is the mass fraction of the young population compared to the old population. A minimum age separation can be enforced between the two populations, and is typically set at $10$ Myr. In the following sub-sections, we consider increasingly complex SFH assumptions and the results of the SED fits.
Preliminary SED-fitting tests
-----------------------------
To gain familiarity and confidence in the SED fitting methods and results, we first considered simple SFH scenarios with very few free parameters (e.g., a single recent burst on top of an old burst, or on top of a constantly star-forming population). We varied grid parameters, sampling densities, and explored the effects these choices had on the outputs of the SED fits. Full details and results of these extensive tests can be found in Chapter 3 of @janowiecki15. In brief, we used a variety of diagnostics to check that the SED fits were reasonable, and also to realistically assess the quality of the outputs. In the following sub-sections, we discuss the grid choices and quality assessments of the fits and their results.
{width="5.65cm"} {width="5.65cm"} {width="5.65cm"}
SFH and grid parameters
-----------------------
In order to characterize a stellar population a SFH must be assumed. We use CIGALE’s double-exponential option to create a two-population SFH. Both the old and young stellar populations are described by declining SFRs which began abruptly and exponentially decline. In these fits, both the age and exponential scale time are allowed to vary, with the sampling of possible values shown in Table \[grid\]. Note that while the old stellar population age is allowed to be as young as $1000$ Myr and the young stellar population is allowed to be as old as $3000$ Myr, we require $\textrm{age}_o > \textrm{age}_y + 10$ Myr, to keep a meaningful distinction between the old and young stellar populations.
In this grid, the mass fraction of the young population is allowed to vary between $1\%$ and $50\%$. The dust attenuation, E$($B$-$V$)$, varies from 0.01 to 1.5 magnitudes, the Calzetti law is modified by a power law slope between $-1$ and $2.5$, and the dust heating parameter $\alpha$ varies between $0.5$ and $2.5$. The stellar metallicity is a free parameter and varies between $Z=0.0001$ and $0.05$. See Section \[input\_models\] for further details on these parameters.
This grid requires $\sim$$1,000,000$ SEDs to be computed, and the process of generating and fitting the SEDs takes about an hour on a laptop computer. Once the entire library of SEDs for each grid point is generated, they can be compared with our observed SEDs. Before carrying out the actual fitting process, we first demonstrate that the model grids are suitably well-matched and appropriate for fitting our observed SEDs.
Following the method described in Section 3.2 of @buat11, we test whether the colors of the model grids sufficiently overlap with our observed SEDs by placing them onto diagnostic color-color diagrams. We consider the FUV$-$NUV and NUV$-$R colors because of their connections with dust attenuation and star formation history. We also show the FUV$-$$70\mu$ color for its very large wavelength baseline, and the B$-$V and B$-$H colors as they are commonly used. Figure \[cc\] shows these diagnostic color-color plots. The galaxy with the most extreme FUV-NUV color is KDG 061, a tidal dwarf galaxy in the M81 group. Its *GALEX* observations are very deep (16,238s exposure time), but its FUV flux is very weak. Its optical colors are typical of dSphs but it has significant amounts of HI and [H$\alpha$ ]{}emission indicating ongoing star formation [@johnson97; @croxall09].
Figure \[cc\] also demonstrates the correlations between these colors in the models. Models are selected from within a single shaded pixel on the left panel (shown in green) and plotted on the other two plots at their appropriate colors. Similarly, models are selected from a single pixel on the right panel (in blue) and are shown on the other two diagrams. These model points demonstrate the connections between the different color-spaces, and the strengths of multi-wavelength SEDs from UV to IR.
Overall, it is clear that our model grids are sampling an adequate amount of color-color space to match most of the observations. Our choice of parameters for the stellar populations and dust attenuation appear to cover an appropriate range of values to be useful in fitting our observations.
Fitting SEDs and determining output values
------------------------------------------
Now that we have shown that the colors of our model grids are suitable comparisons to our observations, we can proceed to fit the SEDs. To determine the best-fit values of each parameter for each observed SED, CIGALE first calculates the $\chi^2$ value between the observations and each model grid point SED. As described by Equation (5) from @noll09, this is calculated using: $$\chi^2 (M_\textrm{gal}) = \Sigma^k_{i=1}
\frac{(M_\textrm{gal} f_\textrm{mod,$i$} - f_\textrm{obs,$i$})^2}
{\sigma^2_{\textrm{obs,$i$}}}
$$ where the difference between each flux measurement ($f_\textrm{obs,$i$}$) and each model flux point ($f_\textrm{mod,$i$}$) is divided by the uncertainty on the observed flux ($\sigma^2_\textrm{obs,$i$}$), in filter $i$ [see also @salim07]. The model fluxes are given per unit M$_\odot$, so are multiplied by the galaxy mass (M$_\textrm{gal}$). This summation is taken over all flux observations ($k$) in the SED. The photometric uncertainties on the flux observations, $\sigma_{\textrm{obs}}$, are included as a weighting factor.
After determining values of $\chi^2$ between a galaxy’s SED and all of the model grid points, CIGALE generates probability distribution functions (PDFs) for selected parameters in a Bayesian-like framework [@kauffmann03; @salim05; @salim07; @noll09]. For each parameter, CIGALE creates a number of bins between its lowest to highest values, and determines which models fall into each bin for that parameter. Among the models in each bin, the model with the highest probability (i.e., best match to observed SED) is found and reported. With these maximum probabilities from each bin, a PDF can be generated which represents the maximum envelope of the probability distribution for that parameter. These PDFs are used to generate expectation values and uncertainties (see Figures 6 and 7 in @noll09 for further details about this method). The key analyzed parameters from the fits are described in Table \[parms\], their error determinations are discussed further in Section \[uncert\], and the best-fit values for the BCD sample are shown in Table \[best1\].
[clc]{} $\langle$SFR$\rangle$$_{10}$ & SFR averaged over 10 Myr & $M_\odot/$yr\
$\langle$SFR$\rangle$$_{50}$ & SFR averaged over 50 Myr & $M_\odot/$yr\
$\langle$SFR$\rangle$$_{100}$ & SFR averaged over 100 Myr & $M_\odot/$yr\
$\langle$SFR$\rangle$$_{500}$ & SFR averaged over 500 Myr & $M_\odot/$yr\
$\langle$SFR$\rangle$$_{1000}$ & SFR averaged over 1000 Myr & $M_\odot/$yr\
$\langle$SFR$\rangle$$_\textrm{all}$ & SFR averaged over lifetime & $M_\odot/$yr\
$M_{\star,o}$ & Stellar mass of old stellar population & $M_\odot$\
$M_{\star,y}$ & Stellar mass of young stellar population & $M_\odot$\
$\tau_o$ & Exponential scale time of old SP & Myr\
age$_o$ & Age of old stellar population & Myr\
$\tau_y$ & Exponential scale time of young SP & Myr\
age$_y$ & Age of young stellar population & Myr\
$f_\textrm{burst}$ & Initial mass fraction of young SP &\
$Z$ & Stellar metallicity & $Z/Z_\odot$\
$A_\textrm{FUV}$ & Attenuation (internal) in $FUV$ & mag\
$A_\textrm{V}$ & Attenuation (internal) in $V$ & mag\
$\alpha$ & Dust heating exponent &\
$\delta$ & Slope modification on extinction law &\
We use CIGALE to fit SEDs using all of the fluxes from UV (FUV/NUV), optical (UBVRI), near-infrared (JHK), and infrared (Spitzer IRAC-MIPS). We examine the distribution of reduced $\chi^2$ values from the best-fitting SEDs to assess the success at fitting our observations. Figure \[x2\] shows the reduced $\chi^2$ distribution for our fits. On average, the reduced $\chi^2$ is $1.97$, and only $11$ galaxies have $\chi^2>5$.
![Distributions of values of reduced $\chi^2$ from our SED fits. The thick black line shows the distribution of the combined LVL and BCD samples, while the BCDs are shown separately in blue and the LVL galaxies are shown separately in red. \[x2\]](redx2-crop.eps){width="7.5cm"}
Reliability and uncertainty estimates {#uncert}
-------------------------------------
We now describe a few verifications and consistency checks which were used to estimate the reliability and uncertainty of the SED fits.
{width="\textwidth"}
It is important to verify that the parameter space covered by the grid is of suitable resolution and range. Inadequate resolution in a particular parameter will decrease the accuracy of that parameter’s determination, while over-sampling a poorly constrained parameter can artificially increase its uncertainty or negatively impact the determinations of other parameters [@buat11]. Figure \[hist1\] shows histograms of the best-fitting results for the most relevant parameters.
We experimented with different ranges and resolution of these parameters until satisfactory distributions were obtained. Initially, our grid was too narrow and some of the histograms showed unrealistically narrow spikes at the extremes of parameter space. We expanded the range of parameter space (by expanding the maximum or minimum values of the parameters) until there were no unrealistic peaks at the edges. Many of the parameters are logarithmically sampled in order to smoothly cover the wide range of parameter space. The resolution in each parameter was also adjusted until a generally smooth and continuous histogram was obtained, to eliminate unrealistically sharp features which were artifacts of inadequate sampling. The practical constraints of computing time were also considered, which prevented the grids from becoming unmanageably large. The resulting grid represents a compromise between covering the necessary parameter space with enough resolution and restricting the computational needs within reason. The final choices for parameter sampling are shown in Table \[grid\].
{width="13.5cm"}
In order to estimate the reliability of the SED fit results for each galaxy, we start with the best-fitting grid point for each object. This best-fitting model SED is treated as a “mock” observation (retaining the original photometric uncertainties on the real observations), and re-fitted with the same grid to re-generate the most likely value of each parameter. This “mock” fitting method allows us to estimate the reliability of the SED fit results [@salim09; @giovannoli10] by comparing “known” input parameters of the model SED with its re-fitted parameters.
Figure \[mock\] shows shows the results of our mock analysis for nine of the analyzed parameters. As before, we only show the SED fits which have a reduced $\chi^2<5$ and which had a complete set of observations for all $15$ flux points. Shown on the x-axes are the exact values of each parameter used to generate the “mock” SEDs. The y-axes show the differences between the exact “mock” values and the best-fit values from the SED re-fitting process. The number in parentheses at the bottom right corner of all panels is the rms scatter for that parameter.
In some cases, the exact parameter values in the “mock” SEDs can be reliability recovered by our fits. For example, the mass of the old stellar population ($M_\textrm{o,stellar}$) shows no systematic trends and a scatter of only $0.10$ dex. However, the mass of the young stellar population ($M_\textrm{y,stellar}$) has more scatter and perhaps a slope or offset at the lowest stellar masses. Furthermore, the best-fit values of the age of the young stellar population (age$_\textrm{y}$) show a small systematic trend where younger ages are likely to be under-estimated and older ages are likely to be over-estimated.
{width="47.00000%"} {width="47.00000%"}
The reliability of these parameters varies in this mock analysis, but these reliability estimates are only useful in conjunction with $\chi^2$ indicators and other independent verifications of the SED fits. The deviations shown here represent the reliability of the fit results, independent from our actual observations. The differences in best-fit parameters of the mock SEDs and the re-fit SEDs give an estimate of the reliability of this grid and method, and show which parameters are more reliably determined in this type of analysis. Some parameters are more reliably fit than others (e.g., the stellar mass of the young population has a great scatter than the stellar mass of the old population), which is incorporated into the uncertainties listed in the error budget in Table \[err1\]. This error budget also includes the full Bayesian uncertainties generated from our fitting procedure.
Further verifications {#external}
---------------------
{width="47.00000%"} {width="47.00000%"}
[cccccccccccc]{} UM323 & 9 & 0.3 & 8.13 (0.77) & 6.10 (0.71) & 0.124 (0.091) & 5.1 (3.3) & 0.42 (0.32) & 3.6 (2.3) & 0.76 (0.82) & 0.16 (0.15) & 0.32 (0.33)\
UM408 & 8 & 0.1 & 8.06 (1.56) & 5.91 (0.93) & 0.081 (0.089) & 4.9 (3.3) & 0.39 (0.31) & 3.4 (2.1) & 0.88 (0.90) & 0.18 (0.16) & 0.19 (0.25)\
Mk600 & 16 & 1.6 & 8.01 (0.37) & 5.50 (0.35) & 0.031 (0.010) & 5.1 (3.3) & 0.35 (0.29) & 3.7 (2.4) & 0.92 (0.85) & 0.19 (0.15) & 0.22 (0.03)\
IIZw40 & 17 & 8.4 & 8.99 (0.64) & 6.73 (0.20) & 0.531 (0.101) & 4.8 (3.3) & 0.36 (0.30) & 3.0 (1.7) & 1.06 (1.01) & 0.18 (0.15) & 0.67 (0.03)\
Mk5 & 15 & 8.9 & 8.21 (0.52) & 5.23 (0.62) & 0.018 (0.010) & 4.1 (3.2) & 0.29 (0.25) & 4.0 (2.6) & 0.89 (0.92) & 0.15 (0.14) & 0.30 (0.16)\
CG10 & 8 & 0.4 & 7.73 (0.90) & 5.45 (0.81) & 0.027 (0.021) & 4.7 (3.3) & 0.44 (0.32) & 3.6 (2.4) & 0.87 (0.86) & 0.14 (0.14) & 0.20 (0.24)\
IZw18 & 12 & 0.7 & 7.20 (0.64) & 5.85 (0.30) & 0.088 (0.027) & 3.6 (3.0) & 0.33 (0.28) & 5.5 (3.0) & 0.33 (0.02) & 0.23 (0.15) & 0.12 (0.06)\
Was5 & 9 & 0.4 & 7.87 (0.95) & 5.58 (1.04) & 0.036 (0.034) & 4.5 (3.3) & 0.46 (0.33) & 4.3 (2.8) & 0.72 (0.78) & 0.11 (0.12) & 0.23 (0.28)\
Mk36 & 14 & 2.5 & 7.02 (1.12) & 5.45 (0.13) & 0.023 (0.004) & 4.0 (3.1) & 0.53 (0.33) & 4.2 (2.7) & 0.33 (0.02) & 0.18 (0.15) & 0.10 (0.08)\
UM439 & 9 & 0.3 & 7.96 (0.44) & 5.93 (0.73) & 0.099 (0.094) & 5.3 (3.3) & 0.38 (0.30) & 3.7 (2.4) & 0.69 (0.78) & 0.20 (0.16) & 0.15 (0.18)\
Mk750 & 7 & 0.1 & 7.15 (1.00) & 5.08 (0.93) & 0.011 (0.011) & 4.7 (3.3) & 0.45 (0.32) & 3.9 (2.6) & 0.68 (0.76) & 0.16 (0.15) & 0.46 (0.39)\
UM461 & 15 & 1.2 & 7.52 (0.56) & 5.22 (0.32) & 0.016 (0.005) & 5.1 (3.3) & 0.45 (0.33) & 4.6 (2.9) & 0.60 (0.59) & 0.17 (0.15) & 0.41 (0.26)\
UM462 & 16 & 2.2 & 8.60 (0.65) & 5.95 (0.20) & 0.089 (0.018) & 4.2 (3.2) & 0.49 (0.33) & 5.3 (3.0) & 0.75 (0.79) & 0.11 (0.12) & 0.66 (0.03)\
Mk67 & 10 & 0.7 & 7.88 (0.63) & 5.70 (1.00) & 0.057 (0.061) & 4.8 (3.3) & 0.37 (0.30) & 4.1 (2.7) & 0.72 (0.82) & 0.17 (0.15) & 0.34 (0.30)\
Mk475 & 15 & 1.4 & 7.47 (0.78) & 4.91 (0.44) & 0.008 (0.003) & 4.1 (3.2) & 0.34 (0.29) & 3.7 (2.4) & 0.92 (0.87) & 0.18 (0.16) & 0.18 (0.13)\
Mk900 & 15 & 3.6 & 9.18 (0.30) & 5.79 (0.31) & 0.060 (0.017) & 2.5 (2.1) & 0.42 (0.32) & 5.3 (3.0) & 2.09 (0.98) & 0.28 (0.16) & 0.21 (0.06)\
Mk324 & 10 & 0.3 & 8.69 (0.71) & 6.23 (1.30) & 0.163 (0.192) & 5.2 (3.3) & 0.40 (0.31) & 4.3 (2.7) & 0.88 (0.85) & 0.14 (0.14) & 0.34 (0.35)\
Mk328 & 10 & 0.4 & 8.87 (0.51) & 6.14 (1.35) & 0.130 (0.160) & 3.9 (3.1) & 0.38 (0.31) & 4.2 (2.7) & 1.17 (1.01) & 0.12 (0.12) & 0.43 (0.40)\
[cccc]{} $\langle$SFR$\rangle$$_{10}$ & 0.62 $\pm$ 0.48 & -0.04 $\pm$ 0.25 &\
$\langle$SFR$\rangle$$_{50}$ & 0.58 $\pm$ 0.39 & -0.03 $\pm$ 0.19 & 0.34$^a$\
$\langle$SFR$\rangle$$_{100}$ & 1.00 $\pm$ 0.42 & 0.06 $\pm$ 0.23 & 0.29$^b$\
$\langle$SFR$\rangle$$_{500}$ & 0.55 $\pm$ 0.26 & 0.03 $\pm$ 0.15 &\
$\langle$SFR$\rangle$$_{1000}$ & 0.72 $\pm$ 0.29 & -0.03 $\pm$ 0.19 &\
SFR$_\textrm{all}$ & 0.40 $\pm$ 0.14 & -0.01 $\pm$ 0.07 &\
$M_{\star,o}$ & 0.43 $\pm$ 0.18 & 0.02 $\pm$ 0.10 & 0.14$^c$\
$M_{\star,y}$ & 0.62 $\pm$ 0.48 & -0.04 $\pm$ 0.25 &\
age$_o$ & 0.54 $\pm$ 0.12 & 0.01 $\pm$ 0.10 &\
$\tau_o$ & 0.66 $\pm$ 0.20 & -0.00 $\pm$ 0.13 &\
age$_y$ & 0.90 $\pm$ 0.21 & -0.01 $\pm$ 0.17 &\
$\tau_y$ & 0.76 $\pm$ 0.13 & -0.02 $\pm$ 0.12 &\
$f_\textrm{burst}$ & 0.88 $\pm$ 0.23 & -0.05 $\pm$ 0.15 &\
$A_\textrm{FUV}$ & 0.35 $\pm$ 0.38 & 0.01 $\pm$ 0.08 &\
$A_\textrm{V}$ & 0.40 $\pm$ 0.36 & 0.00 $\pm$ 0.10 &\
$\alpha$ & 0.15 $\pm$ 0.13 & 0.02 $\pm$ 0.05 &\
$\delta$ & 0.45 $\pm$ 0.21 & -0.03 $\pm$ 0.12 &\
In addition to these internal consistency checks on the SED grids and fitting methods, we can also use comparisons with independent determinations of similar parameters to verify our results. Most importantly, we can compare our SED-derived SFRs with those determined from [H$\alpha$ ]{}observations. We can also compare the SED-derived SFRs with estimates from the UV flux and compare the SED-derived stellar masses with estimates from $3.6\mu$m flux and K$_\textrm{s}$ flux. These are less independent comparisons than in [H$\alpha$ ]{}as these fluxes are already included in the SED fits. Still, these comparisons are an important consistency check of our fits, and also help provide a pathway toward comparing our SED-derived parameters with even broader samples from the literature.
[H$\alpha$ ]{}fluxes have been used to estimate SFRs for the LVL galaxies in @lee09 as part of the 11HUGs project. We use their extinction-corrected [H$\alpha$ ]{}SFRs, which have been determined using the standard @kennicutt98 relationship with a dust correction as described in @lee09, and follow a similar procedure for the BCD [H$\alpha$ ]{}fluxes. Figure \[sfr\] shows the comparison between our SED SFRs and the SFRs from @lee09 for the galaxies in common, and the same comparison using UV photometry. The [H$\alpha$ ]{}emission is most sensitive to star formation occurring over the past $\sim$$10$ Myrs while the shortest meaningful time scale we can measure SFR in the SED fits is $\sim$$50$ Myrs. Despite these timescale differences, the agreement is still quite good. The UV estimates of SFR are in better agreement with the 100 Myr-averaged SFH since the timescales are more closely matched, but the UV fluxes are themselves included in the SED-fitting.
The relationship between SED and [H$\alpha$ ]{}SFRs for the LVL galaxies shows generally good agreement, with a scatter of $\sim$0.3 dex. The [H$\alpha$ ]{}fluxes are not included in the SED fits, as it remains a challenge to incorporate them at low redshifts, but is often required for galaxies at higher redshifts [@ono10; @stark13; @debarros14]. In our case, the [H$\alpha$ ]{}fluxes give an independent estimate of the recent star formation in our sample, and allow us to assess the reliability of our SED fits. However, the SED fits rely primarily on UV fluxes to determine SFRs, while the [H$\alpha$ ]{}SFRs are measuring somewhat different star formation activity (see Section \[extremeMZ\] for further discussion on the systematic offset between SFRs from [H$\alpha$ ]{}fluxes and SED fits for the BCDs).
Consistent with many previous studies, @McGaughSchombert14 find that NIR luminosities of galaxies are a good tracer of their stellar mass, almost independently of their color (e.g., evolutionary state). They give mass-to-light ratios of M$/$L$=$$0.6$ $M_\odot/L_\odot$ in the K$_s$ filter, and M$/$L$=$$0.47$$M_\odot/L_\odot$ in the $3.6\mu$m band of Spitzer-IRAC (using Vega magnitudes). We use these mass-to-light ratios and our observed photometry to determine stellar masses for our galaxies, shown in Figure \[M\]. For BCDs without Spitzer observations, we generate $3.6\mu$m stellar masses from *w1* band photometry from the ALLWISE[^4] catalog from the Wide-field Infrared Survey Explorer [@wright10].
{width="47.00000%"} {width="47.00000%"}
Figure \[M\] shows the generally good agreement between SED masses and luminosity-based estimates. This is unsurprising as the SED fits already use the same fluxes that are used to generate the masses via mass-to-light ratios. Still, these comparisons demonstrate that our SED-fitting methods produce stellar masses which are consistent with those found using simpler methods. Further, the small offset and lack of mass-trends in the relationships indicate that there are no significant mass-dependent systematic effects.
Total uncertainty budget
------------------------
Putting all of these reliability and uncertainty estimates together, we now arrive at a complete error budget for our SED fit results, shown in Table \[err1\]. This includes the Bayesian-like uncertainties from the SED-fits, the mock analysis of reliability, and the external comparison with observations. In general, the uncertainties are small enough that it is possible to fit the SEDs of BCDs and LVL galaxies and reliably determine some of their physical parameters. In particular, we find that our estimates of stellar mass have a typical Bayesian uncertainty of $\sim$$0.5$ dex, but agree with empirical comparisons within $0.14$ dex. Similarly our estimates of SFR on various time scales have formal Bayesian uncertainties between $0.40-1.00$ dex, but show significantly better agreement with empirical comparisons of $0.25-0.30$ dex, and mock deviations that are even smaller. These SED-derived parameters are not “high precision” measurements, but are constrained well enough to allow for meaningful comparisons between the BCD and LVL samples. These uncertainty estimates are included on all subsequent plots, and demonstrate the reliability and accuracy of the parameters.
Results from SED fits {#sed_results}
=====================
Having generated best-fit parameters for the full sample of BCDs (shown in Table \[best1\]) and LVL galaxies, we can now compare these populations. To put the BCDs (from our sample and the LVL sample) into a broader context, we consider the non-BCD late-type galaxies in LVL as “normal” galaxies. We are interested in testing whether BCDs are best described as a phase of dI evolution, or whether BCDs represent a unique type of dwarf galaxy. While some of the SED-derived parameters are difficult to accurately and reliably determine, our fitting method is uniformly applied to the observations of the BCDs and LVL galaxies, so we can study the statistical trends between the two populations in a differential sense. In the following section we discuss the results of our SED fitting and the implications they have on the evolutionary context of BCDs.
{width="47.00000%"} {width="47.00000%"}\
Stellar Mass vs SFR relation {#MSFRsec}
----------------------------
Figure \[MSFR\] shows two versions of the M$_\textrm{stellar}-$SFR diagram, where the the late-type LVL galaxies populate the so-called “main sequence” of star-forming galaxies. We show measurements of the SFR from the SED fit averaged over 50 Myr and from [H$\alpha$ ]{}observations. As found by previous studies [e.g., the shaded area from @peng10], there is a relatively tight relationship between SFR and stellar mass. The BCDs lie at (and above) the upper extreme of this relationship, in the sense that they have larger SFRs than most other galaxies of their stellar mass.
The relatively tight M$-$SFR relationship is typically interpreted as evidence of a universal mode of star formation in star-forming galaxies, where the global amount of star formation in a galaxy is regulated (in some way) by its stellar mass. This strong correlation between SFR and M$_\textrm{stellar}$ is seen across five orders of magnitude of mass. As discussed more extensively in @cook14b, there is good agreement in the overlap between the relationship from the late-type LVL galaxies and that of @peng10. The early-type LVL galaxies are typically offset below the best-fit line, as they have lower SF activity than comparable late-type galaxies. The LVL sample extends to significantly lower stellar masses than the SDSS sample, and confirms that the star-forming main sequence relationship continues at least down to M$_\textrm{stellar}$ $\sim$$10^6M_\odot$, although it may become broader at the lower masses.
As seen in Figure \[MSFR\], regardless of whether SFR is measured via the 50-Myr average from SED fits or from [H$\alpha$ ]{}fluxes, the M$-$SFR relationship for BCDs is offset from that of normal galaxies. When considering SFR$_\textrm{{H$\alpha$}}$ measurements, the BCDs are more than an order of magnitude offset from the typical relationship, in the sense that BCDs have $\sim$$10$ times more star formation per unit stellar mass. When using $\langle$SFR$\rangle_{50}$, the offset between the BCDs and normal galaxies is smaller, but is still nearly an order of magnitude. Regardless of the implicit assumptions in different SFR indicators, BCDs show SFRs elevated above the star-forming “main sequence”, indicating that BCDs are experiencing truly exceptional amounts of star formation. As their [H$\alpha$ ]{}SFRs are elevated beyond the SED SFRs (based on UV fluxes), the current SF in BCDs may also be especially recent.
Evidence of past star formation episodes {#pastSF}
----------------------------------------
To test whether the BCDs or the LVL galaxies show any evidence of past bursts of star formation, we can compare the SFR estimates averaged over multiple time scales. Evidence of past bursts would support the possibility that other dwarf galaxies may have experienced BCD-like star formation events, and may help to identify BCDs which are no longer bursting. However, the time-averaged SFR estimates have substantial uncertainties (typically $\sim$$0.5$ dex, see Table \[err1\]), making this comparison difficult. We calculate a $\chi^2$-style difference between each SFR estimate averaged over time $X$ and the total average SFR for each galaxy, as shown: $$\frac{SFR_X - SFR_\textrm{all}}{\sigma_{SFR_X}}$$ where the absolute difference is normalized by the Bayesian uncertainty on the average SFR value. This normalization prevents poorly-determined SFR estimates from appearing as deviant results. This SFR difference is calculated for the 50, 100, 500, and 1000 Myr time-averaged SFR estimates, and is shown in the panels of Figure \[dSFR\]. The BCDs show a systematic offset toward less-deviant SFR estimates on most time scales, while LVL galaxies have typically more significant deviations. On the longest time scales (1000 Myr average SFR), the BCDs show a much tighter distribution of differences than the LVL galaxies, possibly implying that BCDs have had less variation in their star formation histories. Alternatively, this lack of deviations may mean that their current star formation event has generated a substantial fraction of their total stellar mass. It is difficult to accurately compare these SFH indicators as they all have large uncertainties.
{width="7.4cm"} {width="7.4cm"}\
{width="7.4cm"} {width="7.4cm"}
In order to test whether the decreased scatter of the BCDs in the SFR deviations of Figure \[dSFR\] means that the SFH in the BCDs is dominated by a recent event, we compare SFR estimates in a different way. We subtract SFR estimates from two different time intervals and normalize the difference by the lifetime average SFR. For example, when comparing the 50 Myr SFR estimate with the 10 Myr estimate, the excess is found by: $$\frac{SFR_{50} - SFR_{10}}{SFR_\textrm{all}}$$ and indicates the relative intensity of SF activity in the interval between 10 and 50 Myr ago, compared with the overall SFR. We calculate this quantity for four age-pairings, as shown in the panels of Figure \[dSFRd\]. It appears that the SFR excess on the shortest time scales (between 50 and 10 Myr ago, shown in Figure \[dSFRd\]a, is relatively small, as expected since our broadband SED-derived SFR estimates are not sensitive to star formation on such small time scales. The deviations on 100-50 Myr time scales are the largest (Figure \[dSFRd\]b), and this is also a time region with good sensitivity in our SED fits. When comparing the 1000-500 Myr SFR estimates (Figure \[dSFRd\]d), the differences are relatively small. The lack of differences on the longest time scales may be a result of averaging over more star forming events, or may be a result of the coarseness of our SFR indicators for older stellar populations.
**(a) (b)**\
{width="7.4cm"} {width="7.4cm"}\
**(c) (d)**\
{width="7.4cm"} {width="7.4cm"}\
In general, Figure \[dSFRd\] demonstrates that our SED fits are somewhat sensitive to variations in SFH over modest (50-500 Myr) time scales. Most of the deviations shown in Figure \[dSFRd\]b, between the 100 and 50-Myr averaged SFRs, are positive in the sense that $SFR_{100} > SFR_{50}$. This is complemented by the reverse trend in Figure \[dSFRd\]c, where the differences between the 500 and 100 Myr averaged SFRs are usually negative, so $SFR_{500} < SFR_{100}$. This means in those deviant galaxies, the star formation between 50 and 100 Myr ago was significantly higher than average, while between 100 and 500 Myr, it was below average. Some of these BCDs and LVL galaxies with deviant values are discussed individually in Section \[individuals\]. This kind of SFH diagnostic, while crude, may be the first step toward identifying systems with evidence of past episodes of intense star formation, or possibly finding post-burst BCDs.
Burst strength {#bstrength}
--------------
As part of its SED-fitting process, CIGALE determines the “burst fraction” ($f_\textrm{burst}$) for each galaxy, defined as the initial mass fraction of the young stellar population. More explicitly, this is the mass fraction of the young stellar population compared to the total stellar mass, *which are both taken at the time of the young population’s birth.* The present-day mass ratio of the young stellar population will always be smaller than the initial $f_\textrm{burst}$ as the young high mass stars are perishing at a faster rate than the older stellar populations.
Figure \[bfracM\] shows both the initial $f_\textrm{burst}$ and the present-day young stellar population mass ratio as functions of total stellar mass. There is a substantial overlap between the late-type LVL galaxies and BCDs in terms of their initial $f_\textrm{burst}$ values (left panel), which may mean that BCDs and non-BCDs may have experienced similarly strong starbursts in their past. However, BCDs with lower stellar masses typically have higher initial $f_\textrm{burst}$ values, which is consistent with a single star formation event having a more significant effect on a lower mass galaxy. One of the BCD/E galaxies, Mk 900, has $f_\textrm{burst}=0.28$ and is an exception to this trend. Its young stellar population is the oldest among the BCD sample fits (age$_y$=2 Gyr), so its burst strength is also the most extreme. Note that its young population mass ratio ($M_{\star,y}/M_\star$) is $\sim$$10^{-3}$, which is typical for other galaxies of its mass.
While there is overlap between BCDs and late-type LVL galaxies, the BCD population is offset higher than normal galaxies, suggesting that the star forming events (in terms of the mass involved) may be more substantial in BCDs.
When considering the young stellar population mass ratio (shown in the right panel of Figure \[bfracM\]), the BCDs more clearly populate the extreme of the parameter space. While BCDs and non-BCDs galaxies may have had more similar values of initial burst strengths (left panel), the mass fraction of today’s young stellar population is much greater in BCDs than in LVL galaxies (right panel). Much of this difference is due to the age of their young stellar populations. BCDs have a significant young star-forming population, while the stellar populations in non-BCDs have had more time to age and fade. Again there is a trend for BCDs with lower stellar masses to have larger young population mass fractions.
Combined, the two panels of Figure \[bfracM\] suggest that the young stellar populations of BCDs are different from normal late-type galaxies today. Their SED-derived burst strength does not separate them as clearly, but their young population is usually more significant. Furthermore, the exponential scale times associated with the young and old stellar populations can vary, which can add to the differences between these comparisons.
Further insight can be gained by combining these two estimates of burst strength into Figure \[bfrac\], which shows $f_\textrm{burst}$ as a function of the young stellar population mass ratio. The points on Figure \[bfrac\] are color-coded by their young population burst age (age$_y$) divided by its exponential scale time ($\tau_y$). Blue points show galaxies with young populations younger than $1.5\tau$, which are very recent bursts. Green points show middle-aged bursts, and red points show more evolved bursts. We plot all of the BCDs as dots (from our sample and the LVL sample) and the LVL late-type galaxies as circles. We only show galaxies from LVL in the stellar mass range (M$_\textrm{*}<10^9$[M\_]{}), to match the BCD sample.
![Comparison between best-fit $f_\textrm{burst}$ values and estimates of the current mass ratio of the young stellar population. Each galaxy is plotted with a shape and color that correspond to the age of its young burst population, in units of its scale time, $\tau$. Galaxies with bursts of age/$\tau$$<$1.5 are shown in blue, middle-aged bursts are shown in green, and older bursts are shown in red. Solid dots show BCDs (from our sample and LVL) and open circles show late-type LVL galaxies, selected to be in the same stellar mass range as the BCDs. The dashed diagonal line shows the unity line, which indicates a very recent burst. The galaxy with the highest $f_\textrm{burst}$ value is UGCA 281, an LVL galaxy classified as a BCD and discussed further in Section \[bstrength\]. The galaxy with the highest M$_{*,y}$/M$_*$ is I Zw 18. In general, BCDs are characterized by having both strong burst strength and a significant young population at present day, although there is substantial scatter. \[bfrac\]](BfracM-crop.eps){width="47.00000%"}
If the young population bursts were extremely recent, the points would lie very close to the equality line (shown as a dashed diagonal line, where f$_\textrm{burst}$=M$_\textrm{*,y}$/M$_\textrm{*}$). As the young population evolves, the points move horizontally in this plot as the young stellar population mass ratio decreases (while its initial burst fraction always remains the same). This is seen on the plot as the points closest to the 1:1 line have younger bursts (i.e. are blue points), although there is significant scatter beyond this simple trend due to different age$_y$/$\tau_y$ parameters of the SFHs. Compared with the late-type LVL galaxies, on average the BCDs show higher values of both $f_\textrm{burst}$ and young population mass ratios, as expected.
With this relationship in mind, it appears that the star formation in BCDs is both substantial and typically more recent (than LVL). Note that some of the late-type galaxies in LVL have similar young populations to BCDs. The requirements of recent and substantial young stellar population are necessary, but not sufficient, criteria for a galaxy to be a BCD.
{width="\columnwidth"} {width="\columnwidth"}\
{width="\columnwidth"} {width="\columnwidth"}\
It is interesting to note that Figure \[bfrac\] also shows a population of LVL galaxies with moderate initial $f_\textrm{burst}$ (e.g., $>0.1$) but with low present-day young population mass fraction (e.g., $< 0.5\%$). These characteristics are similar to what we would expect from a population of post-burst BCDs. When the young populations in those galaxies with high $f_\textrm{burst}$ first started forming stars, by definition their young population mass fraction would have been equivalent to $f_\textrm{burst}$. While we do not have any objects in our sample with young population mass fractions as extreme as $10\%$, galaxies in this state would unambiguously be considered starbursting systems (and possibly BCDs).
Figure \[bfrac\] can also be used to crudely estimate a duty cycle on the bursting phase of BCDs (more specifically, in terms of the amount of young stellar mass involved in the burst), based solely on assumptions about the population demographics of starbursts in this sample. We must also assume that $f_\textrm{burst}$ represents a single burst, and not multiple star formation events which are blended together into a single representative population in the SED fit. With these simplifying assumptions, we can pick a typical value of $f_\textrm{burst}$ and ask how many galaxies currently have a current young population mass fraction that large. For example, picking a set of objects with $f_\textrm{burst} > 0.2$ includes 14 galaxies (from both the LVL and BCD samples), only one of which has a present-day young stellar population with $> 0.5\%$ the total mass. The largest young population mass fraction of this set of objects is only $4\%$, and the average young mass fraction is $0.4\%$. Clearly most of these objects show evidence of significant starbursts in their past, but only $1$ in $14$ is experiencing a similarly strong burst at the present day. Figure \[bfrac\] also shows the SED-determined ages of the burst population by the colored rings around points, which can help further constrain the duty cycle of these starbursts. The points nearest to the 1:1 line are more often young bursts (with age$_\textrm{y} < 500$ Myr), while the points in the middle are typically middle-aged bursts, and the points with low present-day young population mass fractions are typically old bursts (with age$_\textrm{y} > 1000$ Myr). In the set of objects with $f_\textrm{burst} > 0.2$, there is a correlation between burst ages and young stellar mass fractions, in the sense that older bursts have significantly smaller young stellar mass fractions. It appears that most of the galaxies with the highest $f_\textrm{burst}$ had their bursts $\gtrsim 1$ Gyr ago, and have a very weak young stellar population today. These types of systems with historically strong but currently faded bursts may be interesting objects to followup with searches for post-burst BCDs.
{width="45.00000%"} {width="45.00000%"}
The young population burst ages of the BCDs (shown in Table \[best1\]) are somewhat larger than might be expected. Some of the most extreme BCDs have young population ages $< 500$ Myr (e.g., I Zw 18 and Mk 36), but many have young population ages between 500 and 1000 Myr, and some even $>1000$ Myr. However, based on the strong [H$\alpha$ ]{}emission from the BCDs, it is unlikely that their young populations can be this old. Instead, this over-estimate of burst age demonstrates a limitation of our SED fits. The broadband fluxes used in the SED fit are most sensitive to star formation on 100-1000 Myr timescales, but only crudely and with low resolution. Even if the BCDs have had recent (50-100 Myr) star formation events, the SED fits are unable to resolve SFH changes on such short timescales. In this way, the SED fits are almost certainly systematically under-estimating the current star formation rates of galaxies which have recently started forming stars, which is consistent with the behavior seen in Figure \[MSFR\].
To look for correlations between these burst strength indicators and the current amount of star formation in these galaxies, Figure \[fracSFR\] shows the values of $f_\textrm{burst}$ and the young stellar population mass ratio as functions of SFR estimates averaged over 50 and 500 Myr. As was the case with Figures \[bfracM\] and \[bfrac\], the BCDs are more distinctive in terms of their young stellar population mass ratio than their initial burst strengths. In contrast with the comparison between burst strengths and stellar mass, however, none of these indicators appear to be correlated with SFR. **\[Jessica says think about SSFR as well, in looking for correlations here.\]** This suggests that the absolute amount of SFR in a BCD or LVL galaxy is not a defining characteristic on its own, but that intense (in terms of mass fraction) starbursts can happen in galaxies with high or low star formation rates. An absolute SFR cutoff will not be useful in identifying BCD-like galaxies which may experience a different mode of star formation than typical star-forming galaxies on the “main sequence”.
As an alternative parametrization of the significance of the current star formation episode, we also calculate the birthrate parameter of @kennicutt83, defined as: $$b = \frac{\textrm{SFR}}{\langle\textrm{SFR}\rangle}$$ where $b$ is the birthrate parameter, SFR is an estimate of the current SFR, and $\langle$SFR$\rangle$ is the total lifetime average SFR. This parameter is related to the specific star formation rate and also to the young stellar population mass fraction and $f_\textrm{burst}$, but now parametrizes the starburst in terms of its star formation rate instead of the mass of the young stellar population .
We make two versions of $b$, one calculated using the most recent 50-Myr averaged SFR ($b_{50}$) as the numerator and one with the SFR determined from [H$\alpha$ ]{}observations ($b_\textrm{{H$\alpha$}}$) as the numerator. In both cases, the denominator is the lifetime average SFR from the SED fits (closely related to the total stellar mass). These birthrate parameters are plotted as a function of stellar mass in Figure \[bSFR\]. As discussed previously, the [H$\alpha$ ]{}SFR is more sensitive to recent intense star formation, while the SED SFR probes longer time scales.
Note that these birthrate parameters are not measuring the stellar mass involved in the star formation event, but represent the significance of the current star formation in comparison with a galaxy’s total lifetime of star formation. For a particular galaxy, the birthrate parameter describes the significance of its current star formation in its long-term evolution. A galaxy with a birthrate parameter of $b \ge
3$ is typically required to be considered a starburst [e.g., @bergvall15 and references therein]. In the smallest dwarf galaxies, the effects of star formation can be more significant (i.e., the effects of feedback on galaxy-wide scales) than the same absolute amount of star formation in larger galaxies. The birthrate parameter gives a good indication of the significance of the current SFR. Starbursting galaxies with $b > 3$ typically are undergoing a truly transformative event, distinct from the low-level fluctuations in SFR which are typical in dwarf galaxies.
In Figure \[bSFR\], the $b$ values for the BCDs show the correlation with stellar mass in the sense that lower stellar masses have larger $b$ values, in the same way as $f_\textrm{burst}$ and the young population mass fraction increased at lower stellar masses. This behavior is expected if the starbursts are similar in size, but can have a greater impact on a lower mass galaxy than on a higher mass galaxy. When comparing the $b$ values for the BCD and LVL samples, the BCDs have values which are at or beyond the maximum values from the LVL non-BCD samples. In particular, one LVL galaxy also populates this extreme area: UGCA 281, which has long been classified as a BCD [@tm81; @lelli14]. On the other hand, some members of our BCD sample (notably the three most massive galaxies: Mk 324, Mk 328, and Mk 900) populate the more normal areas of parameter space. The early-type galaxies from the LVL sample (red triangles) and the BCD/Es (cyan triangles) have some of the lowest values. These are also at somewhat higher masses, so a star formation event may have a smaller impact on their evolution.
Whether measured by the 50-Myr average from SED fits or by [H$\alpha$ ]{}fluxes, the birthrate parameters for BCDs are extreme, and show that the current star formation in BCDs is proceeding at a level which is not typical for normal star-forming galaxies.
Discussion {#sed_discussion}
==========
While our SED fits can only generate coarse estimates of key physical parameters (e.g., M$_\star$, SFR) for the complete LVL sample and our BCDs, they still provide a systematic and meaningful way to compare the BCDs with normal LVL galaxies. Our SED fits have shown that BCDs are forming stars at exceptionally intense rates and may populate an elevated relation parallel to the “star formation main sequence.” However, the SED fits cannot provide the detailed star formation histories which would be needed to clearly identify both past and present intense (but unsustained) levels of star formation in order to separate BCDs (both past and present) from normal dIs. Much of the detailed information about a galaxy’s SFH is not accessible through SED fits alone, but can only be probed via studies of resolved stellar populations [e.g., @mcquinn15c]. However, our SED fits can still be used to discuss and constrain the possible evolutionary connections between dwarf galaxies.
SFR and stellar mass indicators at the extremes {#extremeMZ}
-----------------------------------------------
Most of the empirical SFR and stellar mass indicators [e.g., @kennicutt98; @McGaughSchombert14] have been calibrated using galaxies with large stellar masses (M$_\star$$\approx$M$_\textrm{MWG}$), with modest abundances ($Z$$\approx$$Z_\odot$), and star formation rates (SFR$\approx$$1$$M_\odot/$yr). More recently, some groups have been able to test whether these standard relations apply equally well to galaxies with lower masses, lower abundances, and possibly higher SFRs.
@lee09 found that for galaxies with low SFR, [H$\alpha$ ]{}measurements give systematically lower SFRs than are measured in the UV. Similar findings have been reported by other groups as well [@fumagalli11; @eldridge12; @weisz12]. We can also compare the SFRs for the galaxies in our sample, as shown in Figure \[hauv\], and plot SFR from [H$\alpha$ ]{}against SFR from UV. The best fit relation from @lee09 is also shown, and matches the same general trend as the LVL data.
![This figure shows the correlation between observed SFRs from [H$\alpha$ ]{}fluxes and UV fluxes (corrected for internal extinction). LVL galaxies are shown in black, and BCDs in blue. In black is the equality line. The offset between BCDs and LVL galaxies is likely due to assumptions in the empirical conversion factors about SFH, which the BCDs do not obey. The red line shows the best fit relation from @lee09. \[hauv\]](SFRSFR_hauv2-crop.eps){width="0.90\columnwidth"}
The BCDs appear offset from the main population in Figure \[hauv\], which is similar to the offsets previously seen in the SFR comparisons (e.g., Figure \[sfr\]). As mentioned in Section \[external\], the BCDs have systematically larger SFR$_\textrm{{H$\alpha$}}$ than $\langle$SFR$\rangle$$_{50}$, while their SFR$_\textrm{UV}$ and $\langle$SFR$\rangle$$_{100}$ estimates are in better agreement. A similar offset in the BCD SFR$_\textrm{{H$\alpha$}}$ estimates is seen here in Figure \[hauv\].
A possible explanation for this offset lies in the calibration of these SFR indicators. The empirical SFR estimates used to generate SFR$_\textrm{UV}$ and SFR$_\textrm{{H$\alpha$}}$ both assume a constant SFR. In the FUV, the flux comes from young stars with lifetimes $\sim$$10^8$ years, and correspondingly measures the SFR on those time scales. The [H$\alpha$ ]{}flux comes from gas ionized by massive O-stars and early-type B-stars, with shorter lifetimes than those providing UV flux. As such, the [H$\alpha$]{}-derived SFR probes shorter time scales.
@lee09 argue that the change of this relationship at low SFRs may be due to stochastic variations in the SFH of low mass galaxies, where the UV SFR averages over longer time scales and will typically be larger than the [H$\alpha$ ]{}SFR in the case of a bursty SFH. As shown in Figure \[hauv\], the BCDs (from our sample and from within LVL) nearly always have larger [H$\alpha$ ]{}SFRs than UV SFRs. This suggests that the BCDs may be in a phase of their SFH where intense star formation has only recently begun, and is better traced by the [H$\alpha$ ]{}SFR than by UV. The best-fit SFRs from the SED fits are driven by the UV emission, and as such are not sensitive to the recent star formation which produces [H$\alpha$ ]{}emission. If the BCDs have had a recent enhancement in SF activity, the UV SFR measurement could be diluted by the lower SF which may have preceded the current event.
Outliers and exceptional galaxies {#individuals}
---------------------------------
The comparisons from Section \[external\] between empirical estimates of stellar mass (via $3.6\mu$m and K$_s$ luminosity) and SFR (via [H$\alpha$ ]{}and UV luminosity) with those from the SED fits were primarily intended as consistency checks of the SED fitting process. However, they can also be used to test the reliability of different stellar mass and SFR estimators for particular objects. For most galaxies in our sample, the SED fitting gives SFR and stellar mass values consistent with the empirical methods. However, the galaxies with inconsistent values require further investigation. There can be many reasons for individual galaxies to deviate from the normal relationships: they may be intrinsically unusual objects in a brief evolutionary stage or an uncommon region of parameter space, or may be examples of objects which have SEDs that are not well fit by our methods. Here we look at individual objects from the previous figures as examples of the types of outliers in this work.
As shown in Section \[external\], the SED-derived stellar masses are generally in good agreement with the empirical estimates from $3.6\mu$m and K$_s$ luminosity measurements. We have identified only very slight offsets or non-unity slopes in these relationships. On the $3.6\mu$m plot in the left panel of Figure \[M\], there are no significant outliers from the distribution of points. On the right panel of Figure \[M\] which shows the stellar mass from K$_s$ photometry, the object furthest above the relation is a BCD (UM 462) and the object furthest below the relation is an LVL galaxy (UGC 01056). UM 462 has a SED-derived stellar mass $\sim$$50\%$ larger than its K$_s$ mass, and UGC 01056 has an SED-mass $\sim$$50\%$ smaller than its K$_s$ mass. These deviations are still within the Bayesian estimates of uncertainty on the SED fits, and represent the most deviant mass estimates.
In our comparison of SFR estimates shown in Figure \[sfr\], some galaxies are significantly deviant from the overall relationship. In the comparison between $\langle$SFR$\rangle$$_{100}$ and SFR$_\textrm{UV}$ (right panel), one galaxy (Sextans A) is a significant outlier **\[quantify/??\]**. We now consider what made that outlier drift so far from the relationship. Sextans A has similar values of SFR$_\textrm{{H$\alpha$}}$ and $\langle$SFR$\rangle$$_{50}$ (within 0.4 dex), and it is unclear what has caused its UV and SED SFR estimates to deviate. Its other best-fit parameters are fairly typical, and its SED fit has a reduced $\chi^2 = 3.23$. However, given its diffuse nature and large extent, it is possible that some of the photometry has not been measured carefully enough. **\[Look at Dohm-Palmer2002, Dolphin2003 papers?\]** The scatter in the comparison between SED SFRs and [H$\alpha$ ]{}SFRs is somewhat larger, but may be intrinsic to any [H$\alpha$ ]{}SFR estimate, as discussed earlier and in @lee09.
**\[Jessica says move this to SFR discussion area. also clarify what “not measured carefully enough” means. quantify!\]**
Throughout this analysis it has become clear that some members of our BCD sample are less extreme than others. In particular, Mk 324, Mk 328, and Mk 900 (labeled as “BCD/E” on the figures) typically have properties more similar to the average LVL late-type galaxies. Similarly, @janowiecki14 found that these three were the reddest (B$-$V$\sim$$0.8$ in their outskirts), most luminous, and most metal-rich objects in this BCD sample. The SED fits have shown that these three also have the largest stellar masses (M$_{\star}$$>$$4 \times 10^8 M_\odot$) and the lowest specific star formation rates (SSFR$ < 3 \times
10^{-10}$ /yr) of the BCD sample. Given the dissimilarities between the three objects and the rest of the BCD sample, we advise caution when drawing conclusions about BCDs based on the inclusion of these three objects. They have been included in our sample to cover the broad range of objects commonly referred to as BCDs, but may be more intermediate between extreme BCDs and more typical dwarf irregular galaxies.
Significance of starbursts {#starburst}
--------------------------
With the exception of mergers, galaxies rarely evolve faster than when they are experiencing intense star formation events such as those happening in the BCDs. In order to understand where BCDs (and BCD-like) galaxies fit into dwarf galaxy evolution, we must understand the transformative effects of their active star formation. The term “starburst” is often used to describe a variety of galaxies experiencing star formation events. In different parametrizations, the star formation can be measured in terms of how much higher the SFR is compared to average, or in terms of how much stellar mass the current episode will produce compared to the total stellar mass. Each method will identify a different type of starbursting galaxy. In reality, there is a broad continuum of galaxies with different intensities of star formation, which will have differing impacts on their ongoing evolution.
In this work, we are particularly interested in the transformative effects that intense star formation events can have on dwarf galaxies. This can be discussed in terms of the burst fraction ($f_\textrm{burst}$), which is the mass fraction of the star-forming population when it first formed, as shown in Figure \[bfrac\]. This graph also shows the present-day mass fraction of the star-forming population, which gives an indication of how much of a role the star-forming population still plays in the galaxy. In this parametrization, the BCDs occupy an extreme of parameter space with exceptionally large $f_\textrm{burst}$ (as high as $0.3$) and large mass ratios of the burst population. Some of the BCDs even lie close to the region of parameter space where $f_\textrm{burst}$ and the young population mass ratio are equivalent, which would indicate a very young burst. BCDs are not just sites of young bursts – they are also exceptional in their large values of $f_\textrm{burst}$. It appears from Figure \[bfrac\] that as galaxy evolution proceeds, the points will move to the left (same value of initial $f_\textrm{burst}$, but ever-decreasing young population mass fraction as younger stars perish). In that way, the stellar populations of the BCDs may become less distinctive from the late-type LVL galaxies as their young populations continue to age.
We can also parametrize the impact of the star formation population in terms of its initial mass fraction, at the time it first started forming stars. This birthrate parameter is discussed at the end of Section \[bstrength\] and is a ratio of the current (or recent) SFR to the total average SFR. Figure \[bSFR\] shows the relationship between $b_{50}$ and $b_\textrm{{H$\alpha$}}$ (using the SFR averaged over the most recent $50$ Myr and the SFR from [H$\alpha$ ]{}observations) and the stellar mass. In this parametrization, the BCDs are at the extreme edge of a very broad parameter space. For their mass, BCDs have values of $b_{50}$ and $b_\textrm{{H$\alpha$}}$ at or above the highest values of the LVL sample. Some LVL galaxies also populate this extreme area, notably UGCA 281, which has long been classified as a BCD [@tm81; @lelli14]. On the other hand, some members of our BCD sample (notably Mk 324, Mk 328, and Mk 900) populate the more normal areas of parameter space. This parametrization is a useful way to measure the impact of the current star formation in terms of the galaxy’s star formation history.
In a recent paper, @bergvall15 make a case for using the term “starburst” with care, and only in situations with extreme burst strengths or birthrate parameters. This distinction matters especially when considering the global effects of star formation on galaxies, and trying to identify the galaxies which are being transformed the most rapidly by their current star formation episodes. In the smallest dwarf galaxies, the effects of star formation can be more significant (e.g., feedback) than in larger galaxies, so the absolute amount of star formation is not the relevant parametrization. By setting a threshold of $b \ge 3$, @bergvall15 insure that the galaxies they consider starbursting are experiencing a truly transformative event, distinct from the low-level fluctuations in SFR which are typical in dwarf galaxies.
Deviations in SFH: what does it mean? {#deviations}
-------------------------------------
In order to understand the evolutionary role that BCDs play in dwarf galaxy evolution, information about their past star formation history is essential. Today, studies of resolved stellar populations in large numbers of dwarf galaxies are providing increasingly precise estimates of their SFHs [e.g., @weisz11; @mcquinn15a]. These SFHs are showing that most dwarf galaxies have had similar average SFH over their lifetimes (although somewhat less than the cosmic SFR), and much of their differentiation may have occurred in the most recent few Gyrs, typically as a result of environmental effects [@weisz11] **\[check - is this for all masses? there is a mass dependence on environmental effects?? otherwise these are sort of contradictory and flip-floppy\]**. However, some populations of star-forming dwarf galaxies do not show a correlation between environment and recent star formation events, suggesting that internal processes may be regulating star formation in some of the smallest galaxies [@mcquinn15a and references therein]. For large galaxies, the stellar mass (and, correspondingly, the local environment) may be the primary regulator of star formation (and merger) events. However, smaller star-forming dwarf galaxies are typically more isolated and their star formation may be more strongly regulated by internal processes.
In Section \[pastSF\] we explored whether the results from the SED fits can crudely constrain the SFH of the BCDs and LVL galaxies. While much more coarse than the full resolved stellar population fitting, the SED fit results can still produce broad constraints on SFH. We use the SFR averages over different time scales as approximations to the complete SFH, and discuss some of the individually deviant galaxies in Section \[individuals\]. While the BCDs do not clearly distinguish themselves from the LVL population in these diagnostic plots (e.g., Figures \[dSFR\] and \[dSFRd\]), they are among the extreme objects in those parametrizations. The SED fits are helping to identify galaxies with apparently unusual deviations in their SFH on particular time scales, which may require individual followup study in order to understand their roles in dwarf galaxy evolution.
Future evolution of BCDs
------------------------
With their short gas depletion time scales and rapid star formation rates, BCDs are in an unsustainable evolutionary state, and cannot continue forming stars at their current rate. Many groups have studied or discussed the implications of a population of faded or post-burst BCDs [@sanchezalmeida08; @sanchezalmeida09; @amorin12; @lelli12; @koleva13; @janowiecki14; @meyer14]. Here it is important to separate possible evolutionary pathways based on whether external or internal effects are being considered.
@meyer14 consider the long-term evolution of BCDs in the Virgo Cluster, and find that BCDs may fade into galaxies similar to the extant population of early type dwarfs in the Virgo Cluster. They further suggest that in the cluster environment, a compact early type dwarf may have its star formation re-ignited if it acquires fresh gas. In lower density environments, @sanchezalmeida08 searched for BCDs in quiescence (QBCDs) by selecting galaxies with compact structural parameters and other BCD-like properties, but without the current intense star formation event. Within their search criteria, they found a very large population of QBCDs, and suggest that perhaps 1/3 of all local dwarf galaxies may be capable of hosting a BCD-like burst. Whether such intense starbursts are this common is still an open question.
In this work, some of the SED-derived parameters for the BCDs and LVL galaxies will evolve with time, while others will not. For example, the initial burst fraction ($b_\textrm{frac}$) of the current star star forming population will not decrease with time, while the young population mass fraction will continue to decline. Similarly, the total stellar mass will not change quickly, while the color will shift from blue to red as the hottest youngest stars perish. We can generate model predictions for the future evolution of BCDs, but this requires that we assume there are no external effects driving this evolution.
In the simple case of internal evolution only, the stellar populations of the BCDs would age normally, and follow the characteristic isochrones for older ages. The resulting objects may resemble the so-called “postburst” galaxies [or E+A, k+a, a+k, @dresslergunn83; @dressler99]. These galaxies are characterized by strong Balmer line absorption (from the most massive stars still alive, the A stars), but no [H$\alpha$ ]{}emission (as the O and B stars have perished). Other groups have used the $4000$Å break [@kauffmann03] or star formation histories from resolved stellar populations [@mcquinn10a; @mcquinn10b] to identify galaxies which have experienced strong star formation events. Further followup on individual post-BCD-like galaxies is necessary to explore this connection further.
Summary {#sed_summary}
=======
In this work we have used panchromatic observations (from UV to FIR) of a large sample of 258 LVL galaxies and 18 BCDs to generate and fit SEDs in order to determine key physical parameters about these galaxies. These fits have allowed us to explore the role that BCDs play in the larger context of dwarf galaxy evolution. In particular, our main results are as follows:
- Our SED fits are able to reliably and robustly estimate key physical parameters of these galaxies, including M$_\star$ and various SFRs.
- SED model grids must be chosen to be broad and well-sampled enough to cover appropriate parameter space or else the best-fit values may have systematic offsets or errors.
- Two-burst stellar population models fit the observed SEDs significantly better than one-burst models, suggesting that most or all dwarf galaxies have an old stellar population.
- While powerful, the inability of SED fits to return a detailed SFH makes it difficult to answer questions about the evolutionary histories of individual BCDs.
- When using [H$\alpha$ ]{}SFRs in place of SED SFRs, the BCDs appear even more extreme than typical galaxies, suggesting that the SF activity in BCDs is uniquely intense or recent, and that SED fits are not sensitive to the recent star formation traced by [H$\alpha$ ]{}emission.
- We have identified unusual and potentially extreme objects for further study, which may represent brief evolutionary stages in dwarf galaxy evolution.
Further work is needed to more completely explore the evolutionary relationships between LVL galaxies and BCDs, and to connect the SED-fit results with the compactness estimates of @janowiecki14. Once available, the optical surface photometry of the LVL galaxies will provide an invaluable resource for a more complete comparison between the BCDs and LVL galaxies. We will be able to determine how extreme the BCDs are both in terms of the structural parameters of their underlying hosts and in terms of their stellar populations. BCDs appear to represent an extreme class of dwarf galaxy with physical properties that place them at the edges of the broad continuum of dwarf galaxy properties.
Acknowledgments
===============
SJ is grateful to D. Cook for providing the LVL SEDs and metallicities in a convenient format, and to S. Salim and H. Evans for useful discussions and advice. We thank the anonymous referee for helping to improve, clarify, and expand the context of this work. All SED analysis was run on a laptop purchased with funds from a McCormick Science Grant from the Indiana University College of Arts and Sciences. Some results from this work were also presented in @janowiecki15. Support is acknowledged from GALEX Cycle 3 program GI3-089. SJ acknowledges support from the Australian Research Council’s Discovery Project funding scheme (DP150101734).
This research has made use of NASA’s Astrophysics Data System Bibliographic Services. This research has also made use of the NASA/IPAC Extragalactic Database (NED), which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.
Funding for SDSS-III has been provided by the Alfred P. Sloan Foundation, the Participating Institutions, the National Science Foundation, and the U.S. Department of Energy Office of Science. The SDSS-III web site is http://www.sdss3.org.
SDSS-III is managed by the Astrophysical Research Consortium for the Participating Institutions of the SDSS-III Collaboration including the University of Arizona, the Brazilian Participation Group, Brookhaven National Laboratory, University of Cambridge, Carnegie Mellon University, University of Florida, the French Participation Group, the German Participation Group, Harvard University, the Instituto de Astrofisica de Canarias, the Michigan State/Notre Dame/JINA Participation Group, Johns Hopkins University, Lawrence Berkeley National Laboratory, Max Planck Institute for Astrophysics, Max Planck Institute for Extraterrestrial Physics, New Mexico State University, New York University, Ohio State University, Pennsylvania State University, University of Portsmouth, Princeton University, the Spanish Participation Group, University of Tokyo, University of Utah, Vanderbilt University, University of Virginia, University of Washington, and Yale University.
Alam, S., Albareti, F. D., Allende Prieto, C., et al. 2015, , 219, 12
Aloisi, A., Clementini, G., Tosi, M., et al. 2007, , 667, L151
Amor[í]{}n, R., P[é]{}rez-Montero, E., V[í]{}lchez, J. M., & Papaderos, P. 2012, , 749, 185
Ashley, T., Elmegreen, B. G., Johnson, M., et al. 2014, , 148, 130
Baldwin, J. A., Spinrad, H., & Terlevich, R. 1982, , 198, 535
Bauer, A. E., Hopkins, A. M., Gunawardhana, M., et al. 2013, , 434, 209
Bekki, K. 2008, , 388, L10
Bell, E. F., & de Jong, R. S. 2001, , 550, 212
Berg, D. A., Skillman, E. D., Marble, A. R., et al. 2012, , 754, 98
Bergvall, N., Marquart, T., Way, M. J., et al. 2015, , accepted, arXiv:1501.06928
Bordalo, V., Plana, H., & Telles, E. 2009, , 696, 1668
Brinchmann, J., Charlot, S., White, S. D. M., et al. 2004, , 351, 1151
Brinchmann, J., Pettini, M., & Charlot, S. 2008, , 385, 769
Bruzual, G., & Charlot, S. 2003, , 344, 1000
Buat, V., Giovannoli, E., Takeuchi, T. T., et al. 2011, , 529, AA22
Calzetti, D., Armus, L., Bohlin, R. C., et al. 2000, , 533, 682
Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, , 345, 245
Carigi, L., Colin, P., Peimbert, M., & Sarmiento, A. 1995, , 445, 98
Cook, D. O., Dale, D. A., Johnson, B. D., et al. 2014, , 445, 899
Cook, D. O., Dale, D. A., Johnson, B. D., et al. 2014, , 445, 881 Croxall, K. V., van Zee, L., Lee, H., et al. 2009, , 705, 723
Daddi, E., Dickinson, M., Morrison, G., et al. 2007, , 670, 156
Dale, D. A., Cohen, S. A., Johnson, L. C., et al. 2009, , 703, 517
Dale, D. A., Helou, G., Magdis, G. E., et al. 2014, , 784, 83
de Barros, S., Schaerer, D., & Stark, D. P. 2014, , 563, A81
Deason, A., Wetzel, A., & Garrison-Kimmel, S. 2014, , 794, 115
Dressler, A., & Gunn, J. E. 1983, , 270, 7
Dressler, A., Smail, I., Poggianti, B. M., et al. 1999, , 122, 51
Eldridge, J. J. 2012, , 422, 794
Elmegreen, B. G., Zhang, H.-X., & Hunter, D. A. 2012, , 747, 105
Fumagalli, M., da Silva, R. L., & Krumholz, M. R. 2011, , 741, L26
Gil de Paz, A., Madore, B. F., & Pevunova, O. 2003, , 147, 29
Giovannoli, E., Buat, V., Noll, S., Burgarella, D., & Magnelli, B. 2010, SF2A-2010: Proceedings of the Annual meeting of the French Society of Astronomy and Astrophysics, 41
Guseva, N. G., Izotov, Y. I., Papaderos, P., et al. 2001, , 378, 756
Hunter, D. A., & Hoffman, L. 1999, , 117, 2789
Inoue, A. K. 2011, , 415, 2920
Inoue, A. K., Buat, V., Burgarella, D., et al. 2006, , 370, 380
Izotov, Y. I., & Thuan, T. X. 2004, , 616, 768
Izotov, Y. I., & Thuan, T. X. 1999, , 511, 639
Izotov, Y. I., Thuan, T. X., & Lipovetsky, V. A. 1994, , 435, 647
Izotov, Y. I., Thuan, T. X., & Stasi[ń]{}ska, G. 2007, , 662, 15
Janowiecki, S., & Salzer, J. J. 2014, , 793, 109
Janowiecki, S. 2015, Ph.D Thesis (Indiana University) [ProQuest, 290, 3718472](http://pqdtopen.proquest.com/doc/1712659805.html?FMT=ABS)
Johnson, R. A., Lawrence, A., Terlevich, R., & Carter, D. 1997, , 287, 333
Kauffmann, G., Heckman, T. M., White, S. D. M., et al. 2003, , 341, 33
Kennicutt, R. C., Jr. 1983, , 272, 54
Kennicutt, R. C., Jr. 1998, , 36, 189
Kennicutt, R. C., Jr., Lee, J. C., Funes, S. J., Jos[é]{} G., Sakai, S., & Akiyama, S. 2008, , 178, 247
Kewley, L. J., & Ellison, S. L. 2008, , 681, 1183
Kobulnicky, H. A., & Kewley, L. J. 2004, , 617, 240
Koleva, M., Bouchard, A., Prugniel, P., De Rijcke, S., & Vauglin, I. 2013, , 428, 2949
Lee, J. C., Gil de Paz, A., Tremonti, C., et al. 2009, , 706, 599
Lee, J. C., Gil de Paz, A., Kennicutt, R. C., Jr., et al. 2011, , 192, 6
Leitherer, C., Li, I.-H., Calzetti, D., & Heckman, T. M. 2002, , 140, 303
Lelli, F., Verheijen, M., Fraternali, F., & Sancisi, R. 2012, , 544, AA145
Lelli, F., Verheijen, M., & Fraternali, F. 2014, , 566, AA71
Loose, H.-H., & Thuan, T. X. 1986, , 309, 59
Mac Low, M.-M., & Ferrara, A. 1999, , 513, 142
Marble, A. R., Engelbracht, C. W., van Zee, L., et al. 2010, , 715, 506
Martin, D. C., Fanson, J., Schiminovich, D., et al. 2005, , 619, L1
Matteucci, F., & Chiosi, C. 1983, , 123, 121
McGaugh, S. S. 1991, , 380, 140
McGaugh, S. S., & Schombert, J. M. 2014, , 148, 77
McQuinn, K. B. W., Skillman, E. D., Cannon, J. M., et al. 2009, , 695, 561
McQuinn, K. B. W., Skillman, E. D., Cannon, J. M., et al. 2010, , 724, 49
McQuinn, K. B. W., Skillman, E. D., Cannon, J. M., et al. 2010, , 721, 297
McQuinn, K. B. W., Cannon, J. M., Dolphin, A. E., et al. 2015a, , 802, 66
McQuinn, K. B. W., Skillman, E. D., Dolphin, A., et al. 2015b, , 815, L17
McQuinn, K. B. W., Lelli, F., Skillman, E. D., et al. 2015, , 450, 3886
Meiksin, A. 2006, , 365, 807
Meixner, M., Smee, S., Doering, R. L., et al. 2010, , 122, 451
Meyer, H. T., Lisker, T., Janz, J., & Papaderos, P. 2014, , 562, AA49
Morrissey, P., Conrow, T., Barlow, T. A., et al. 2007, , 173, 682
Moustakas, J., & Kennicutt, R. C., Jr. 2006, , 164, 81
Moustakas, J., Kennicutt, R. C., Jr., Tremonti, C. A., et al. 2010, , 190, 233
Noeske, K. G., Faber, S. M., Weiner, B. J., et al. 2007, , 660, L47
Noeske, K. G., Papaderos, P., Cair[ó]{}s, L. M., & Fricke, K. J. 2005, , 429, 115
Noeske, K. G., Papaderos, P., Cair[ó]{}s, L. M., & Fricke, K. J. 2003, , 410, 481
Noll, S., Burgarella, D., Giovannoli, E., et al. 2009, , 507, 1793
Norton, S. 1997, “Structural Parameters of Blue Compact Dwarf Galaxies” (Honors Thesis). Wesleyan University
Ono, Y., Ouchi, M., Shimasaku, K., et al. 2010, , 724, 1524
Papaderos, P., Guseva, N. G., Izotov, Y. I., & Fricke, K. J. 2008, , 491, 113
Papaderos, P., Loose, H.-H., Fricke, K. J., & Thuan, T. X. 1996, , 314, 59
Peng, Y.-j., Lilly, S. J., Kova[č]{}, K., et al. 2010, , 721, 193
Pilyugin, L. S., & Thuan, T. X. 2005, , 631, 231
S[á]{}nchez Almeida, J., Mu[ñ]{}oz-Tu[ñ]{}[ó]{}n, C., Amor[í]{}n, R., et al. 2008, , 685, 194
S[á]{}nchez Almeida, J., Aguerri, J. A. L., Mu[ñ]{}oz-Tu[ñ]{}[ó]{}n, C., & Vazdekis, A. 2009, , 698, 1497
Salim, S., Charlot, S., Rich, R. M., et al. 2005, , 619, L39
Salim, S., Rich, R. M., Charlot, S., et al. 2007, , 173, 267
Salim, S., Dickinson, M., Michael Rich, R., et al. 2009, , 700, 161
Salim, S., Lee, J. C., Janowiecki, S., et al. 2016, , 227, 2
Salzer, J. J., & Elston, R. 1992, The Stellar Populations of Galaxies, 149, 482
Sargent, W. L. W., & Searle, L. 1970, , 162, L155
Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, , 500, 525
Searle, L., & Sargent, W. L. W. 1972, , 173, 25
Stark, D. P., Schenker, M. A., Ellis, R., et al. 2013, , 763, 129
Sudarsky, D. L. 1995, “Stellar Populations of Blue Compact Dwarf Galaxies” (Honors Thesis). Wesleyan University
Terlevich, R., Melnick, J., Masegosa, J., Moles, M., & Copetti, M. V. F. 1991, , 91, 285
Thuan, T. X., & Martin, G. E. 1981, , 247, 823
Tolstoy, E., Hill, V., & Tosi, M. 2009, , 47, 371
Tremonti, C. A., Heckman, T. M., Kauffmann, G., et al. 2004, , 613, 898
Valcke, S., de Rijcke, S., & Dejonghe, H. 2008, , 389, 1111
van Zee, L., Skillman, E. D., & Salzer, J. J. 1998, , 116, 1186
Verbeke, R., De Rijcke, S., Koleva, M., et al. 2014, , 442, 1830
Walcher, J., Groves, B., Budav[á]{}ri, T., & Dale, D. 2011, , 331, 1
Weisz, D. R., Dalcanton, J. J., Williams, B. F., et al. 2011, , 739, 5
Weisz, D. R., Johnson, B. D., Johnson, L. C., et al. 2012, , 744, 44
Witt, A. N., & Gordon, K. D. 2000, , 528, 799
Wright, E. L., Eisenhardt, P. R. M., Mainzer, A. K., et al. 2010, , 140, 1868-1881
Zhao, Y., Gao, Y., & Gu, Q. 2010, , 710, 663
Scatter in the M-SFR relation
=============================
We briefly consider systematic effects and biases across the M$_\star$ and SFR indicators used in this work. As a diagnostic, we start from the assumption that galaxies on the star-forming main sequence [e.g., @brinchmann04; @salim07] populate a narrow range of star formation rates at a given stellar mass. To assess the quality of of our various M$_\star$ and SFR indicators, we can compare the dispersion about that relation using each combination of the values of M$_\star$ and SFR produced in this work. In this work, our M$_\star$ estimates include those using simple mass-to-light ratios from monochromatic photometry (in both $3.6\mu$m and K$_\textrm{s}$), and from our SED fitting analysis (e.g., Figure \[M\]). Our SFR estimates include those from narrow-band [H$\alpha$ ]{}photometry and from our SED fitting (e.g., Figure \[sfr\]). In order to consistently quantify the scatter in the M$_\star$-SFR relation across different estimators, we first select LVL galaxies which are not identified as early-types or BCDs. In this type of comparison, galaxies with extreme (high or low) star formation properties will add scatter to the main sequence M$_\star$-SFR relation, and should not be included. Additionally, [@salim16] found that stellar mass determinations based on UV+optical+MIR SED fits are systematically affected at different sSFRs by $\sim0.1$ dex (see their Figure 9). We also require that galaxies in this comparison have a complete set of observations required to generate these estimates (i.e., fluxes measured at $3.6\mu$m, K$_\textrm{s}$, [H$\alpha$]{}, and a best-fitting SED with reduced $\chi^2<5$). This yields a common sample of 97 LVL galaxies which we use to compare M$_\star$ and SFR estimates.
Figure \[rms\] shows the M$_\star$-SFR relations for each combination of indicators in this common sample. The top row compares the [H$\alpha$ ]{}SFRs with our three estimates of mass (SED-fit, K$_\textrm{s}$, and $3.6\mu$m), and the bottom row shows the 50-Myr-averaged SFR from our SED fitting. The results of linear least squares fits to the SFR-M$_\star$ relations in each panel are shown, and the RMS scatters are given. Within this sample, the scatter about the fit increases from M$_\star$($3.6\mu$m) to M$_\star$(K$_\textrm{s}$) to M$_\star$(SED); and also increases from SFR([H$\alpha$]{}) to SFR(SED). The relation with the lowest scatter (0.39 dex) is SFR([H$\alpha$]{}) vs M$_\star$($3.6\mu$m), and the SED-fit quantities have the largest scatter (0.50 dex) Given that the scatter in the SFR-M$_\star$ relation increases when using SED fits, one might question whether the SED-fitting is improving the determination of SFR and M$_\star$ in this analysis. However, the goal of this project is not to simply re-measure SFR and M$_\star$ in normal main sequence star-forming galaxies using standard monochromatic indicators [for works which focus on statistical samples of star-forming galaxies, see @brinchmann04; @salim07; @salim16 and references therein]. Rather, we are adopting a consistent SED-fitting methodology in order to determine SFR and M$_\star$ even when considering more extreme galaxies which have not previously been included in calibrating M$_\star$ and SFR indicators. This approach allows us to quantify the differences between BCDs (with extreme star formation) and the population of main sequence star-forming galaxies, using consistent estimates of SFR and M$_\star$. Strong and/or recent star formation can bias monochromatic stellar mass estimates of galaxies. For example, if a dwarf galaxy undergoes a strong starburst, some of its K$_\textrm{s}$ and $3.6\mu$m luminosity will come from the young stellar populations and those monochromatic estimates of stellar mass could be systematically larger than for an otherwise similar passive (or normal star-forming) galaxy. Also note the relative positions of the four LVL galaxies classified as BCDs (black stars) on the top row of the Figure \[rms\]. In each panel those four galaxies have the same SFR([H$\alpha$]{}), but their M$_\star$ values are different. When using $3.6\mu$m or K$_\textrm{s}$ luminosity to estimate M$_\star$, these four galaxies are closer to the best-fit main sequence line for normal galaxies. However, if the full SED-fitting is used, their M$_\star$ estimates are reduced – it appears that their strong star formation events may be artificially enhancing their $3.6\mu$m and K$_\textrm{s}$ stellar masses. By including UV, optical, and IR photometry, our SED-fits produce more robust M$_\star$ and SFR estimates for our diverse sample of galaxies which have a wide range of star formation histories. This systematic and consistent approach allows us to compare BCDs with a large sample of “normal” galaxies, without a bias toward main-sequence star-forming systems.
{width="\textwidth"}
[^1]: Kitt Peak National Observatory, National Optical Astronomy Observatory, which is operated by the Association of Universities for Research in Astronomy (AURA) under cooperative agreement with the National Science Foundation.
[^2]: The WIYN Observatory is a joint facility of the University of Wisconsin-Madison, Indiana University, the University of Missouri, and the National Optical Astronomy Observatory.
[^3]: This research has made use of the NASA/IPAC Extragalactic Database (NED) which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.
[^4]: <http://wise2.ipac.caltech.edu/docs/release/allwise/>
|
{
"pile_set_name": "ArXiv"
}
|
MORE
Homo hookers busted in Atlanta police sting
The hunt for a hooker in Midtown proved a little more difficult over the weekend as Atlanta police posed as gay men seeking sex and busted scores of alleged prostitutes.
The arrests of 50 alleged male hookers along corridors long known for gay sex for sale—Cypress Street near Peachtree Street, as well as Ponce de Leon Avenue—came as the Atlanta Police Department notched 79 arrests in a days-long sting that ended Sunday.
Operation Summer Heat included 61 arrests for prostitution—50 men and 11 women, eight “johns” seeking a prostitute, and 10 narcotics arrests. In addition, seven vehicles were impounded and $1,000 was seized, according to the Atlanta Police Department. The sting started May 5, though it did not operate every day, and included the department’s undercover vice squad and Atlanta Proactive Enforcement & Interdiction (APEX) unit, which earlier this year replaced APD’s troubled Red Dog Unit.
“We’ve received many complaints about hustling and prostitution in these corridors,” Chief George Turner (bottom photo) said in a prepared statement on Monday. “We hope these arrests send a signal to the community that the APD is listening, and to lawbreakers that we are not going to tolerate this conduct.”
Officer Brian Sharp, one of two LGBT liaisons for Atlanta police, was on hand during portions of the sting, according to Carlos Campos, a spokesperson for the department. At other times, one of the agency’s LGBT liaisons was on call.
The sting included areas known for transgender prostitutes, though none of the 61 people charged with prostitution were identified as transgender, according to Campos.
“While our LGBT liaisons were present and assisted us with identifying any potential issues within those communities, none of the arrestees expressed a desire to be treated differently,” Campos said.
The relationship between Atlanta police and the city’s LGBT residents has been strained since its botched raid of the Eagle in September 2009. LGBT activists have also criticized the department over allegations that some officers mistreat transgender sex workers, a criticism leveled at commanders in May 2010 when they introduced the department’s new LGBT liaison.
The issue came up again during a meeting of the police department’s LGBT advisory board in January as well as when it was reported that the department’s lesson plans for its LGBT diversity training includes the term “transvestite,” which is considered pejorative and offensive by some gay and transgender people.
The criticism prompted police to have Tracee McDaniel, a transgender activist and member of the department’s LGBT advisory board, and other trans activists take part in training for the new APEX unit.
On Monday, police warned that prostitution stings would continue as the activity increases in part due to warmer weather.
“This operation was a major success,” Vice commander Lt. Scott Kreher said in a prepared statement. “And it highlights our efforts at proactive policing, working with the community to address quality of life issues and APD units working together towards a common goal.”
'Tis the season for holiday parties with a charitable twist. So Impulse Group Atlanta offered its take on Saturday with 'A Candy Coated Affair Holiday Party,' the group's fifth annual holiday celebration. READ MORE »
|
{
"pile_set_name": "Pile-CC"
}
|
---
abstract: |
Compressed sensing (CS) demonstrates that sparse signals can be estimated from under-determined linear systems. Distributed CS (DCS) further reduces the number of measurements by considering joint sparsity within signal ensembles. DCS with jointly sparse signals has applications in multi-sensor acoustic sensing, magnetic resonance imaging with multiple coils, remote sensing, and array signal processing. Multi-measurement vector (MMV) problems consider the estimation of jointly sparse signals under the DCS framework. Two related MMV settings are studied. In the first setting, each signal vector is measured by a different independent and identically distributed (i.i.d.) measurement matrix, while in the second setting, all signal vectors are measured by the same i.i.d. matrix. Replica analysis is performed for these two MMV settings, and the minimum mean squared error (MMSE), which turns out to be identical for both settings, is obtained as a function of the noise variance and number of measurements. To showcase the application of MMV models, the MMSE’s of complex CS problems with both real and complex measurement matrices are also analyzed. Multiple performance regions for MMV are identified where the MMSE behaves differently as a function of the noise variance and the number of measurements.
Belief propagation (BP) is a CS signal estimation framework that often achieves the MMSE asymptotically. A phase transition for BP is identified. This phase transition, verified by numerical results, separates the regions where BP achieves the MMSE and where it is suboptimal. Numerical results also illustrate that more signal vectors in the jointly sparse signal ensemble lead to a better phase transition.
author:
- 'Junan Zhu, , Dror Baron, , and Florent Krzakala [^1] [^2] [^3]'
title: |
Performance Limits\
for Noisy Multi-Measurement Vector Problems
---
[*Keywords*]{}:Approximate message passing, multi-measurement vector problem, replica analysis.
Introduction
============
Compressed sensing (CS) [@CandesRUP; @DonohoCS; @BaraniukCS2007] demonstrates that sparse signals can be estimated from under-determined linear measurements. Owing to the potential for radically reduced measurement rates, CS has become an active research area within signal processing. CS has many application areas including magnetic resonance imaging [@JuYeKi07; @JuSuNaKiYe09], communication [@Cotter2002scemp], and remote sensing [@Ma2009deblur].
Distributed CS (DCS) [@HN05; @DuarteWakinBaronSarvothamBaraniuk2013] is based on the premise that joint sparsity within signal ensembles enables a further reduction in the number of measurements. Motivated by sensor networks [@Pottie2000], preliminary work in DCS [@Duarte2006IPSN; @HN05; @DuarteWakinBaronSarvothamBaraniuk2013] showed that the number of measurements required per sensor must account for the minimum features unique to that sensor while features that are common to multiple sensors are amortized. DCS led to a proliferation of research on the multi-measurement vector (MMV) problem [@chen2006trs; @cotter2005ssl; @Mishali08rembo; @Berg09jrmm; @LeeKimBreslerYe2011; @LeeBreslerJunge2012; @YeKimBresler2015]. The MMV problem considers the estimation of a set of sparse signal vectors that share common supports, and has applications such as radar array signal processing, acoustic sensing with multiple speakers, magnetic resonance imaging with multiple coils [@JuYeKi07; @JuSuNaKiYe09], and diffuse optical tomography using multiple illumination patterns. In MMV, thanks to the common support, the number of sparse coefficients that can be successfully estimated increases with the number of measurements. This property was evaluated rigorously for noiseless measurements using $l_0$ minimization [@DuarteWakinBaronSarvothamBaraniuk2013]. To address measurement noise, estimation approaches for MMV problems have included greedy algorithms such as SOMP [@tropp2006ass; @chen2006trs], $l_1$ convex relaxation [@malioutov2005ssr; @tropp2006ass2], and M-FOCUSS [@cotter2005ssl]. REduce MMV and BOost (ReMBo) has been shown to outperform conventional methods [@Mishali08rembo], and subspace methods have also been used to solve MMV problems [@LeeBreslerJunge2012; @YeKimBresler2015]. Statistical approaches [@ZinielSchniter2011] often achieve the oracle minimum mean squared error (MMSE). However, the performance limits of MMV signal estimation in the presence of measurement noise have not been studied.
Replica analysis is a statistical physics method that can be used to analyze the MMSE and phase transition for inverse problems [@Tanaka2002; @GuoVerdu2005; @Montanari2006; @Krzakala2012probabilistic; @krzakala2012statistical; @MezardMontanariBook; @Barbier2015; @Lesieur2015]. Barbier and Krzakala [@Barbier2015] studied the MMSE for estimating superposition codes using replica analysis. In this paper, we extend the derivation in Barbier and Krzakala [@Barbier2015] to two related yet different MMV settings: ([*i*]{}) $J$ jointly sparse signals are measured by $J$ different dense matrices that are independent and identically distributed (i.i.d.), and ([*ii*]{}) $J$ jointly sparse signals are measured by $J$ identical i.i.d. matrices. We only consider dense i.i.d. Gaussian matrices in this work, while our analysis can be extended to other i.i.d. matrices easily.
We make several contributions in this paper. First, we obtain the information theoretic MMSE for the two MMV settings above under the Bayesian setting. Second, we show that in the large system limit the MMSE’s for these two settings are identical to the single measurement vector (SMV) problem with a dense measurement matrix and a block sparse signal with fixed length blocks. Third, we derive the MMSE for SMV complex CS problems by noticing that SMV complex CS is essentially an MMV problem. Fourth, we identify several performance regions for MMV, where the MMSE has different characteristics based on channel noise variance and measurement rate. Finally, we find a phase transition for belief propagation algorithms (BP) [@DMM2009; @CSBP2010; @Bayati2011; @Montanari2012; @Krzakala2012probabilistic; @krzakala2012statistical; @Barbier2015] applied to MMV problems, which separates regions where BP achieves the MMSE asymptotically and where it is suboptimal. BP simulation results confirm the phase transition results.
The remainder of the paper is organized as follows. We introduce our signal and measurement models in Section \[sec:model\], followed by replica analyses for two MMV settings as well as two SMV complex CS problems in Section \[sec:main\]. Section \[sec:proof\] proves the results of Section \[sec:main\]. Numerical results are discussed in Section \[sec:numeric\] and we conclude in Section \[sec:conclusion\].
[**Notations:**]{} In this paper, bold capital letters represent matrices, bold lower case letters represent vectors, and normal font lower case letters represent scalars. The entry (scalar) in the $\mu$-th row, $l$-th column of a matrix $\F$ is denoted by $F_{\mu,l}$, where the comma is often omitted. The $\mu$-th entry (scalar) in a vector $\z$ is denoted by $z_{\mu}$.
Signal and Measurement Models {#sec:model}
=============================
[**Signal model**]{}: We consider an ensemble of $J$ signal vectors, $\underline{\s}^j\in\mathbb{R}^N,\ j\in\{1,...,J\}$, where $j$ is the index of the signal. Consider a [*super symbol*]{} $\s_l=[\underline{s}_l^1,...,\underline{s}_l^J]^T,\ l\in\{1,...,N\}$, where $[\cdot]^T$ denotes the transpose. The super symbol $\s_l$ follows a $J$-dimensional Bernoulli-Gaussian distribution, $$\label{eq:jsm}
P(\s_l)=\rho \phi(\s_l)+(1-\rho)\delta(\s_l),$$ where $\rho$ is the sparsity rate, $\phi(\s_l)$ is a $J$-dimensional Gaussian distribution with zero mean and identity covariance matrix, and $\delta(\s_l)$ is the delta function for $J$-dimensional vectors.
\[def:jointly\_sparse\] [*Ensembles of signals that obey are called jointly sparse.*]{}
[**Measurement models**]{}: Each signal $\underline{\s}^j$ is measured by an i.i.d. Gaussian measurement matrix $\underline{\F}^j\in\mathbb{R}^{M\times N}$, $\underline{F}_{\mu l}^j \sim \mathcal{N}(0,1/N)$, where $\mu$ refers to the row index and $l$ is the column index. The measurements $\underline{\y}^j$ are corrupted by i.i.d. Gaussian noise $\underline{\z}^j$ consisting of entries $\underline{z}_{\mu}^j\sim \mathcal{N}(0,\Delta)$, $$\label{eq:MMVmodel}
\underline{\y}^j=\underline{\F}^j\underline{\s}^j+\underline{\z}^j,\quad j\in\{1,\cdots,J\}.$$ When the number of signal vectors $J=1$, this MMV model becomes an SMV problem. Our analyses in this paper are readily extended to other i.i.d. matrices, jointly sparse signals , and other i.i.d. noise distributions.
\[def:MMV\_set1\] [*The setting MMV-1 refers to the measurement model in with all matrices $\underline{F}^j$ being different.*]{}
\[def:MMV\_set2\] [*The setting MMV-2 refers to the measurement model in with all matrices $\underline{F}^j$ being equal.*]{}
In the signal model and measurement model , the sparsity rate $\rho$, channel noise variance $\Delta$, and number of channels $J$ are constant.
\[def:largeSystemLimit\] The signal length $N$ scales to infinity, and the number of measurements $M=M(N)$ depends on $N$ and also scales to infinity, where the ratio approaches a positive constant $R$ for practical problems, $$\label{eq:measurementRate}
\lim_{N\rightarrow\infty} \frac{M(N)}{N} = R>0.$$
We call $R$ the measurement rate.
Replica Analyses for MMV Settings {#sec:main}
=================================
Section \[sec:model\] discussed two MMV settings. Both settings have applications in real-world problems such as magnetic resonance imaging [@JuYeKi07; @JuSuNaKiYe09] and sensor networks [@Pottie2000]. Although numerous algorithms for MMV signal estimation have been proposed [@tropp2006ass; @chen2006trs; @malioutov2005ssr; @tropp2006ass2; @cotter2005ssl; @Mishali08rembo; @ZinielSchniter2011], what is missing is an information theoretic analysis of the best possible mean squared error (MSE) performance. Throughout this paper, we only consider the MSE as our performance metric.
Statistical physics background and replica method {#sec:set1}
-------------------------------------------------
![Illustration of MMV channel with $J=3$ signal vectors (left), and one of its possible SMV forms (right). Different background patterns differentiate entries from different channels, and blank space denotes zeros.[]{data-label="fig:channel"}](MMV_channel_v2.png){width="8.5cm"}
In order to express using a single channel, we transform it to an SMV form. One possible way to do so is illustrated in Fig. \[fig:channel\]. The equivalent SMV problem is $$\label{eq:MMVchannel}
\y=\F\s+\z,$$ where $\F\in\mathbb{R}^{MJ\times NJ}$ is the matrix, $\y\in\mathbb{R}^{MJ}$ are the measurements, and the noise is $\z\in\mathbb{R}^{MJ}$. Entries of the signal vectors $\underline{\s}^j$, measurement vectors $\underline{\y}^j$, and noise vectors $\underline{\z}^j$ form the SMV signal $\s$, measurements $\y$, and noise $\z$ with $$s_{(l-1)J+j}=\underline{s}^{j}_l,\ y_{(j-1)M+\mu}=\underline{y}^j_{\mu},\ \text{and}\ z_{(j-1)M+\mu}=\underline{z}^j_{\mu},$$ respectively. Entries of the matrix $\underline{\F}^j$ form the SMV matrix $\F$ with $F_{(j-1)M+\mu,(l-1)J+j}=\underline{F}^j_{\mu l}$; other entries of $\F$ are zeros. The posterior for the estimate $\x\in\mathbb{R}^{NJ}$, comprised of super symbols $\x_l=[x_{(l-1)J+1},...,x_{lJ}]^T,\ l\in\{1,...,N\}$, is $$\label{eq:pxy2}
P(\x|\y)=\frac{1}{Z}{\prod_{l=1}^{N}}P(\x_l){\prod_{\mu=1}^{MJ}}\l[\frac{\operatorname{e}^{-\frac{1}{2\Delta}(\\y_{\mu}-{\sum_{l=1}^{N}}\F_{\mu l}\x_l)^2}}{\sqrt{2\pi\Delta}}\r],$$ where $\F_{\mu l}=[F_{\mu,(l-1)J+1}, \ldots, F_{\mu,lJ}]$ is a super symbol highlighted by the dashed area in Fig. \[fig:channel\], and the denominator $Z$ is the partition function [@Tanaka2002; @GuoVerdu2005; @Krzakala2012probabilistic; @krzakala2012statistical; @MezardMontanariBook; @Barbier2015], $$\label{eq:partition}
Z=\int {\prod_{l=1}^{N}}P(\x_l){\prod_{\mu=1}^{MJ}}\l[\frac{\operatorname{e}^{-\frac{1}{2\Delta}(y_{\mu}-{\sum_{l=1}^{N}}\F_{\mu l}\x_l)^2}}{\sqrt{2\pi\Delta}}\r]{\prod_{l=1}^{N}}d\x_l.$$ Note that multi-dimensional integrations such as are denoted by a single $\int$ operator for brevity. Confining our attention to the Bayesian setting [@Krzakala2012probabilistic; @krzakala2012statistical; @Barbier2015], $P(\x_l)$ follows the true distribution , $P(\x_l)=\rho \phi(\x_l)+(1-\rho)\delta(\x_l)$.
By creating an analogy between the channel and a many-body thermodynamic system [@Tanaka2002; @GuoVerdu2005; @Krzakala2012probabilistic; @krzakala2012statistical; @MezardMontanariBook; @Barbier2015], the posterior can be interpreted as the Boltzmann measure on a disordered system with the following Hamiltonian, $$\label{eq:Hamiltonian}
H(\x)=\sum_{l=1}^N \log [P(\x_l)]+\sum_{\mu=1}^{MJ} \frac{1}{2\Delta}\l(y_{\mu}-\sum_{l=1}^N\F_{\mu l} \x_l\r)^2.$$
The averaged free energy of the disordered system given by characterizes the thermodynamic properties of the system. Evaluating the fixed points (local maxima) in the free energy expression provides the MMSE for the channel [@Tanaka2002; @GuoVerdu2005; @Krzakala2012probabilistic; @krzakala2012statistical; @MezardMontanariBook; @Barbier2015]. [*Under the assumption of self-averaging*]{} [@Tanaka2002; @GuoVerdu2005; @Krzakala2012probabilistic; @krzakala2012statistical; @MezardMontanariBook; @Barbier2015], the free energy is defined as[^4] $$\label{eq:free_energy}
\mathcal{F}=\lim_{N\rightarrow\infty}\frac{1}{N}\mathbb{E}_{\F,\s,\z}[\log (Z)],$$ which is difficult to evaluate. The replica method [@Tanaka2002; @GuoVerdu2005; @Krzakala2012probabilistic; @krzakala2012statistical; @MezardMontanariBook; @Barbier2015] introduces $n$ replicas of the estimate $\x$ as $\x^a,\ a\in\{1,...,n\}$, and the free energy can be approximated by the replica trick [@Krzakala2012probabilistic; @krzakala2012statistical; @MezardMontanariBook; @Barbier2015], $$\label{eq:replicaTrick}
\mathcal{F}=\lim_{N\rightarrow\infty}\lim_{n\rightarrow 0} \frac{\mathbb{E}_{\F,\s,\z}[ Z^n]-1}{Nn}.$$ Note that the self-averaging property that leads to and the replica trick , as well as the replica symmetry assumptions that appear in latter parts of this paper, are assumed to be valid in this work, and their rigorous justification is still an open problem in mathematical physics [@Tanaka2002; @GuoVerdu2005; @Krzakala2012probabilistic; @krzakala2012statistical; @MezardMontanariBook; @Barbier2015].[^5]
**Evaluating the free energy**: To evaluate the free energy , we calculate $\mathbb{E}_{\F,\s,\z}\l[Z^n\r]$, where $\cdot_{\F,\s,\z}$ denotes expectation with respect to (w.r.t.) $\F,\s$, and $\z$, and $Z$ is given in : $$\label{eq:EZn1}
\mathbb{E}_{\F,\s,\z}\l[Z^n\r]\!=\!\frac{\mathbb{E}_{\s}\!\l[\displaystyle{\int {\prod_{l=1}^{N}}{\prod_{a=1}^{n}}\! P(\x_l^a)\!{\prod_{\mu=1}^{M}}\!\mathbb{X}_{\mu}\!{\prod_{l=1}^{N}}{\prod_{a=1}^{n}}d\x_l^a}\r]}{(2\pi\Delta)^{\frac{nMJ}{2}}},$$ where $$\label{eq:Xmu}
\mathbb{X}_{\mu}=\mathbb{E}_{\F,\z}\l[\operatorname{e}^{-\frac{1}{2\Delta}{\sum_{j=1}^{J}}{\sum_{a=1}^{n}}(v_{\mu j}^a)^2}\r],$$ $a$ is the replica index, $\x^a_l$ is the $l$-th super symbol of $\x^a$, and $$\label{eq:v_mu_a}
v_{\mu j}^a={\sum_{l=1}^{N}}\F_{\mu+M(j-1),l}(\s_l-\x_l^a)+z_{\mu+M(j-1)}.$$
\[lemma:covIsSame\] In the large system limit, the quantity $\mathbb{X}_{\mu}$ is the same for both MMV-1 and MMV-2.
Lemma \[lemma:covIsSame\] is proved in Section \[sec:proof\]. Because of Lemma \[lemma:covIsSame\], the free energy expressions for MMV-1 and MMV-2 should be identical in the large system limit. We state the result as a theorem and the detailed derivations appear in the Appendix.
\[th:free\_energy\] For settings MMV-1 and MMV-2, the free energy expressions as functions of $E$ are identical in the large system limit and are given in .[^6]
$$\begin{aligned}
\mathcal{F}(E)&=&-\frac{J}{2}R\l\{\log[2\pi(\Delta+E)]+\frac{\rho+\Delta}{E+\Delta}\r\}\!+\!\int\!P(\s_1)\! \int\! \log\! \l[ \int P(\x_1)\operatorname{e}^{-\frac{\widehat{Q}+\widehat{q}}{2}\x_1^T\x_1+\widehat{m}\x_1^T\s_1+
\sqrt{\widehat{q}}\h^T\x_1}\!d\x_1\!\r]\!\mathcal{D}\h \ d\s_1\label{eq:free_energy3}\\
&=&-\frac{J}{2}R\l\{\log[2\pi(\Delta+E)]+\frac{\Delta}{E+\Delta}\r\}+\frac{JR(1-\rho)}{2(R+E+\Delta)}+\rho\int \log \Bigg[ \rho \l(\frac{E+\Delta}{R+E+\Delta}\r)^{J/2}+\nonumber\\
&\ &(1-\rho)\operatorname{e}^{-\frac{R}{2(E+\Delta)}\g^T\g}\Bigg]\mathcal{D}\g+(1-\rho)\int \log \l[ \rho \l(\frac{E+\Delta}{R+E+\Delta}\r)^{J/2}+(1-\rho)\operatorname{e}^{-\frac{R}{2(R+E+\Delta)}\h^T\h}\r]\mathcal{D}\h.\label{eq:free_energy4}\end{aligned}$$
**MMSE**: The $E$ that maximizes the free energy the MMSE [@Krzakala2012probabilistic; @krzakala2012statistical; @Barbier2015]. After finding the $E_0$ that maximizes the free energy , we obtain the MMSE, $D_0=E_0$, in the large system limit.
The MMSE for MMV-1 and MMV-2 is the same for the same measurement rate $R$, noise variance $\Delta$, and number of signal vectors $J$.
[**Remark 1:**]{} As the reader can see from the proof of Lemma \[th:free\_energy\] in Section \[sec:proof\], the key reason that both MMV-1 and MMV-2 have an identical MMSE is that the entries in the super symbols $\s_l$ and $\x_l^{\cdot}$ are i.i.d. That said, we suspect that the MMSE for MMV-1 and MMV-2 could differ by some higher order terms. If the entries of these super symbols are not i.i.d., which is true in some practical MMV applications [@ZinielSchniter2013MMV], then it becomes more difficult to analyze the covariance matrix $\G_{\mu}$ as in Section \[sec:proof\]. Therefore, we do not have an analysis for non-i.i.d. entries within $\s_l$ and $\x_l^{\{\cdot\}}$. However, we speculate that MMV-1 might have lower MMSE than MMV-2 in that case.
[**Link to SMV with block sparse signal:**]{} The signal $\s$ in is a block sparse signal comprised of $N$ blocks of length $J$. We study a single measurement vector (SMV) problem by replacing the measurement matrix $\F$ in with an i.i.d. Gaussian matrix ${\mathbf A}\in\mathbb{R}^{MJ\times NJ}$, i.e., $\y={\mathbf A}\s+\z$. The entries of ${\mathbf A}$ follow the distribution, $A_{\mu l}\sim \mathcal{N}(0,\frac{1}{NJ})$. This SMV is similar to the setting in Barbier and Krzakala [@Barbier2015], except for the different priors and different $\ell_2$ norms in each row of ${\mathbf A}$. We consider these differences while following their derivation [@Barbier2015], and obtain the same free energy expression as . We have also shown that MMV-1 and MMV-2 have the same MMSE in the large system limit. Hence, the three settings have the same free energy expression and their MMSE’s are the same under the same noise variance $\Delta$ and measurement rate $R$ in the large system limit.
Extension to complex SMV {#sec:complex}
------------------------
MMV with jointly sparse signals is a versatile model that can be adapted to other problems. As an example, we show how the MMV model can be used to analyze the MMSE of a complex SMV.[^7] Consider the complex CS channel, $\y^{\mathcal{C}}=\F^{\mathcal{C}}\s^{\mathcal{C}}+\z^{\mathcal{C}}$, where $\s^{\mathcal{C}}=\s^{\mathcal{R}}+i\s^{\mathcal{I}}\in\mathbb{C}^N$, $\F^{\mathcal{C}}=\F^{\mathcal{R}}+i\F^{\mathcal{I}}\in\mathbb{C}^{M\times N}$, $\z^{\mathcal{C}}=\z^{\mathcal{R}}+i\z^{\mathcal{I}}\in\mathbb{C}^M$, $\y^{\mathcal{C}}=\y^{\mathcal{R}}+i\y^{\mathcal{I}}\in\mathbb{C}^M$, $i=\sqrt{-1}$, and $\mathcal{R}$ and $\mathcal{I}$ refer to the real and imaginary parts, respectively. The real and imaginary parts of the entries of $\z^{\mathcal{C}}$ both follow a Gaussian distribution, $z_l^{\mathcal{R}}, z_l^{\mathcal{I}}\sim\mathcal{N}(0,\Delta), l\in\{1,...,M\}$. Assume that the complex signal $\s^{\mathcal{C}}$ is comprised of two jointly sparse signals, $\s^{\mathcal{R}}$ and $\s^{\mathcal{I}}$, that satisfy the $J=2$ dimensional Bernoulli-Gaussian distribution . We can extend the analysis of Section \[sec:set1\] to two settings of complex CS: ([*i*]{}) the measurement matrix $\F^{\mathcal{C}}$ is real, and ([*ii*]{}) $\F^{\mathcal{C}}$ is complex.[^8]
[**Real measurement matrix:**]{} Suppose that $\F^{\mathcal{C}}$ is real, $\F^{\mathcal{C}}=\F^{\mathcal{R}}\in \mathbb{R}^{M\times N}$, and the entries of $\F^{\mathcal{R}}$ follow a Gaussian distribution, $F^{\mathcal{R}}_{\mu l}\sim \mathcal{N}(0,\frac{1}{N})$. Complex CS with a real measurement matrix can be written as real-valued MMV, $$\label{eq:complexRealMat}
\y^{\mathcal{R}}=\F^{\mathcal{R}}\s^{\mathcal{R}}+\z^{\mathcal{R}}\ \text{and}\ \y^{\mathcal{I}}=\F^{\mathcal{R}}\s^{\mathcal{I}}+\z^{\mathcal{I}},$$ where $\s^R$ and $\s^I$ are jointly sparse and follow . This formulation fits into MMV-2 for $J=2$. Hence, we can obtain the MMSE according to .[^9]
[**Complex measurement matrix:**]{} Consider a complex $\F^{\mathcal{C}}=\F^{\mathcal{R}}+i\F^{\mathcal{I}}\in\mathbb{C}^{M\times N}$ with entries $F_{\mu l}^{\mathcal{R}}, F_{\mu l}^{\mathcal{I}}\sim \mathcal{N}(0,\frac{1}{2N})$. Expanding out the complex channel, $\y^{\mathcal{C}}=\F^{\mathcal{C}}\s^{\mathcal{C}}+\z^{\mathcal{C}}$, we obtain the equivalent real-valued SMV channel, $$\label{eq.realComplexChannel}
\begin{bmatrix}
\y^{\mathcal{R}} \\
\y^{\mathcal{I}}
\end{bmatrix}
=
\begin{bmatrix}
\F^{\mathcal{R}} & -\F^{\mathcal{I}} \\
\F^{\mathcal{I}} & \F^{\mathcal{R}}
\end{bmatrix}
\begin{bmatrix}
\s^{\mathcal{R}} \\
\s^{\mathcal{I}}
\end{bmatrix}
+
\begin{bmatrix}
\z^{\mathcal{R}} \\
\z^{\mathcal{I}}
\end{bmatrix}.$$
We re-arrange as follows, $$\label{eq:complexMatRearrange}
\underbrace{\begin{bmatrix}
\y^{\mathcal{R}} \\
\y^{\mathcal{I}}
\end{bmatrix}}_{\overline{\y}}
\!=\!
\underbrace{\begin{bmatrix}
\F_{:,1}^{\mathcal{R}},-\F_{:,1}^{\mathcal{I}},...,\F_{:,N}^{\mathcal{R}}, -\F_{:,N}^{\mathcal{I}} \\
\F_{:,1}^{\mathcal{I}},\ \ \F_{:,1}^{\mathcal{R}},...,\F_{:,N}^{\mathcal{I}},\ \ \F_{:,N}^{\mathcal{R}}
\end{bmatrix}}_{\overline{\F}}
\underbrace{\begin{bmatrix}
s_1^{\mathcal{R}}\\
s_1^{\mathcal{I}}\\
\vdots\\
s_N^{\mathcal{R}}\\
s_N^{\mathcal{I}}
\end{bmatrix}}_{\overline{\s}}
\!+\!
\underbrace{\begin{bmatrix}
\z^{\mathcal{R}} \\
\z^{\mathcal{I}}
\end{bmatrix}}_{\overline{\z}},$$ where $\{:\}$ refers to all the rows. In the re-arranged channel , the measurement matrix $\overline{\F}$ consists of super symbols, $$\label{eq:SMV_F}
\overline{\F}_{\mu l}=\left\{
\begin{array}{ll}
&[F_{\mu l}^{\mathcal{R}},-F_{\mu l}^{\mathcal{I}}],\ \mu\in\{1,...,M\}\\
&[F_{\mu l}^{\mathcal{I}}, F_{\mu l}^{\mathcal{R}}],\ \mu\in\{M+1,...,2M\}
\end{array}
\right.,\\$$ and the signal $\overline{\s}$ consists of $\overline{\s}_{l}=\begin{bmatrix}
s_l^{\mathcal{R}} \\
s_l^{\mathcal{I}}
\end{bmatrix},\ l\in\{1,...,N\}$. The measurements and noise are $\overline{\y}=\begin{bmatrix}
\y^{\mathcal{R}} \\
\y^{\mathcal{I}}
\end{bmatrix}$ and $\overline{\z}=\begin{bmatrix}
\z^{\mathcal{R}} \\
\z^{\mathcal{I}}
\end{bmatrix}$, respectively. Hence, $\overline{y}_{\mu}=\sum_{l=1}^N \overline{\F}_{\mu l}\overline{\s}_l+\overline{z}_{\mu},\ \mu\in\{1,...,2M\}$.
Section \[sec:proof\] shows that the free energy and MMSE for SMV complex CS with complex measurement matrices are the same as MMV-1 with $J=2$. Note that in the free energy expression of MMV-1, the MSE, $D=E$ , is the average MSE of the $J$ entries of $\s_l$. Therefore, in this complex CS setting, $D$ is the average MSE of the real and imaginary parts of the signal entries.
Proof of Lemma \[lemma:covIsSame\] {#sec:proof}
==================================
In this section, we show that the quantity $\mathbb{X_{\mu}}$ is the same for MMV-1 and MMV-2. Moreover, we show that complex SMV with a complex measurement matrix also yields the same $\mathbb{X_{\mu}}$ with $J=2$.
First, we re-write in the vector form $$\label{eq:XMuVector}
\mathbb{X}_{\mu}\!=\!\mathbb{E}_{\v_{\mu}}\!\left[\operatorname{e}^{-\frac{1}{2\Delta}{\sum_{j=1}^{J}}{\sum_{a=1}^{n}}(v_{\mu j}^a)^2}\right]\!
=\!\mathbb{E}_{\v_{\mu}}\!\left[\operatorname{e}^{-\frac{1}{2\Delta}\v_{\mu}^T\v_{\mu}}\right],$$ where $\v_{\mu}=[v_{\mu 1}^1,...,v_{\mu 1}^a,...,v_{\mu J}^1$, $...,v_{\mu J}^n]^T$ and $v_{\mu j}^a$ is given in . In order to calculate the expectation w.r.t. $\v_{\mu}$ in , we calculate the distribution of $\v_{\mu}$, which is approximated by a Gaussian distribution, due to the central limit theorem. The mean is $\mathbb{E}_{\F,\z}[v_{\mu j}^a]=0$.
We now calculate the covariance matrix, $\G_{\mu}=\mathbb{E}[\v_{\mu}\v_{\mu}^T]$. The matrix is separated into $J\times J$ blocks of size $n\times n$, as shown in Fig. \[fig.cov\]. The main diagonal of $\G_{\mu}$ consists of entries $w_1=\mathbb{E}_{\F,\z}[(v_{\mu j}^a)^2]$. The entries in the blocks along the main diagonal (other than entries along the main diagonal itself) are $w_3=\mathbb{E}_{\F,\z}[v_{\mu j}^a v_{\mu j}^b]$. The main diagonals of other blocks have entries $w_2=\mathbb{E}_{\F,\z}[v_{\mu j}^a v_{\mu\eta}^a]$, and other entries in these blocks are $w_4=\mathbb{E}_{\F,\z}[v_{\mu j}^a v_{\mu \eta}^b]$. We now calculate each of these values as follows for MMV-1, MMV-2, and complex SMV with a complex measurement matrix.
![Covariance matrix $\G_{\mu}\in\mathbb{R}^{nJ\times nJ}$. Each block in $\G_{\mu}$ has a size of $n\times n$. The entries in the heavily marked blocks take the value $w_3$, except that entries along the dashed diagonal are $w_1$. The entries in the lightly marked blocks take the value $w_4$, except that entries along the dotted diagonal are $w_2$.[]{data-label="fig.cov"}](MMV_cov_mat.png){width="3.5cm"}
[**MMV-1:**]{} We begin by calculating the diagonal entries of the covariance matrix $\G_{\mu}=\mathbb{E}[\v_{\mu}\v_{\mu}^T]$, $$\label{eq:vVar1}
\begin{split}
&w_1=\mathbb{E}_{\F,\z}\!\l[(v_{\mu j}^a)^2\r]\!=\!{\sum_{l,k=1}^{N,N}}\!\Bigg[(\s_l-\x_l^a)^T\times\\
& \mathbb{E}_{\F}\l\{\F_{\mu+M(j-1), l}^T\F_{\mu+M(j-1), k}\r\}(\s_k-\x_k^a)\Bigg]+\Delta.
\end{split}$$ In , $\mathbb{E}_{\F}\l\{\F_{\mu+M(j-1), l}^T\F_{\mu+M(j-1), k}\r\}=\frac{\delta_{k,l}}{N}\widetilde{\mathbf{I}}_J$ (cf. Fig. \[fig:channel\]), where $\widetilde{\mathbf{I}}_J$ is a $J\times J$ matrix with only one 1 located at row $j$ and column $j$, and $\delta_{k,l}=1$ when $k=l$, else zero. Hence, becomes $$\begin{aligned}
w_1&=& \mathbb{E}_{\F,\z}\l[(v_{\mu j}^a)^2\r]=\frac{1}{N}{\sum_{l=1}^{N}} (s_{l,j}-x_{l,j}^a)^2+\Delta\label{eq:vVar2}\\
&=&\!\frac{1}{NJ}\!{\sum_{l=1}^{N}} \!(\s_l-\x_l^a)^T (\s_l-\x_l^a)\!+\!\Delta,\label{eq:vVar2_1}\end{aligned}$$ where $s_{l,j}$ and $x_{l,j}^a$ denote the $j$-th entries in super symbols $\s_l$ and $\x_l^a$, respectively, and holds because all $J$ entries within the same super symbol ($\s_l$ or $\x_l^a$) are i.i.d.
Similarly, we obtain $$\begin{split}
w_2=&\mathbb{E}_{\F,\z}[v_{\mu j}^a v_{\mu\eta}^{a}] = \frac{1}{N}{\sum_{l=1}^{N}}(s_{l,j}-x_{l,j}^a)(s_{l,\eta}-x_{l,\eta}^{a})\\
&= \frac{1}{NJ}{\sum_{l=1}^{N}}(\s_l-\x_l^a)^T (\s_l^a-\x_l^b),\label{eq:sx_iid1}
\end{split}$$ where entries of $\s_l^a$ follow the same distribution as entries of $\s_l$ given $l$, and is due to ([*i*]{}) entries of $\s_l$ being i.i.d., ([*ii*]{}) entries of $\x_l^{\{\cdot\}}$ being i.i.d. for fixed $l$, and ([*iii*]{}) the replica symmetry assumption [@Krzakala2012probabilistic; @krzakala2012statistical]. We also obtain $$\label{eq:v_j_eta_a}
\begin{split}
w_3=\mathbb{E}_{\F,\z}[v_{\mu j}^a v_{\mu j}^b]\!&=\!\frac{1}{NJ}\!{\sum_{l=1}^{N}}(\s_l\!-\!\x_l^a)^T \!(\s_l-\x_l^b)+\Delta.\\
w_4=\mathbb{E}_{\F,\z}[v_{\mu j}^a v_{\mu\eta}^{b}]&=\frac{1}{NJ}{\sum_{l=1}^{N}}(\s_l-\x_l^a)^T (\s_l^a-\x_l^b),
\end{split}$$
We now define the following auxiliary parameters $$\label{eq:auxParamsSet1}
\begin{split}
m_a=\frac{\displaystyle{\sum_{l=1}^{N}} (\x_l^a)^T\s_l}{NJ},&\quad Q_a=\frac{\displaystyle{\sum_{l=1}^{N}} (\x_l^a)^T\x_l^a}{NJ}, \\
q_{ab}=\frac{\displaystyle{\sum_{l=1}^{N}} (\x_l^a)^T\x_l^b}{NJ},&\quad
q_0=\frac{1}{NJ}{\sum_{l=1}^{N}}(\s_l^a)^T \s_l,
\end{split}$$ which allow us to express – as $$w_1=\rho-2m_a+Q_a+\Delta,$$ $$\label{eq:ws2}
w_2=q_0-(m_a+m_b)+q_{ab},$$ $$w_3 = \rho-(m_a+m_b)+q_{ab}+\Delta,$$ $$\label{eq:ws4}
w_4=q_0-(m_a+m_b)+q_{ab}.$$
Plugging the distribution of $\v_{\mu}$, approximated by $P(\v_{\mu})=[(2\pi)^n\det (\G_{\mu})]^{-\frac{1}{2}}\exp(-\frac{1}{2}\v_{\mu}^T\G_{\mu}^{-1}\v_{\mu})$, into , we obtain $$\label{eq:Xmu2}
\mathbb{X}_{\mu}=\l[\det(\mathbb{I}_n+\frac{1}{\Delta}\G_{\mu})\r]^{-1/2}.$$
[**MMV-2:**]{} For the matrix $\F$ in this setting, rows $jM+1,...,(j+1)M,\ 2\leq j \leq J$, will be the right-shift of rows $(j-1)M+1,...,jM$. We express $v_{\mu j}^a$ as $$\label{eq:v_mu_j_a_mmv2}
v_{\mu j}^a={\sum_{l=1}^{N}}\F_{\mu l}{\bf T}_{j}(\s_l-\x_l^a)+z_{\mu+M(j-1)},\ \mu\in \{1,\cdots,M\},$$ where $\T_{j}$ is a $J\times J$ transform matrix with the $j$-th entry of the first row being one and all other entries in $\T_j$ being zeros. Using the same derivations as in MMV-1, it can be proved that the covariance matrix $\G_{\mu}=\mathbb{E}[\v_{\mu}\v_{\mu}^T]$ in MMV-2 is identical to that of MMV-1. Therefore, $\mathbb{X}_{\mu}$ in MMV-1 and MMV-2 are identical in the large system limit.
[**Complex SMV with complex measurement matrix:**]{} The derivations are the same as in MMV-2 above, except that we need to change $\F_{\mu l}$ in to $\overline{\F}_{\mu l}$ and replace $\T_j$ by $$\label{eq:TransMat}
{\bf T}=\begin{bmatrix}
0 & 1 \\
-1 & 0
\end{bmatrix},$$ because $\overline{\F}_{(\mu+M)l}=\overline{\F}_{\mu l}{\bf T},\ \mu\in\{1,...,M\}$. Using similar steps as above, we obtain that the covariance matrix $\G_{\mu}$ in this case, is also the same as that of MMV-1 with $J=2$.
**Solving $\mathbb{X}_{\mu}$**: For such a structured matrix $\G_{\mu}$ (Fig. \[fig.cov\]), elementary transforms show that the eigen-values (EV’s) are comprised of one EV equal to $\alpha_1=[w_1+(J-1)w_2]+(n-1)[w_3+(J-1)w_4],\ (J-1)$ EV’s equal to $\alpha_2=(w_1-w_2)+(n-1)(w_3-w_4),\ (n-1)$ EV’s equal to $\alpha_3=[w_1+(J-1)w_2]-[w_3+(J-1)w_4]$, and $(J-1)(n-1)$ EV’s equal to $\alpha_4=(w_1-w_2)-(w_3-w_4)$.
Owing to replica symmetry [@Krzakala2012probabilistic; @krzakala2012statistical], we have $m_a=m_b=m$, $Q_a=Q$, and $q_{ab}=q$. Also, in the Bayesian setting, we have $m=q_0=q$ and $Q=\rho$. Thus, $w_2=w_4=0$ ( and ), and $$\label{eq:detSet2}
\begin{split}
\det (\mathbb{I}_{nJ}+&\frac{1}{\Delta}\G_{\mu})= \l(1+\frac{\alpha_1}{\Delta}\r)\l(1+\frac{\alpha_2}{\Delta}\r)^{J-1}\times\\
&\quad \l(1+\frac{\alpha_1}{\Delta}\r)^{n-1}\l(1+\frac{\alpha_1}{\Delta}\r)^{(n-1)(J-1)}\\
&=\l(1+n\frac{w_3}{\Delta+\alpha_4}\r)^J\!\l(1+\frac{1}{\Delta}\alpha_4\r)^{Jn}\!.
\end{split}$$ Considering , we simplify , $$\label{eq:XmuNew}
\lim_{n\rightarrow 0}\mathbb{X}_{\mu}=\operatorname{e}^{-\frac{nJ}{2}\l[\frac{\rho-2m+\Delta+q}{Q-q+\Delta}+\log(Q-q+\Delta)-\log(\Delta)\r]},$$ where we rely on the following Taylor series, $$\label{eq:firstOrder1}
\operatorname{e}^{nk}\approx 1+nk\Rightarrow \operatorname{e}^{-\frac{n}{2}k}\approx (1+nk)^{-1/2},\ n\rightarrow 0.$$
Numerical Results {#sec:numeric}
=================
Given a free energy expression for a CS problem, the MMSE can be obtained by evaluating the largest free energy [@Tanaka2002; @GuoVerdu2005; @Krzakala2012probabilistic; @krzakala2012statistical; @MezardMontanariBook; @Barbier2015]. Having derived the free energy for the two MMV settings in Section \[sec:main\], this section calculates the MMSE under various cases. Different performance regions of MMV are identified, where the MMSE behaves differently as a function of the noise variance $\Delta$ and measurement rate $R$. We identify a phase transition of belief propagation (BP) that separates regions where BP is optimal asymptotically or not. Simulation results match the predicted performance of BP.
Performance regions: Definitions and numerical results {#sec:PerfRegion}
------------------------------------------------------
When calculating the MMSE for different settings from the free energy expression , four different [*performance regions*]{} will appear, as illustrated in Fig. \[fig:freeEnergyProf\] and discussed below.
![Free energy as a function of MSE for different measurement rates $R$ (number of jointly sparse signal vectors $J=3$ and noise variance $\Delta=-35$ dB). The black circles mark the largest free energy, and so they correspond to the MMSE.[]{data-label="fig:freeEnergyProf"}](freeEnergyProfiles_J3.pdf){width="8cm"}
[**Regions 1 and 4:**]{} The free energy has one local maximum point w.r.t. the MSE $D$ . This $D$ leads to the globally maximum free energy and is the MMSE.
[**Regions 2 and 3:**]{} There are 2 local maxima in the free energy, $D_1$ and $D_2$, where $D_1<D_2$. In Region 2, the smaller MSE, $D_1$, leads to the larger local maximum free energy (hence, $\mathcal{F}(D_1)$ is the global maximum), and is the MMSE. In Region 3, the larger MSE, $D_2$, is the MMSE.
[**Boundaries between regions:**]{} We denote the boundary separating regions 1 and 2 by the [*BP threshold*]{} $R_{BP}(\Delta)$, the boundary separating regions 2 and 3 by the [*low noise threshold*]{} $R_l(\Delta)$, and the boundary separating regions 3 and 4 by the [*critical threshold*]{} $R_c(\Delta)$.
[**Numerical results:**]{} Consider $J$-dimensional Bernoulli-Gaussian signals with sparsity rate $\rho=0.1$. Evaluating the free energy with the noise variance $\Delta$ from -20 dB to -50 dB and measurement rate $R$ from 0.11 to 0.24, we obtain the MMSE as a function of $\Delta$ and $R$ for $J=1,3$, and $5$, as shown in Fig. \[fig:PerformanceRegions\].[^10] The darkness of the shades represents the natural logarithm of the MMSE, $\ln$(MMSE). In all panels, the critical threshold $R_c(\Delta)$, low noise threshold $R_l(\Delta)$, and BP threshold $R_{BP}(\Delta)$, as well as Regions 1-4, are marked.
In Regions 3 and 4, the best-possible algorithm yields a large MMSE for all noise variances. In contrast, in Regions 1 and 2, the optimal algorithm yields an MMSE that decreases with the noise variance $\Delta$. To summarize, the optimal algorithm yields poor estimation performance below the low noise threshold $R_{l}(\Delta)$, and good performance above $R_{l}(\Delta)$.
We further examine the MMSE as a function of the number of jointly sparse signal vectors $J$ and the measurement rate $R$. We plotted the MMSE in dB scale in Fig. \[fig:MMSE\_R\_J\]. The noise variance is -35 dB. We can see that the MMSE decreases with more signal vectors $J$ and greater measurement rate $R$. However, the MMSE depends less on $J$ as $J$ is increased. Note that the discontinuity in the MMSE surface in Fig. \[fig:MMSE\_R\_J\] is a result of the different performance regions that the various settings (different $J$ and $R$) lie in.
BP phase transition {#sec:phaseTrans}
-------------------
Belief propagation (BP) [@DMM2009; @CSBP2010; @Montanari2012; @Bayati2011; @Krzakala2012probabilistic; @krzakala2012statistical; @Barbier2015] is an algorithmic framework motivated by statistical physics, which can often achieve the optimal estimation performance (MMSE). When there are multiple local maxima $D_1<D_2$ in the free energy , BP converges to the local maximum with the larger MSE, $D_2$ [@DMM2009; @Montanari2012; @Bayati2011; @Krzakala2012probabilistic; @krzakala2012statistical]. Hence, $D_2$ characterizes the [*predicted MSE*]{} for BP. Moving from Region 1 to Region 2 by decreasing the measurement rate $R$ with fixed noise variance $\Delta$, the number of local maxima increases from 1 to 2. Therefore, BP estimation performance experiences a sudden deterioration (increase in MSE) when the measurement rate $R$ drops such that the combination of the noise variance $\Delta$ and measurement rate $R$ moves from Region 1 to Region 2. The BP threshold, $R_{BP}(\Delta)$, is the boundary between Regions 1 and 2, and is where the BP phase transition happens. That is, BP achieves poor estimation performance below $R_{BP}(\Delta)$, and good performance above $R_{BP}(\Delta)$.
[**Remark 2**]{}: In Fig. \[fig:PerformanceRegions\], we see that increasing $J$ reduces the BP threshold $R_{BP}(\Delta)$. Since BP achieves the MMSE when $R>R_{BP}(\Delta)$, increasing $J$ is beneficial to applications that use BP as the estimation algorithm.
[**Remark 3**]{}: We further analyzed the low noise ($\Delta\rightarrow 0$) and zero noise ($\Delta=0$) cases. The critical threshold $R_c(\Delta)$ converges to $\rho$ as the noise variance $\Delta$ is decreased for $J=1,3$, and $5$. We believe that this numerical result holds for every $J$. Moreover, this result matches the theoretical robust threshold of Wu and Verd[ú]{} [@WuVerdu2012] for $J=1$ in the low noise limit. Our numerical results also show that the BP threshold $R_{BP}(\Delta)$ converges to some value for different $J$ as $\Delta\rightarrow 0$. Analyzing these observations rigorously is left for future work.
![MMSE in dB as a function of measurement rate $R$ and number of jointly sparse signal vectors $J$ (noise variance $\Delta=-35$ dB).[]{data-label="fig:MMSE_R_J"}](MMSE_vs_R_J.pdf){width="8cm"}
BP simulation {#sec:AMPsim}
-------------
After obtaining the theoretic MMSE for MMV, as well as the predicted MSE for BP, we run some simulations to estimate the $\underline{\s}^j$ of channel in a Bayesian setting. The algorithm we use is approximate message passing (AMP) [@DMM2009; @Montanari2012; @Bayati2011; @Krzakala2012probabilistic; @krzakala2012statistical; @Barbier2015], which is an approximation to the BP algorithm; related algorithms have been proposed by Ziniel and Schniter [@ZinielSchniter2013MMV] and Kim et al. [@KimChangJungBaronYe2011]. In the SMV case, when the measurement matrix and the signal have i.i.d. entries, AMP has the state-evolution (SE) property [@DMM2011; @Bayati2011; @JavanmardMontanari2012; @Donoho2013; @Bayati2015] that tracks the evolution of the MSE at each iteration. Recently, Javanmard and Montanari proved that SE tracks AMP rigorously in an SMV setting with a spatially coupled measurement matrix [@JavanmardMontanari2012]. According to our transform in Fig. \[fig:channel\], we can see that the proof [@JavanmardMontanari2012] could be extended to the MMV setting. Note that SE allows to compute the highest equilibrium of Gibbs free energy [@DMM2011; @Bayati2011; @JavanmardMontanari2012; @Donoho2013; @Bayati2015], which corresponds to the local optimum $D_2$ in Section \[sec:phaseTrans\]. Hence, AMP often achieves the same MSE as BP and we use AMP simulation results to demonstrate that the MMSE can often be achieved.[^11] Considering the structure of $\F$, we simplify the AMP algorithm in Barbier and Krzakala [@Barbier2015] to obtain Algorithm \[algo:AMP\_MMV\],[^12] where $\{\Sigma_j\}_{j=1}^J$, $\{R^j_l\}_{j=1}^J$, $\{a_l^j\}_{j=1}^J$ and $\{v_l^j\}_{j=1}^J$ refer to sets of all intermediate variables $\Sigma_j$, pseudodata $R^j_l$, estimates $a_l^j$, and variances $v^j_l,\ j\in\{1,...,J\},\ l\in\{1,...,N\}$, respectively. The current iteration $t$, change in the estimate $\delta$, and intermediate variables $\Theta_j,\ j\in\{1,...,J\}$, are scalars. The intermediate variables $\q^j$ and $\w^j$ are vectors of length $M$. The functions $f_{a_l}(\{\Sigma_j\}_{j=1}^J,\{R^j_l\}_{j=1}^J)$ and $f_{v_l}(\{\Sigma_j\}_{j=1}^J,\{R^j_l\}_{j=1}^J)$ are given by $$\begin{split}
&f_{a_l}(\{\Sigma_j\}_{j=1}^J,\{R^j_l\}_{j=1}^J)=\\
&\frac{\rho\frac{1}{\Sigma_j+1}\{R^j_l\}_{j=1}^J}{\rho+(1-\rho){\prod_{j=1}^{J}}\l\{\sqrt{1+\frac{1}{\Sigma_j}}\exp\l[-\frac{(R^j_l)^2}{2\Sigma_j(\Sigma_j+1)}\r]\r\}},
\end{split}$$ $$\begin{split}
&f_{v_l}(\{\Sigma_j\}_{j=1}^J,\{R^j_l\}_{j=1}^J)=-\l[f_{a_l}(\{\Sigma_j\}_{j=1}^J,\{R^j_l\}_{j=1}^J)\r]^2\\
&+\frac{\rho\frac{1}{\Sigma_j+1}\l[(\{R^j_l\}_{j=1}^J)^2\frac{1}{\Sigma_j+1}+\Sigma_j\r]}{\rho+(1-\rho){\prod_{j=1}^{J}}\l\{\sqrt{1+\frac{1}{\Sigma_j}}\exp\l[-\frac{(R^j_l)^2}{2\Sigma_j(\Sigma_j+1)}\r]\r\}},
\end{split}$$ for $J$-dimensional Bernoulli-Gaussian signals .
\
[**Inputs:**]{} Maximum number of iterations $t_{max}$, threshold $\epsilon$, sparsity rate $\rho$, noise variance $\Delta$, measurements $\y^j$, and measurement matrices $\F^j, \forall j$\
[**Initialize:**]{} $t=1,\delta=\infty,\w^j=\y^j,\Theta_j=0,v^j_l=\rho\Delta,a^j_l=0,\forall l,j$\
$\q^j=\frac{\y^j-\w^j}{\Delta+\Theta_j}$\
$\Theta_j=\frac{1}{N}\sum_{l=1}^N v^j_l$\
$\w^j=\F^j \a^j-\Theta_j \q^j$\
$\Sigma_j=\frac{N(\Delta+\Theta_j)}{M}$ // Scalar channel noise variance\
$\R^j=\a^j+\Sigma_j (\F^j)^T \frac{\y^j-\w^j}{\Delta+\Theta_j}$ // Pseudodata\
$\widehat{\a}^j=\a^j$ // Save current estimate\
$\{v^j_l\}_{j=1}^J=f_{v_l}(\{\Sigma_j\}_{j=1}^J,\{R^j_l\}_{j=1}^J)$ // Variance\
$\{a^j_l\}_{j=1}^J=f_{a_l}(\{\Sigma_j\}_{j=1}^J,\{R^j_l\}_{j=1}^J)$ // Estimate\
$t=t+1$ // Increment iteration index.\
$\delta=\frac{1}{NJ}\sum_{l=1}^N\sum_{j=1}^J(\widehat{a}^j_l-a^j_l)^2$ // Change in estimate\
[**Outputs:**]{} Estimate $\a^j,\forall j$
We simulated the signals in with $J=3$ signal vectors and sparsity rate $\rho=0.1$ measured by a channel with measurement rate $R\in[0.11,0.24]$ and noise variance $\Delta\in[-20,-50]$ dB. For each setting, we generated 50 signals of length $N=5000$, and the resulting MSE compared to the predicted BP MSE is shown in Fig. \[fig:AMPoverMSE\].[^13]
![AMP simulation results ($\text{MSE}_{\text{AMP}}$) compared to the predicted BP MSE ($\text{MSE}_{\text{BP}}$) with $J=3$ jointly sparse signal vectors. The dashed curve, solid curve, and the curve comprised of little circles correspond to thresholds $R_c(\Delta),\ R_l(\Delta)$, and $R_{BP}(\Delta)$, respectively. Regions 1-4 are also marked. The shade denotes $\ln \l(\frac{\text{MSE}_{\text{AMP}}}{\text{MSE}_{\text{BP}}}\r)$, which we expect to be 0 (completely dark shades) in the entire $R$ versus $\Delta$ plane. The narrow bright band above the BP threshold indicates the mismatch of AMP simulated MSE to the BP predicted MSE.[]{data-label="fig:AMPoverMSE"}](simMSE_over_predMSE_J3_TSP_v2.pdf){width="8cm"}
The labels of the thresholds are omitted for brevity. We can see that AMP simulation results match with the predicted MSE of BP and BP phase transition from the replica analysis of Section \[sec:phaseTrans\]. Note that there is a narrow band of light shades above the BP threshold, $R_{BP}(\Delta)$ (the top threshold), meaning that the simulated MSE is greater than the predicted MSE; this is due to randomness in our generated signals and channels. Note that we also compared the AMP simulation results to that of the M-SBL algorithm [@YeKimBresler2015], a widely used algorithm to solve the MMV problem. The M-SBL results were not as good. Indeed, because AMP is often an achievable to the MMSE, other algorithms are expected to provide greater MSE.
Conclusion {#sec:conclusion}
==========
We analyzed the minimum mean squared error (MMSE) for two settings of multi-measurement vector (MMV) problems, where the entries in the signal vectors are independent and identically distributed (i.i.d.), and share the same support. One MMV setting has i.i.d. Gaussian measurement matrices, while the other MMV setting has identical i.i.d. Gaussian measurement matrices. Replica analyses yield identical free energy expressions for these two settings in the large system limit when the signal length goes to infinity and the number of measurements scales with the signal length. Because of the identical free energy expressions, the MMSE’s for both MMV settings are identical. By numerically evaluating the free energy expression, we identified different performance regions for MMV where the MMSE as a function of the channel noise variance and the measurement rate behaves differently. We also identified a phase transition for belief propagation algorithms (BP) that separates regions where BP achieves the MMSE asymptotically and where it is suboptimal. Simulation results of an approximated version of BP matched with the MSE predicted by replica analysis. As a special case of MMV, we extended our replica analysis to single measurement vector (SMV) complex CS, so that we can calculate the MMSE for SMV complex CS with real or complex measurement matrices.
Acknowledgments {#acknowledgments .unnumbered}
===============
The work in this paper is based in part on preliminary work with Jong Min Kin, Woohyuk Chang, Bangchul Jung, and Jong Chul Ye [@KimChangJungBaronYe2011]. The authors thank Lenka Zdeborov[á]{} for useful discussions about replica analysis, and Yanting Ma and Ryan Pilgrim for helpful comments. Junan Zhu also thanks Shikai Luo for helpful discussions.
This appendix follows the derivation of Barbier and Krzakala [@Barbier2015], except for some nuances. Our compressed derivation makes the presentation self-contained.
Plugging and the following identity [@Barbier2015; @Krzakala2012probabilistic], $$\begin{split}
&1=\int \exp\Bigg\{-{\sum_{a=1}^{n}} \l[\widehat{m}_a\l(m_a NJ-{\sum_{l=1}^{N}}(\x_l^a)^T\s_l\r)\r]+{\sum_{a=1}^{n}}\Bigg[\\
&\widehat{Q}_a\!\l(Q_a\frac{NJ}{2}\!-\!\frac{1}{2}{\sum_{l=1}^{N}}(\x_l^a)^T\x_l^a\r)\Bigg]\!-\!{\sum_{1\leq a< b\leq n}^{}}\Bigg[\widehat{q}_{ab}\Bigg(q_{ab} NJ-\\
&{\sum_{l=1}^{N}}(\x_l^a)^T\x_l^b\Bigg)\Bigg]\!\Bigg\}\!{\prod_{a=1}^{n}}dQ_a\ d\widehat{Q}_{a}\ dm_a \ d\widehat{m}_{a}\! {\prod_{1\leq a<b\leq n}^{}}\! dq_{ab}\ d\widehat{q}_{ab},
\end{split}$$ into , we obtain $$\label{eq:EZn3}
\begin{split}
\mathbb{E}_{\F,\s,\z}&[Z^n]\!=\!(2\pi\Delta)^{-\frac{nMJ}{2}}\bigintsss\! \exp\Bigg[NJ\Bigg(\frac{1}{2}{\sum_{a=1}^{n}}\widehat{Q}_aQ_a\\
& -\frac{1}{2}{\sum_{\substack{1\leq a,b\leq n\\a\neq b}}^{}}\widehat{q}_{ab}q_{ab}-{\sum_{a=1}^{n}}\widehat{m}_am_a\Bigg)\Bigg]\l[{\prod_{\mu=1}^{M}} \mathbb{X}_{\mu}\!\r]\times\\
&\Gamma^N{\prod_{a=1}^{n}}dQ_a\ d\widehat{Q}_{a}\ dm_a\ d\widehat{m}_{a}\!{\prod_{\substack{1\leq a,b\leq n\\a\neq b}}^{}}\! dq_{ab}\ d\widehat{q}_{ab},
\end{split}$$ where $$\label{eq:Gamma_original}
\begin{split}
&\Gamma=\!\!\!\int\! P(\s_1)\! \l[{\prod_{a=1}^{n}}P(\x^a_1)\r]\!\exp\!\Bigg[\!-\frac{1}{2}{\sum_{a=1}^{n}}\widehat{Q}_a(\x^a_1)^T\x^a_1+\\
&\frac{1}{2}{\sum_{\substack{1\leq a,b\leq n\\a\neq b}}^{}}\widehat{q}_{ab}(\x^a_1)^T\x^b_1+{\sum_{a=1}^{n}}\widehat{m}_a(\x^a_1)^T\s_1\Bigg]d\s_1 {\prod_{a=1}^{n}}d\x^a_1.
\end{split}$$
$$\label{eq:PhiJ}
\begin{split}
\widetilde{\Phi}_J&(m,\widehat{m},q,\widehat{q},Q,\widehat{Q})=\frac{J}{2}(Q\widehat{Q}+q\widehat{q}-2m\widehat{m})-\frac{MJ}{2N}\l[\frac{\rho-2m+\Delta+q}{Q-q+\Delta}+\log(Q-q+\Delta)-\log(\Delta)\r]+\\
&\int P(\s_1) \l\{ \int \log \l\{ \int P(\x_1)\exp\l[-\frac{1}{2}(\widehat{Q}+\widehat{q})\x_1^T\x_1+\widehat{m}\x^T_1\s_1+
\sqrt{\widehat{q}}\h^T\x_1\r]d\x_1\r\} \mathcal{D}\h \r\} d\s_1-\frac{MJ}{2N}\log(2\pi\Delta).
\end{split}$$
**Further simplification of **: The Stratanovitch transform [@Stratanovitch-Wiki] in $J$ dimensions is given by $$\label{eq:strat}
\begin{split}
&\exp\!\l[\frac{\widehat{q}}{2}\!{\sum_{\substack{1\!\leq\! a,b\!\leq n\\a\neq b}}^{}}\!(\x^a_1)^T\!\x^b_1\!\r]\!=\!{\prod_{j=1}^{J}}\!\exp\!\l[\frac{\widehat{q}}{2}{\sum_{\substack{1\leq a,b\leq n\\a\neq b}}^{}}x^a_{1,j} x^b_{1,j}\r]\\
&={\prod_{j=1}^{J}}\int\exp\l[\sqrt{\widehat{q}}h_j{\sum_{a=1}^{n}}x^a_{1,j}-\frac{\widehat{q}}{2}{\sum_{a=1}^{n}}(x^a_{1,j})^2\r]\mathcal{D}h_j\\
&=\int\exp\l[\sqrt{\widehat{q}}\h^T{\sum_{a=1}^{n}}\x^a_1-\frac{\widehat{q}}{2}{\sum_{a=1}^{n}}(\x^a_1)^T\x^a_1\r]\mathcal{D}\h,
\end{split}$$ where $\h=[h_1,...,h_J]^T$, and the differential $\mathcal{D}h_j=\frac{1}{\sqrt{2\pi}}\operatorname{e}^{-h_j^2/2}d h_j$. With the Stratanovitch transform , we simplify $\Gamma$ as follows, $$\label{eq:Gamma}
\Gamma=\int P(\s_1)\int\l[f(\h)\r]^n \mathcal{D}\h\ d\s_1,$$ where $f(\h)=\int P(\x_1)\operatorname{e}^{-\frac{\widehat{Q}+\widehat{q}}{2}\x_1^T\x_1+\widehat{m}\x_1^T\s_1+\sqrt{\widehat{q}}\h^T\x_1}d\x_1$, and we drop the super-script $a$ of $\x_1^a$ owing to the replica symmetry assumption [@Krzakala2012probabilistic; @krzakala2012statistical]. In the limit of $n\rightarrow 0$, using another Taylor series $[f(\h)]^n\approx 1+n\log [f(\h)]$, we have $\int [f(\h)]^n \mathcal{D}\h\approx 1+n\int \log [f(\h)] \mathcal{D}\h\approx\operatorname{e}^{n\int \log [f(\h)]\mathcal{D}\h}$, so that $\mathbb{E}\{\int[f(\h)]^n\mathcal{D}\h\}\approx\mathbb{E}\{1+n\int \log [f(\h)] \mathcal{D}\h\}\approx\operatorname{e}^{\mathbb{E}\{n\int \log [f(\h)]\mathcal{D}\h\}}$. Hence, we can approximate as $$\label{eq:Gamma1}
\Gamma=\exp\l\{n\int P(\s_1) \int\log [f(\h)] \mathcal{D}\h \ d\s_1\r\}.$$ Considering , we rewrite as $$\label{eq:EZn4}
\mathbb{E}_{\F,\s,\z}[Z^n]=\int \operatorname{e}^{nN\widetilde{\Phi}_J(m,\widehat{m},q,\widehat{q},Q,\widehat{Q})}
dm\ d\widehat{m}\ dq\ d\widehat{q}\ dQ\ d\widehat{Q},$$ where $\widetilde{\Phi}_J(m,\widehat{m},q,\widehat{q},Q,\widehat{Q})$ is given in .
**Free energy expression**: We now substitute into . Assuming that the limits in commute and that we only evaluate at optimum points of $\widetilde{\Phi}_J$ [@Barbier2015; @Krzakala2012probabilistic; @krzakala2012statistical], we have $\mathcal{F}=\widetilde{\Phi}_J(m^*,\widehat{m}^*,q^*,\widehat{q}^*,Q^*,\widehat{Q}^*)$, where the asterisks denote stationary points. Next, we calculate the stationary points: $$\frac{\partial \widetilde{\Phi}_J }{\partial m}=0 \Rightarrow \widehat{m}^*=\frac{R}{Q^*-q^*+\Delta},$$ $$\frac{\partial \widetilde{\Phi}_J }{\partial q}=0 \Rightarrow \widehat{q}^*=R\frac{\Delta+\rho-2m^*+q^*}{(Q^*-q^*+\Delta)^2},$$ $$\frac{\partial \widetilde{\Phi}_J }{\partial Q}=0 \Rightarrow \widehat{Q}^*=R\frac{2m^*-\rho-2q^*+Q^*}{(Q^*-q^*+\Delta)^2},$$ where $R$ is the measurement rate. Because we are analyzing the MMSE, we must assume that the estimated prior matches the true underlying prior, which is a Bayesian setting. Thus, $q^*=m^*$ and $Q^*=\rho$ . Let $E=q^*-2m^*+Q^*=Q^*-q^*$, then we obtain $\widehat{q}^*=\widehat{m}^*=\frac{R}{E+\Delta}$ and $\widehat{Q}^*=0$. Therefore, we solve for the free energy as a function of $E$ in . Using a change of variables, we obtain , which is a function of $E$. Using , the MSE is $$\label{eq:DandE}
D=E+Q-q=E+\rho/N\overset{N\rightarrow \infty}{\longrightarrow} E.$$ Hence, in the large system limit, we can regard the free energy as a function of the MSE, $D$.
[10]{}
E. Candès, J. Romberg, and T. Tao, “Robust uncertainty principles: [E]{}xact signal reconstruction from highly incomplete frequency information,” , vol. 52, no. 2, pp. 489–509, Feb. 2006.
D. Donoho, “Compressed sensing,” , vol. 52, no. 4, pp. 1289–1306, Apr. 2006.
R. G. Baraniuk, “A lecture on compressive sensing,” , vol. 24, no. 4, pp. 118–121, July 2007.
H. Jung, J. C. Ye, and E. Y. Kim, “[Improved k-t [BLAST]{} and k-t [SENSE]{} using [FOCUSS]{}]{},” , vol. 52, no. 11, pp. 3201–3226, May 2007.
H. Jung, K. Sung, K. S. Nayak, E. Y. Kim, and J. C. Ye, “[k-t [FOCUSS]{}: a general compressed sensing framework for high resolution dynamic [MRI]{}]{},” , vol. 61, no. 1, pp. 103–116, Jan. 2009.
S. F. Cotter and B. D. Rao, “Sparse channel estimation via matching pursuit with application to equalization,” , vol. 50, no. 3, pp. 374–377, Mar. 2002.
J. Ma and F. X. Le Dimet, “[Deblurring from highly incomplete measurements for remote sensing]{},” , vol. 47, no. 3, pp. 792–802, Mar. 2009.
J. Haupt and R. Nowak, “[S]{}ignal reconstruction from noisy random projections,” , vol. 52, no. 9, pp. 4036–4048, Sept. 2006.
M. F. Duarte, M. B. Wakin, D. Baron, S. Sarvotham, and R. G. Baraniuk, “Measurement bounds for sparse signal ensembles via graphical models,” , vol. 59, no. 7, pp. 4280–4289, July 2013.
G. J. Pottie and W. J. Kaiser, “[Wireless integrated network sensors]{},” , vol. 43, no. 5, pp. 51–58, May 2000.
M. F. Duarte, M. B. Wakin, D. Baron, and R. G. Baraniuk, “[Universal distributed sensing via random projections]{},” in [*Proc. Inf. Process. Sensor Networks (IPSN)*]{}, Nashville, TN, Apr. 2006.
J. Chen and X. Huo, “Theoretical results on sparse representations of multiple measurement vectors,” , vol. 54, no. 12, pp. 4634–4643, Dec. 2006.
S. F. Cotter, B. D. Rao, K. Engan, and K. Kreutz-Delgado, “Sparse solutions to linear inverse problems with multiple measurement vectors,” , vol. 53, no. 7, pp. 2477–2488, July 2005.
M. Mishali and Y. C. Eldar, “Reduce and boost: recovering arbitrary sets of jointly sparse vectors,” , vol. 56, no. 10, pp. 4692–4702, Oct. 2009.
E. Berg and M. P. Friedlander, “Joint-sparse recovery from multiple measurements,” , Apr. 2009.
O. Lee, J. M. Kim, Y. Bresler, and J. C. Ye, “Compressive diffuse optical tomography: [N]{}oniterative exact reconstruction using joint sparsity,” , vol. 30, no. 5, pp. 1129–1142, May 2011.
K. Lee, Y. Bresler, and M. Junge, “Subspace methods for joint sparse recovery,” , vol. 58, no. 6, pp. 3613–3641, June 2012.
J. C. Ye, J. M. Kim, and Y. Bresler, “Improving [M-SBL]{} for joint sparse recovery using a subspace penalty,” , vol. 63, no. 24, pp. 6595–6605, Dec. 2015.
J. A. Tropp, A. C. Gilbert, and M. J. Strauss, “Algorithms for simultaneous sparse approximation. part [I]{}: Greedy pursuit.,” , vol. 86, no. 3, pp. 572–588, Mar. 2006.
D. Malioutov, M. Cetin, and A. S. Willsky, “A sparse signal reconstruction perspective for source localization with sensor arrays,” , vol. 53, no. 8, pp. 3010–3022, Aug. 2005.
J. A. Tropp, “Algorithms for simultaneous sparse approximation. part [II]{}: Convex relaxation,” , vol. 86, no. 3, pp. 589–602, Mar. 2006.
J. Ziniel and P. Schniter, “Efficient message passing-based inference in the multiple measurement vector problem,” in [*Proc. IEEE 45th Asilomar Conf. Signals, Syst., and Comput.*]{}, Nov. 2011, pp. 1447–1451.
T. Tanaka, “[A statistical-mechanics approach to large-system analysis of CDMA multiuser detectors]{},” , vol. 48, no. 11, pp. 2888–2910, Nov. 2002.
D. Guo and S. Verd[ú]{}, “Randomly spread [CDMA]{}: [A]{}symptotics via statistical physics,” , vol. 51, no. 6, pp. 1983–2010, June 2005.
A. Montanari and D. Tse, “Analysis of belief propagation for non-linear problems: The example of [CDMA]{} (or: How to prove [T]{}anaka’s formula),” in [*IEEE Inf. Theory Workshop*]{}, Mar. 2006, pp. 160–164.
F. Krzakala, M. M[é]{}zard, F. Sausset, Y. Sun, and L. Zdeborov[á]{}, “Probabilistic reconstruction in compressed sensing: [A]{}lgorithms, phase diagrams, and threshold achieving matrices,” , vol. 2012, no. 08, pp. P08009, Aug. 2012.
F. Krzakala, M. M[é]{}zard, F. Sausset, Y. Sun, and L. Zdeborov[á]{}, “Statistical-physics-based reconstruction in compressed sensing,” , vol. 2, no. 2, pp. 021005, May 2012.
M. M[é]{}zard and A. Montanari, , Oxford University press, 2009.
J. Barbier and F. Krzakala, “Approximate message-passing decoder and capacity-achieving sparse superposition codes,” , Mar. 2015.
T. Lesieur, F. Krzakala, and L. Zdeborov[á]{}, “Phase transitions in sparse [PCA]{},” in [*Proc. Int. Symp. Inf. Theory (ISIT)*]{}, July 2015, pp. 1635–1639.
D. L. Donoho, A. Maleki, and A. Montanari, “[Message passing algorithms for compressed sensing]{},” , vol. 106, no. 45, pp. 18914–18919, Nov. 2009.
D. Baron, S. Sarvotham, and R. G. Baraniuk, “Bayesian compressive sensing via belief propagation,” , vol. 58, pp. 269–280, Jan. 2010.
M. Bayati and A. Montanari, “The dynamics of message passing on dense graphs, with applications to compressed sensing,” , vol. 57, no. 2, pp. 764–785, Feb. 2011.
A. Montanari, “Graphical models concepts in compressed sensing,” , pp. 394–438, 2012.
D. Guo and C. C. Wang, “Multiuser detection of sparsely spread [CDMA]{},” , vol. 26, no. 3, pp. 421–431, Apr. 2008.
J. Barbier, M. Dia, N. Macris, and F. Krzakala, “The mutual information in random linear estimation,” , July 2016.
G. Reeves and H. D. Pfister, “The replica-symmetric prediction for compressed sensing with gaussian matrices is exact,” , July 2016.
J. Ziniel and P. Schniter, “Efficient high-dimensional inference in the multiple measurement vector problem,” , vol. 61, no. 2, pp. 340–354, Jan. 2013.
J. Zhu and D. Baron, “Performance regions in compressed sensing from noisy measurements,” in [*Proc. 2013 Conf. Inference Sci. Syst. (CISS)*]{}, Baltimore, MD, Mar. 2013.
Y. Wu and S. Verd[ú]{}, “Optimal phase transitions in compressed sensing,” , vol. 58, no. 10, pp. 6241 – 6263, Oct. 2012.
J. Kim, W. Chang, B. Jung, D. Baron, and J. C. Ye, “Belief propagation for jointly sparse recovery,” , Feb. 2011.
D. L. Donoho, A. Maleki, and A. Montanari, “The noise-sensitivity phase transition in compressed sensing,” , vol. 57, no. 10, pp. 6920–6941, Oct. 2011.
A. Javanmard and A. Montanari, “State evolution for general approximate message passing algorithms, with applications to spatial coupling,” , Dec. 2012.
D. Donoho, I. Johnstone, and A. Montanari, “Accurate prediction of phase transitions in compressed sensing via a connection to minimax denoising,” , vol. 59, no. 6, pp. 3396–3433, June 2013.
M. Bayati, M. Lelarge, and A. Montanari, “Universality in polytope phase transitions and message passing algorithms,” , vol. 25, no. 2, pp. 753–822, Feb. 2015.
“Hubbard–[S]{}tratonovich transformation,” https://en.wikipedia.org/wiki/\
Hubbard%E2%80%93Stratonovich\_transformation.
[^1]: The work was supported in part by the National Science Foundation under the Grant CCF-1217749, the U.S. Army Research Office under the Contract W911NF-14-1-0314, and the European Research Council under the European Union’s 7th Framework Programme (FP/2007- 2013)/ERC Grant Agreement 307087-SPARCS.
[^2]: Junan Zhu and Dror Baron are with the Department of Electrical and Computer Engineering, NC State University, Raleigh, NC 27695. E-mail: {jzhu9, barondror}@ncsu.edu.
[^3]: Florent Krzakala is with Sorbonne Universit[é]{}s, Universit[é]{} Pierre et Marie Curie Paris 6 and Ecole Normale Superieure, 75005 Paris, France. E-mail: florent.krzakala@ens.fr.
[^4]: Part of the literature [@Tanaka2002; @GuoVerdu2005] defines the free energy as the negative of , so that fixed points of the free energy correspond to local minima.
[^5]: Recently, the replica Gibbs free energy has been proven rigorously for the SMV case by Barbier et al. [@BDMK2016] and Reeves and Pfister [@ReevesPfister2016]. We conjecture that by generalizing these two works [@BDMK2016; @ReevesPfister2016], our MMV analysis can be made rigorous; we leave it for future work.
[^6]: The $J$-dimensional integrals in can be simplified to one-dimensional integrals using a change of coordinate to $J$-sphere coordinate. Note also that $E$ approaches the MSE in the large system limit; details appear in the appendix.
[^7]: In Section \[sec:complex\], we only deal with SMV CS, and omit the word “SMV."
[^8]: A replica analysis for complex CS with a real measurement matrix appears in Guo and Verd[ú]{} [@GuoVerdu2005]. Their derivation does not cover complex matrices.
[^9]: As a reminder, the free energy of MMV-2 is identical to that of MMV-1 in the large system limit.
[^10]: The MMV with $J=1$ becomes an SMV. The MMSE results in Fig. \[fig:MMV\_J1\] match with the SMV MMSE in Krzakala et. al. [@Krzakala2012probabilistic; @krzakala2012statistical] and Zhu and Baron [@ZhuBaronCISS2013].
[^11]: When the assumptions about the measurement matrix and signal [@DMM2009; @Montanari2012; @Bayati2011; @Krzakala2012probabilistic; @krzakala2012statistical; @Barbier2015] are violated, AMP might suffer from convergence issues.
[^12]: Note that Algorithm \[algo:AMP\_MMV\] is a straightforward simplification of the AMP algorithm in Barbier and Krzakala [@Barbier2015].
[^13]: We simulated both $J$ different measurement matrices $\underline{F}^j$ and $J$ identical $\underline{F}^j$. Both results match the predicted BP MSE, which support our conclusion that the MMSE’s of both settings are the same. Fig. \[fig:AMPoverMSE\] is with $J$ different $\underline{F}^j$.
|
{
"pile_set_name": "ArXiv"
}
|
Russia began its biggest war games since the fall of the Soviet Union on Tuesday close to its border with China, mobilizing 300,000 troops in a show of force that will include joint exercises with the Chinese army.
China and Russia have staged joint drills before but not on such a large scale, and the Vostok-2018 exercise signals closer military ties as well as sending an unspoken reminder to Beijing that Moscow is able and ready to defend its sparsely populated far east.
Vostok-2018 is taking place at a time of heightened tension between the West and Russia, and NATO has said it will monitor the exercise closely, as will the United States which has a strong military presence in the Asia-Pacific region.
With its Vostok 2018 exercise, Russia sends a message that it regards the U.S. as a potential enemy and China as a potential ally. - Dmitri Trenin, Carnegie Moscow Centre
Russia's Ministry of Defence broadcast images on Tuesday of military trucks being transported on trains, columns of tanks, armoured vehicles and warships on the move, and combat helicopters and fighter aircraft taking off.
In one clip, marines from Russia's Northern Fleet and a motorized Arctic brigade were shown disembarking from a large landing ship on a barren shore opposite Alaska.
This activity was part of the first stage of the exercise, which runs until Sept. 17, the ministry said in a statement. It involved deploying additional forces to Russia's far east and a naval build-up involving its northern and pacific fleets.
The main aim was to check the military's readiness to move troops large distances, to test how closely infantry and naval forces co-operated, and to perfect command and control procedures. Later stages will involve rehearsals of both defensive and offensive scenarios.
Russia also staged a major naval exercise in the eastern Mediterranean this month and its jets resumed bombing the Syrian region of Idlib, the last major enclave of rebels fighting its ally President Bashar al-Assad.
China, Mongolia participating
The location of the main training range for Vostok-2018 5,000 kilometres east of Moscow means it is likely to be watched closely by Japan, North and South Korea as well as by China and Mongolia, both of whose armies will take part in the manoeuvres later this week.
Analysts say Moscow had to invite the Chinese and Mongolian militaries given the proximity of the war games to their borders and because the scale meant the neighbouring countries would probably have seen them as a threat had they been excluded.
Russian armoured personnel carriers roll along during the military exercises in eastern Siberia. One of the aims of the drills is to check the military's readiness to move troops large distances. (Russian Defence Ministry Press Service pool photo via Associated Press)
By chance or design, the exercise — which will involve more than 1,000 military aircraft, two Russian naval fleets and all Russian airborne units — is taking place while President Vladimir Putin holds talks with Chinese President Xi Jinping in the Russian port city of Vladivostok.
Relations between Moscow and Beijing have long been marked by mutual wariness with Russian nationalists warning of encroaching Chinese influence in the country's mineral-rich far east.
But Russia pivoted east toward China after the West sanctioned Moscow over its annexation of Ukraine's Crimea region in 2014 and trade links between the two, who share a land border over 4,200 kilometres long, have blossomed since.
Putin appearance expected
Russia said 24 helicopters and six jets belonging to the Chinese air force had moved to Russian air bases for the exercise. Beijing has said 3,200 members of the People's Liberation Army (PLA) will join in.
Some experts see the war games as a message to Washington, with which both Moscow and Beijing have strained ties.
"With its Vostok 2018 exercise, Russia sends a message that it regards the U.S. as a potential enemy and China as a potential ally," wrote Dmitri Trenin, a former Russian army colonel and director of the Carnegie Moscow Centre think tank.
Russian military helicopters fly in the Chita region during the military exercises on Tuesday. The drills are scheduled to run until next Monday. (Russian Defence Ministry Press Service via Associated Press)
"China, by sending a PLA element to train with the Russians, is signalling that U.S. pressure is pushing it toward much closer military co-operation with Moscow."
Putin, who is armed forces commander-in-chief, is expected to observe the exercises this week alongside Defence Minister Sergei Shoigu, who is overseeing them.
Shoigu has said they are the biggest since a Soviet military exercise, Zapad-81 in 1981.
|
{
"pile_set_name": "OpenWebText2"
}
|
Microsoft cloud across Europe…
Businesses and institutions in Europe are embracing the cloud to digitally transform… Microsoft state that while many are initially drawn to the cloud for cost savings, that greater agility and reduced IT complexity are quickly becoming fundamental to their competitiveness.
Microsoft cloud services – including Azure, Office 365 and Dynamics CRM online – became the first major cloud services to make the EU model clauses available to enterprise customers. Microsoft was also the first major provider to have its approach to commercial agreements validated by the European Union’s Article 29 Working Party, and adopted the world’s first international standard for cloud privacy. The British Standards Institute (BSI) has also independently verified that in addition to Microsoft Azure, both Office 365 and Dynamics CRM online are aligned with the standard’s code of practice for the protection of Personally Identifiable Information (PII) in the public cloud.
Hover over the dots on the map below to learn how customers across Europe are using the Microsoft cloud to help them achieve more.
|
{
"pile_set_name": "Pile-CC"
}
|
[Activate vitamin D3 or bisphosphonate in glucocorticoid-induced osteoporosis].
Glucocorticoid therapy is associated with bone loss starting soon after the therapy is initiated and an increased risk of fracture. Activated forms of vitamin D(3), which are frequently prescribed for postmenopausal osteoporosis in Japan, are one of a useful agent to prevent glucocorticoid-induced osteoporosis. In this review, the effect of activate form vitamin D(3) is compared with plain vitamin D or bisphosphonate.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
1. Technical Field
The present invention relates generally to semiconductor technology and more specifically to reducing the number of steps in forming MirrorBit(copyright) Flash memory.
2. Background Art
Different types of memories have been developed in the past as electronic memory media for computers and similar systems. Such memories include electrically erasable programmable read only memory (EEPROM) and electrically programmable read only memory (EPROM). Each type of memory had advantages and disadvantages. EEPROM can be easily erased without extra exterior equipment but with reduced data storage density, lower speed, and higher cost. EPROM, in contrast, is less expensive and has greater density but lack erasability.
A newer type of memory called xe2x80x9cFlashxe2x80x9d EEPROM, or Flash memory, has become extremely popular because it combines the advantages of the high density and low cost of EPROM with the electrical erasability of EEPROM. Flash memory can be rewritten and can hold its contents without power. It is used in many portable electronic products, such as cell phone, portable computers, voice recorders, etc. as well as in many larger electronic systems, such as cars, planes, industrial control systems, etc.
In Flash memory, bits of information are programmed individually as in the older types of memory, such as dynamic random access memory (DRAM) and static random access memory (SRAM) memory chips. However, in DRAMs and SRAMs where individual bits can be erased one at a time, Flash memory must currently be erased in fixed multi-bit blocks or sectors.
Conventionally, Flash memory is constructed of many Flash memory cells where a single bit is stored in each memory cell and the cells are programmed by hot electron injection and erased by Fowler-Nordheim tunneling. However, increased market demand has driven the development of Flash memory cells to increase both the speed and the density. Newer Flash memory cells have been developed that allow more than a single bit to be stored in each cell.
One memory cell structure involves the storage of more than one level of charge to be stored in a memory cell with each level representative of a bit. This structure is referred to as a multi-level storage (MLS) architecture. Unfortunately, this structure inherently requires a great deal of precision in both programming and reading the differences in the levels to be able to distinguish the bits. If a memory cell using the MLS architecture is overcharged, even by a small amount, the only way to correct the bit error would be to erase the memory cell and totally reprogram the memory cell. The need in the MLS architecture to precisely control the amount of charge in a memory cell while programming also makes the technology slower and the data less reliable. It also takes longer to access or xe2x80x9creadxe2x80x9d precise amounts of charge. Thus, both speed and reliability are sacrificed in order to improve memory cell density.
An even newer technology allowing multiple bits to be stored in a single cell is known as xe2x80x9cMirrorBit(copyright)xe2x80x9d Flash memory has been developed. In this technology, a memory cell is essentially split into two identical (mirrored) parts, each of which is formulated for storing one of two independent bits. Each MirrorBit Flash memory cell, like a traditional Flash cell, has a gate with a source and a drain. However, unlike a traditional Flash cell in which the source is always connected to an electrical source and the drain is always connected to an electrical drain, each MirrorBit Flash memory cell can have the connections of the source and drain reversed during operation to permit the storing of two bits.
The MirrorBit Flash memory cell has a semiconductor substrate with implanted conductive bitlines. A multilayer storage layer, referred to as a xe2x80x9ccharge-trapping dielectric layerxe2x80x9d, is formed over the semiconductor substrate. The charge-trapping dielectric layer can generally be composed of three separate layers: a first insulating layer, a charge-trapping layer, and a second insulating layer. Wordlines are formed over the charge-trapping dielectric layer perpendicular to the bitlines. Programming circuitry controls two bits per cell by applying a signal to the wordline, which acts as a control gate, and changing bitline connections such that one bit is stored by source and drain being connected in one arrangement and a complementary bit is stored by the source and drain being interchanged in another arrangement.
Programming of the cell is accomplished in one direction and reading is accomplished in a direction opposite that in which it is programmed.
A major problem has been that the diffusion of the bitlines limits how closely the bitlines can be placed and this in turn limits how far the memory cells can be reduced in size.
With the urgency of reducing device size, a solution to this problem has been long sought but has long eluded those skilled in the art.
The present invention provides a method of manufacturing an integrated circuit, which includes providing a semiconductor substrate and depositing a charge-trapping dielectric layer and a gate dielectric layer over the semiconductor substrate. Bitlines are implanted closely in the semiconductor substrate and annealed using a rapid thermal anneal, which allows the bitlines to be more closely spaced than possible with the prior art. Wordlines and gates are formed and source/drain junctions are implanted in the semiconductor substrate. An interlayer dielectric layer is deposited and the integrated circuit completed.
|
{
"pile_set_name": "USPTO Backgrounds"
}
|
Elevation of body temperature to between 41.8 and 42.4 degrees centigrade is now being used both alone and together with chemotherapeutic agents to treat patients with cancer. The mechanisms by which hyperthermia is cytotoxic are poorly understood as are the ways in which elevated temperature affects cell killing by chemotherapeutic agents. We propose to study the interaction between hyperthermia and selected chemotherapeutic drugs utilizing both in vitro and in vivo systems. Our general approach will be to test for synergistic cytotoxicity by measuring in vitro colony survival in chinese hamster ovary cells and by measuring the in vitro colony survival of RT9 astrocytoma cells after in vivo treatments. We then plan to study the mechanism behind the observed increased cytotoxicity by examining such things as altered drug uptake and enzyme inactivation. We will also study the scheduling of the drugs with heat in order to maximize cell kill. These studies should suggest clinically useful combinations of drugs with heat as well as provide basic cellular pharmacologic data. This work is already underway and we have discovered an interesting relationship between methotrexate and adriamycin cytotoxicity when each is combined with heat. In addition we have found that the rate of heating to peak treatment temperature is a critical determinant of the cell kill achieved. Our ultimate aim is to use these studies as starting points in the selection of drugs and hyperthermia schedules for clinical use.
|
{
"pile_set_name": "NIH ExPorter"
}
|
the technology leader you deserve
Projects
GlobeChat
UCSD
Contract to upgrade Inventory control systems from .Net 1.1 to current technologies and structure. Reducing code by as much as 50% through implementation of common API and reusable Nuget and clientside components.
Victiv
Intense front end development to bring the action from the sports field to the browser in under six seconds! Latest front end technologies used to five daily Fantasy Sports the best user experience possible.
DefenseWeb
Building applications to support the Department of Defense. From Family and Relocation Services to HealthCare and Wounded Warrior Counseling, applications built on top of a proprietary content management system used to help our countries service men and woman.
I can help!
DAVID WAYNE BAXTER
My career has been dedicated to start-up and growing companies by building their technology teams and increase product offering. The thrill of creating, leading and contributing to a team from inception to successful project delivery is my primary goal with any organization. Helping to choose the right technologies, the proper architecture, and implement the necessary process to produce the right solution for the organization is an essential skill that I enjoy continuing to strengthen through every project. I pride myself in being able to quickly learn, adapt and change to meet the needs of the project, the team, and the organization.
In my free time enjoy keeping current with changing technology through online course-ware and personal projects. In addition to professional Microsoft development, I spend much of my personal time researching and working with open source and emerging technologies.
Microsoft .Net, C#, WebAPI, MVC, SQL
Microsoft Certified Professional and .Net architect for nearly two decades I've seen it all from Classic ASP to ASP Core. Whether you've got an existing Microsoft Based project or you're looking to dive into the Microsoft universe, I have the experience you need to make sure you're project is a success.
Linux, NodeJS, RoR, BackBone, React and More!
Open Source and Emerging technologies are impossible to ignore. They're rapidly taking over the market, and for good reason. With speed, agility and minimal cost, they make the ideal platform for companies wishing to run with low resources and cost. I make it a point to stay on top of current technologies through self-study and personal projects to make sure my clients get the tech stack they want and I can hit the ground running.
Continuous Delivery & Automation
According to the Three Virtues of a Great Programmer laziness, at least proverbially, is a virtue that drives me. Automation can free developers of the tedious tasks, while making sure they're executed with machine precision. Implementing Continuous integration, Delivery and Deployment, gives any team the confidence in knowing that their customers will receive updated projects quickly and accurately. I have extensive experience introducing and maintaining automation to new projects with great success.
Team Building
Interviewing is time consuming. Misappropriated and underutilized resources are costly.
Endorsements
David is exceptional both professionally and personally. He is highly organized, productive, motivated, self-driven, detailed with a work ethic that is unmatched. His technical skills are supurb and he is highly efficient and effective in both a CTO or Lead Developer Role. He is a problem solver and plain and simply gets things done. David takes initiative and and one of the finest people I have ever worked with in the past 30 years. He has produced incredible results for our company. In addition to David’s consistent top performance, he is a pleasure to work with personally with a positive attitude and fantastic sense of humor. Bottom line, David Baxter is an A+.
I have had the pleasure of working with David spanning several companies over the last decade. He is truly one of the most talented engineers and leaders I have worked with. David has a broad skill set ranging from front end, backend, server level, and management both of architectures and people. This broad knowledge combined with his drive to keep learning, means there is virtually no problem he can’t tackle. I have seen him challenged repeatedly and he always finds a way to bring success to the project and team he is working with.
David is also a great team member and brings a senior level vision to projects. He is great with mentoring more junior teammates and raising their level of achievement. He adds positively to the team and company culture, and is great to work with.
I would highly recommend David for any company needing to add a senior level architect or manager. He will bring a wealth of experience with him, and drive the team to accomplish great things.
|
{
"pile_set_name": "Pile-CC"
}
|
Teaching and Learning: from staff nurse to nurse consultant. Part 7: teaching a skill.
I am always fascinated to watch people who are very competent at a particular task performing that skill. There is a confident ease that marks out a skilled performer but if you look deeper below the actual physical movements you can observe that the almost intuitive actions are guided by visual cues and sensory feedback. However, it is one thing to perform a skill but it is another to teach it.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Q:
Image value was null when posted to controller
in controller image comes null i tried this code but it did not work
function bindForm(dialog) {
$("form", dialog).submit(function () {
var formdata = new FormData($('form').get(0));
$.ajax({
url: "/Books/Create",
type: "POST",
dataType: "JSON",
data: formData,
contentType: false,
processData: false,
success: function (result) {
if (result.success) {
$("#myModal").modal("hide");
location.reload();
} else {
$("#myModalContent").html(result);
bindForm();
}
}
});
return false;
});
}
A:
$(function () {
$.ajaxSetup({ cache: false });
$("a[data-modal]").on("click", function (e) {
$("#myModalContent").load(this.href, function () {
$("#myModal").modal({
keyboard: true
}, "show");
bindForm(this);
});
return false;
});
});
function bindForm(dialog) {
$("#crtForm", dialog).submit(function () {
var myform = document.getElementById("crtForm");
var fdata = new FormData(myform);
$.ajax({
url: this.action,
data: fdata,
cache: false,
processData: false,
contentType: false,
type: "POST",
success: function (result) {
if (result.success) {
$("#myModal").modal("hide");
location.reload();
} else {
$("#myModalContent").html(result);
bindForm();
}
}
});
return false;
});
}
|
{
"pile_set_name": "StackExchange"
}
|
Pages
Monday, November 2, 2009
JCI Junior ChungLing Butterworth Installation (Ep.1)
On the 24th of October I had join my society JCI Junior attended the Installation at ChungLing Butterworth. I was surprise that 40 peoples could turn up on that day and really made me glad. From that on, I realized that our members mostly like the outing activities. XD! By the way, we took bus there. It's not a near distance, I just realized that their school was located at Permatang Pauh which very deep inside rural area. Along the way we could see the factories and paddy fields.
Our Board Of Directors were preparing for their costumes and adjusted for their clothes neatly.
She's the organizer who named JoXian. She is a good organizer and she planned everything with her committees to arrange all the programs to make sure it run smoothly. We were very glad to get the organizer invitation at the registration counter! Thank you, XD!
As you can see my junior Li Yang was writing his name and put his signature on the invitation paper. Let me introduce to you, he is my president's brother which is now Form 1 and holding the Form Representatives position. I think he will be a good leader on next time! Add oil Li Yang...
They provided us a simple but nice booklet. This let me think about reduce campaign and economic crisis case. Nowadays don't waste too much money on making such things not so important. I really like their design which all hand made! Although it's simple but the contents all included! It can let me simply flip through and get into the information and new B.O.D contacts.
The JCI Junior of ChungLing Butterworth were performing a dance with the theme song of "I Got The Feeling" to open up the whole ceremony. The dance were presented by the juniors which aged between 14 and 15 years old.
Later it's speeching time, I managed to take a shot while the JCI Senator Teoh Eng Keat gave his speech. He was talking about the joint installation concept in order to gather up all JCI Junior and held a grand convention next year. Because of the request and curious from the Senior Assistant of ChungLing Butterworth, they had think of this concept. So I am looking forward for the grand event soon in year 2010! XD!
If I'm not mistaken, most of the Installation were using the buffet concept. It's more low costing and most of the people will like it. Besides that, if you feel that not enough still can get again from the stall. I love the food actually because I didn't taste this delicious food before in all the installation. I was full enough after having this buffet dinner, it's worth it!
If any installation are upcoming and you would like to invited me, you can simply email me at least 1 week earlier to let me arrange my time. I will try my best to attend on the occasion. *try to get my phone number via email, thank you! Posted by GenYong 02.11.2009
2 Month Loan Lender is such kind of financial support that is arranged for those who are carrying a bad credit impairments including those omitted installments, arrears, insolvency, bankruptcy or even CCJs etc. they are not considered as good indicators of a person when he is trying to get money but these loans supporting such people without any difficulty.Payday Loan For People On Benefits
Nuffnang Ads
GenYong UNITY ChatBox
Follow Me
Twitter
Subscribe via email
Search This Blog
About Me
I am already a 17 years old boy and although I have Facebook account, I still love in blogging. As well if you're my reader that you found me since September of 2007. Anyway, just a short introduction here and welcome you to visit my blog. There's more features bring by me on next! Want to see more photos? Check it out for the updates! Thanks.
|
{
"pile_set_name": "Pile-CC"
}
|
David McCarty
David Andrew "Dave" McCarty (born November 23, 1969) is a former first baseman and outfielder in Major League Baseball. From 1993 through 2005, McCarty played with the Minnesota Twins (1993–1995), San Francisco Giants (1995–1996), Seattle Mariners (1998), Kansas City Royals (2000–2002), Tampa Bay Devil Rays (2002), Oakland Athletics (2003) and Boston Red Sox (2003–2005). He batted right-handed and threw left-handed.
Career
McCarty attended Stanford University, and in 1989 he played collegiate summer baseball with the Cotuit Kettleers of the Cape Cod Baseball League. He was drafted by the Twins in the 1st round (3rd overall) of the 1991 Major League Baseball draft. In an eleven-season career, he was a .242 hitter with 36 home runs and 175 RBI in 630 games.
A utility player in the purest sense, McCarty was a defensive first baseman and outfielder who occasionally was called upon to pitch. His most productive season came in 2000 for the Royals, when he posted career-highs in batting average (.278), home runs (12), RBI (68), runs (34), hits (75), doubles (14) and games played (103).
McCarty, who was released by the Boston Red Sox in May 2005 after the team signed first baseman, John Olerud, refused a minor league assignment. He retired and was a Red Sox analyst on NESN from July 1, 2005 until the end of the 2008 season.
McCarty was an oddity in MLB in that he batted right-handed and threw left-handed, and was a position player as opposed to a pitcher.
Pitching
McCarty made three pitching appearances for the Red Sox in 2004. The first was during the April 9 home opener against the Toronto Blue Jays. The second was in the June 12 game against the Los Angeles Dodgers in which he struck out Jayson Werth and finally, in the final game of the season, McCarty went 2 scoreless innings against the Baltimore Orioles in which he struck out Rafael Palmeiro, Larry Bigbie and David Newhan.
Personal life
Born in Houston, Texas, McCarty graduated from Sharpstown High School in 1988 before attending Stanford University. He lives in Piedmont, California with his wife, Monica, and their two children.
References
External links
McCarty expands repertoire - Jayson Stark, ESPN
Category:1969 births
Category:Living people
Category:Boston Red Sox players
Category:Kansas City Royals players
Category:Major League Baseball first basemen
Category:Major League Baseball outfielders
Category:Baseball players from Texas
Category:Minnesota Twins players
Category:Oakland Athletics players
Category:People from Houston
Category:San Francisco Giants players
Category:Seattle Mariners players
Category:Stanford Cardinal baseball players
Category:Cotuit Kettleers players
Category:Tampa Bay Devil Rays players
Category:People from Piedmont, California
Category:Orlando Sun Rays players
Category:Visalia Oaks players
Category:Portland Beavers players
Category:Salt Lake Buzz players
Category:Indianapolis Indians players
Category:Phoenix Firebirds players
Category:Tacoma Rainiers players
Category:Toledo Mud Hens players
Category:Durham Bulls players
Category:Sacramento River Cats players
Category:Lowell Spinners players
Category:Pawtucket Red Sox players
Category:All-American college baseball players
|
{
"pile_set_name": "Wikipedia (en)"
}
|
US Tax Writer: Overhaul Can't be Piecemeal
An overhaul of the voluminous and complex U.S. tax code can not be tackled piecemeal, the Republican chairman of the tax-writing committee in the U.S. House of Representatives said Tuesday.
"There is quite a bit of groundwork that needs to be done," Representative Dave Camp, chairman of the House Ways and Means Committee, told Reuters before starting a congressional debate on a tax code revamp with a hearing on Thursday.
"The concern about addressing certain issues that affect larger business is you leave out small businesses or others, which is much of our economy," said Camp, who became chairman of the powerful committee when Republicans took control of the House this month.
Big multinational companies say the top 35 percent corporate tax rate hamstrings them competitively against foreign-based rivals, most of whom are subject to far lower rates.
Camp said a reworking of the tax code must address the individual and corporate tax rates together because so many businesses are organized as "pass through" entities that allow them to pay the individual tax rate, which also tops out at 35 percent but gives them other tax advantages.
It took several years to cement a deal that led to the last major tax reform act, brokered in 1986 between then-President Ronald Reagan, a Republican, and a Democratic-led House.
President Barack Obama, a Democrat, has said he wants a rewrite of the tax code, and started a debate on the corporate tax rate.
Republicans and Democrats agree the top corporate rate may hurt U.S. competitiveness but Democrats want the myriad of deductions and credits in the code -- many of which they dub as "loopholes" -- trimmed to fund any corporate tax cut.
There is less consensus on the personal tax rate.
In December, both sides agreed to extend for two years the top 35 percent rate but that was fervently opposed by many Democrats who say wealthier individuals can afford to pay more, especially in an age of deficits topping $1 trillion.
Procter & Gamble board chairman Robert McDonald will be among the witnesses at the Thursday hearing.
"This is going to be the first step in a long debate and discussion," Camp said.
An overhaul of the voluminous and complex U.S. tax code can not be tackled piecemeal, the Republican chairman of the tax-writing committee in the U.S. House of Representatives said Tuesday.
There is quite a bit of groundwork that needs to be done, Representative Dave...
|
{
"pile_set_name": "Pile-CC"
}
|
Q:
Getting local hour with NodaTime
New to NodaTime and I chose to use it instead of the BCL libraries because it removes a lot of ambiguity when dealing with dates. However, I can't seem to get it to do what I want. I have a date and time specified by year, month, day, hours, and minutes. I also have two timezones for which I need to display the "clock time". For example, if my input is December 15, 2015 at 3:30 PM and my two timezones are central standard and eastern standard (one hour apart), I expect my output to be
12/15/2015 3:30 PM Central
12/15/2015 4:30 PM Eastern
But I can only seem to get the central (local to me, if that matters) timezone. Here's my code:
var localDateTime = new LocalDateTime(
year: 2015,
month: 12,
day: 15,
hour: 15,
minute: 30,
second: 0
);
var centralTimeZone = DateTimeZoneProviders.Bcl.GetZoneOrNull("Central Standard Time");
var easternTimeZone = DateTimeZoneProviders.Bcl.GetZoneOrNull("Eastern Standard Time");
var centralTime = centralTimeZone.AtLeniently(localDateTime);
var easternTime = easternTimeZone.AtLeniently(localDateTime);
It seems that centralTime and easternTime are both ZonedDateTime objects whose times are 2015-12-10T15:30 with the correct offset i.e. centralTime is -6 and easternTime is -5.)
I just can't figure out how to get the output I want.
A:
It sounds like your initial date/time is actually in Central time - so once you've performed the initial conversion, you can just say "Hey, I want the same instant in time, but in a different time zone" - which isn't the same as "I want a different instant in time for the same local time". You want:
var centralTime = centralTimeZone.AtLeniently(localDateTime);
var easternTime = centralTime.InZone(easternTimeZone);
|
{
"pile_set_name": "StackExchange"
}
|
Q:
Unable to align sidebar and it contents in Element-UI in vue-js
I am a newbie in vue-js and currently using element-ui as UI component library. I am trying to use sidebar and have following code(click here for jsfiddle):-
<template>
<div>
<el-menu default-active="1" class="el-menu-vertical-demo" :collapse="true">
<el-menu-item index="1">
<i class="el-icon-money"></i>
<span slot="title">Consent</span>
</el-menu-item>
<el-menu-item index="2" @click="signOut">
<i class="el-icon-switch-button"></i>
<span slot="title">Logout</span>
</el-menu-item>
</el-menu>
<el-table
:data="tableData"
style="width: 100%">
<el-table-column
prop="arn"
label="ARN"
width="180">
</el-table-column>
<el-table-column
prop="phoneNumber"
label="Phone Number"
width="180">
</el-table-column>
<el-table-column
label="Status"
prop="address">
<el-badge value="Applied" class="item" type="primary"></el-badge>
</el-table-column>
</el-table>
</div>
</template>
<style>
</style>
<script>
import * as firebase from "firebase/app";
import "firebase/auth";
export default {
data() {
return {
tableData: [{
arn: 'XXXXXX1',
phoneNumber: 'XXX0',
address: 'No. 189, Grove St, Los Angeles'
}, {
arn: 'XXXXXX1',
phoneNumber: 'XXX0',
address: 'No. 189, Grove St, Los Angeles'
}, {
arn: 'XXXXXX1',
phoneNumber: 'XXX0',
address: 'No. 189, Grove St, Los Angeles'
}, {
arn: 'XXXXXX1',
phoneNumber: 'XXX0',
address: 'No. 189, Grove St, Los Angeles'
}]
}
},
methods: {
...
}
}
</script>
As shown in the above jsfiddle link, table isn't getting aligned towards right of sidebar. I searched about it in its documentation but couldn't get much of help.
Would appreciate any hint.
A:
Why not using the layout system of element-ui?
<el-container>
<el-aside></el-aside>
<el-main></el-main>
</el-container>
example: https://jsfiddle.net/dreijntjens/ba9zq7wt/10/
note: by default the aside has a width of 300px, if you only want to use the el-menu width you have to unset the width <el-aside width="unset"> or as I did on the tag itself or
.el-aside {
width: unset !important; // !important
}
|
{
"pile_set_name": "StackExchange"
}
|
Anti-inflammatory effect of a novel synthetic compound 1-((4-fluorophenyl)thio)isoquinoline in RAW264.7 macrophages and a zebrafish model.
The compound, 1-((4-fluorophenyl)thio)isoquinoline (FPTQ), is a synthetic isoquinoline derivative. To test the anti-inflammatory effect of FPTQ, we used neutrophil-specific transgenic zebrafish Tg(mpx::EGFP)i114 line and lipopolysaccharide (LPS)-stimulated RAW264.7 cells. We also used two different methods, involving tail transection and LPS stimulation in the zebrafish model. Neutrophils translocation in the zebrafish tail-transected model was inhibited by FPTQ. Neutrophil aggregation was also inhibited by FPTQ in the LPS-stimulated zebrafish model. Decreased mRNA expression of the pro-inflammatory cytokine genes, interleukin-1β (il-1β) and interleukin-6 (il-6), was found in zebrafish larvae injected with FPTQ. Additionally, production of nitric oxide was inhibited by FPTQ in RAW264.7 macrophage cells treated with LPS. Moreover, the mRNA expression of Il-1β and Il-6 suppressed by FPTQ treatment in RAW264.7 macrophage cells, and an enzyme immunoassay showed that FPTQ suppressed the secretion of IL-1β and IL-6 in RAW264.7 cells. These results demonstrate that FPTQ reduced inflammatory responses and, therefore, suggest that it may be effective as an anti-inflammatory agent.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Augusta Louise zu Stolberg-Stolberg
Countess Louise Augusta zu Stolberg-Stolberg (7 January 1753 in Bramstedt, Duchy of Holstein30 May 1835 in Kiel) is known for her lively correspondence with the poet and thinker Johann Wolfgang von Goethe; she is known as Goethes Gustchen in the history of literature.
She was the younger sister of Goethe's friends Count Friedrich Leopold zu Stolberg-Stolberg ("Fritz") and Count Christian zu Stolberg-Stolberg. She lived in a pension for young, unmarried girls from 1770 to 1783 along with the older Baroness Metta von Oberg. Her letters to the young Goethe date to 1775 and 1776. They never met in person.
In all her correspondence she was a lively writer. "Augusta – vom Morgen bis in Abend laufen die Depeschen bey ihr ein, wie bey einem Staatsminister, und werden sorgfältiger abgefertigt, als in einer Canzelley"
noted Friedrich Gottlieb Klopstock.
On 7 August 1783 Augusta Louise moved to Copenhagen, and later married the Danish Minister of State Andreas Peter Bernstorff.
Sources
Plath-Langheinrich, Elsa: Als Goethe nach Uetersen schrieb: Das Leben der Conventualin Augusta Louise Gräfin zu Stolberg-Stolberg. (de)
Koopmann, Helmut: Goethe und Frau von Stein. (de)
Goethe an Auguste Gräfin zu Stolberg, [Frankfurt, January 18.- 30. 1775]. (de)
External links
Brief biography and excerpts of Goethe, also with Gustchen(de)
Biographical Highlights: Auguste Louise, Gräfin zu Stolberg-Stolberg (de)
Category:1753 births
Category:1835 deaths
Category:German letter writers
Category:Women letter writers
Category:German philosophers
Augusta Louise
Category:People from Bad Bramstedt
Category:People from the Duchy of Holstein
|
{
"pile_set_name": "Wikipedia (en)"
}
|
2011 Wilmington Hammerheads FC season
The 2011 Wilmington Hammerheads season, is the club's 16th season in existence, and their second-consecutive year playing in the third division of American soccer, following a hiatus in 2009. This year marks the club's debut in the newly created USL Pro League.
Review and events
The club was on a hiatus in 2010, and returned to fielding a professional side this year. They joined the USL Pro, and were allocated into the National Division, which primarily consisted of Southern clubs.
The club signed four players from Ventura County Fusion on March 2, 2011: Dylan Riley, Ivan Becerra, Manny Guzman, and Jack Avesyan.
Match results
Legend
USL Pro
U.S. Open Cup
Club
Roster
As of June 10, 2011.
Management and staff
David Irving - Head Coach/Director of Soccer Operations
Roxanne DeMonte - Game Day Operations
League standings
American Division
Statistics
{|| class="wikitable sortable"
|-
!
!
!Player
!
!
!
!
!
!
!
!Acquired
!Salary
|- align=center
| ||2 ||align=left data-sort-value="Harvey" | Jordan Harvey
|DF ||15 ||15 ||0 ||1 ||3 ||1 ||align=left|Expansion Draft||$63,125
Transfers
In
Awards
None.
See also
2011 USL Pro season
2011 U.S. Open Cup
2011 in American soccer
Wilmington Hammerheads
References
Category:Wilmington Hammerheads FC seasons
Category:American soccer clubs 2011 season
Category:2011 USL Pro season
Category:2011 in sports in North Carolina
|
{
"pile_set_name": "Wikipedia (en)"
}
|
---
category: Components
subtitle: 穿梭框
type: 数据录入
cols: 1
title: Transfer
cover: https://gw.alipayobjects.com/zos/alicdn/QAXskNI4G/Transfer.svg
---
双栏穿梭选择框。
## 何时使用
- 需要在多个可选项中进行多选时。
- 比起 Select 和 TreeSelect，穿梭框占据更大的空间，可以展示可选项的更多信息。
穿梭选择框用直观的方式在两栏中移动元素，完成选择行为。
选择一个或以上的选项后，点击对应的方向键，可以把选中的选项移动到另一栏。其中，左边一栏为 `source`，右边一栏为 `target`，API 的设计也反映了这两个概念。
## API
### Transfer
| 参数 | 说明 | 类型 | 默认值 | 版本 |
| --- | --- | --- | --- | --- |
| DataSource | 数据源，其中的数据将会被渲染到左边一栏中，`targetKeys` 中指定的除外。 | List\<TransferItem> | null | |
| Disabled | 是否禁用 | bool | false | |
| Footer | 底部渲染函数 | OneOf<string, RenderFragment> | null | |
| Style | 两个穿梭框的自定义样式 | string | | |
| Operations | 操作文案集合，顺序从上至下 | string\[] | \['right', 'left'] | |
| Render | 每行数据渲染函数，该函数的入参为 `DataSource` 中的项，返回值为 OneOf<string, RenderFragment>。 | Func<TransferItem, OneOf<string, RenderFragment>> | | |
| SelectedKeys | 设置哪些项应该被选中 | string\[] | \[] | |
| ShowSearch | 是否显示搜索框 | bool | false | |
| ShowSelectAll | 是否展示全选勾选框 | bool | true | |
| TargetKeys | 显示在右侧框数据的 key 集合 | string\[] | \[] | |
| Titles | 标题集合，顺序从左至右 | string\[] | \['', ''] | |
| SelectAllLabels | 自定义顶部多选框标题的集合 | | | |
| OnChange | 选项在两栏之间转移时的回调函数 | TransferSelectChangeArgs | | |
| OnScroll | 选项列表滚动时的回调函数 | TransferScrollArgs | | |
| OnSearch | 搜索框内容时改变时的回调函数 | TransferSearchArgs | - | |
| OnSelectChange | 选中项发生改变时的回调函数 | TransferSelectChangeArgs | | |
### Render Props
Transfer 支持接收 `ChildContent` 自定义渲染列表，并返回以下参数：
| 参数 | 说明 | 类型 | 版本 |
| --------------- | -------------- | ----------------------------------- | ---- |
| direction | 渲染列表的方向 | `left` \| `right` | |
| disabled | 是否禁用列表 | bool | |
| filteredItems | 过滤后的数据 | List\<TransferItem> | |
| onItemSelect | 勾选条目 | (key: string, selected: boolean) | |
| onItemSelectAll | 勾选一组条目 | (keys: string[], selected: boolean) | |
| selectedKeys | 选中的条目 | string[] | |
## 注意
按照 Blazor 的[规范](http://facebook.github.io/react/docs/lists-and-keys.html#keys)，所有的组件数组必须绑定 key。在 Transfer 中，`DataSource`里的数据值需要指定 `Key` 值。对于 `DataSource` 默认将每列数据的 `Key` 属性作为唯一的标识。
如果你的数据没有这个属性，务必使用 `rowKey` 来指定数据列的主键。
|
{
"pile_set_name": "Github"
}
|
1. Introduction {#sec1-cells-09-00275}
===============
Calcium ion (Ca^2+^) is essential for life, regulating many crucial cellular functions like contraction, migration and proliferation \[[@B1-cells-09-00275]\]. Cytosolic Ca^2+^ signals are mainly controlled through a combination of two closely linked processes--Ca^2+^ release from the endoplasmic reticulum (ER) and Ca^2+^ entry across the plasma membrane (PM) \[[@B2-cells-09-00275]\]. Much of the former event is mediated by inositol 1,4,5-trisphosphate receptor (IP~3~Rs), Ca^2+^ release channels resident in ER membranes \[[@B3-cells-09-00275],[@B4-cells-09-00275],[@B5-cells-09-00275],[@B6-cells-09-00275]\]. As a major Ca^2+^ release channel on ER, the largest intracellular Ca^2+^ store in mammalian cells \[[@B7-cells-09-00275]\], IP~3~Rs often form large complexes with other Ca^2+^ handling proteins and Ca^2+^ channels \[[@B8-cells-09-00275]\]. Numerous reports have showed biochemical or co-localization evidence indicating that IP~3~Rs interact with PM Ca^2+^-transporting ATPase (PMCA) \[[@B9-cells-09-00275],[@B10-cells-09-00275]\], Sarco/endoplasmic reticulum Ca^2+^ ATPase 2 (SERCA2) or SERCA3 \[[@B10-cells-09-00275],[@B11-cells-09-00275]\], STIM1 \[[@B12-cells-09-00275]\], Orai1 channels \[[@B13-cells-09-00275],[@B14-cells-09-00275]\], and handling organelles like mitochondria \[[@B15-cells-09-00275]\] in different types of cells. Thus, IP~3~Rs are key elements of Ca^2+^ signalling machinery. IP~3~Rs interact with these proteins to spatiotemporally control Ca^2+^ signals and corresponding downstream events \[[@B8-cells-09-00275]\], aberrant IP~3~R-regulated-signalling are related to many types of human diseases \[[@B16-cells-09-00275]\]. Yet it still remains elusive how IP~3~Rs would coordinate these Ca^2+^ handling proteins to maintain ER Ca^2+^ homeostasis.
In mammalian cells, one ubiquitous Ca^2+^ entry pathway mediated by Ca^2+^ release activated Ca^2+^ (CRAC) channels is also essential for the maintaining of ER Ca^2+^ homeostasis \[[@B17-cells-09-00275],[@B18-cells-09-00275]\]. Its activation is closely linked with the function of IP~3~Rs \[[@B19-cells-09-00275],[@B20-cells-09-00275]\]. Activation of G-protein coupled receptors or Tyrosine kinase on PM would lead to the opening of immobile IP~3~Rs adjacent to ER-PM junctions \[[@B21-cells-09-00275]\], resulting in a transient increase in cytosolic Ca^2+^ levels and a decrease in ER Ca^2+^ content. The lowering of ER Ca^2+^ store would induce Ca^2+^ influxes from extracellular space through a process called store operated Ca^2+^ entry (SOCE), resulting in a sustained increase cytosolic Ca^2+^ levels \[[@B2-cells-09-00275]\]. SOCE is mediated by CRAC channels. Briefly, a decrease in ER Ca^2+^ levels is sensed by STIM1, the ER-resident Ca^2+^ sensing component of the CRAC channel. Activated STIM1 will then translocate into ER-PM junctions to bind with and activate Orai1, the pore forming subunit of CRAC channels, causing Ca^2+^ influxes \[[@B18-cells-09-00275]\]. Thus the activation of IP~3~Rs and CRAC channels is spatiotemporally coupled around ER-PM junctions \[[@B22-cells-09-00275]\]. Even though Ca^2+^ release signals generated by IP~3~Rs and Ca^2+^ influxes through CRAC channels occur at different subcellular regions, mediating different cellular processes \[[@B23-cells-09-00275],[@B24-cells-09-00275]\], IP~3~Rs and CRAC channels are functionally coupled \[[@B24-cells-09-00275],[@B25-cells-09-00275],[@B26-cells-09-00275]\]. There is one report showing that mutation in IP~3~Rs could inhibit SOCE in *Drosophila* cells, likely through impairments of coupling between STIM1-Orai1 \[[@B27-cells-09-00275]\]. However, it still remains elusive whether IP~3~Rs could regulate SOCE through other means.
To further dissect IP~3~Rs-centered, ER-mediated Ca^2+^ signalling in a genetic clean background, we generated IP~3~Rs triple and double knockout HEK cell lines (IP~3~Rs-TKO and IP~3~Rs-DKO) using CRISPR/Cas9 genome-editing technology. Using these engineered cells together with ER Ca^2+^ indicator CEPIA1ers (Calcium-measuring organelle-Entrapped Protein IndicAtor 1 in the ER) \[[@B28-cells-09-00275]\], we demonstrated that even though IP~3~Rs-TKO cells managed to keep normal ER Ca^2+^ homeostasis, they had impaired ER-Ca^2+^ dynamics and diminished SOCE. Our results showed that the expression of IP~3~R3 correlated with the rate of ER Ca^2+^ leakage and refilling, and that IP~3~R3 affected SOCE by regulating NEDD4L (neural precursor cell expressed developmentally downregulated gene 4-like)-mediated ubiquitination of Orai1 protein. Overall, our findings suggest that IP~3~R3 maybe one key player in coordinating ER-mediated Ca^2+^ signalling.
2. Results and Discussion {#sec2-cells-09-00275}
=========================
2.1. With IP~3~R2 Being the Dominant Isoform, IP~3~Rs Regulate Growth and Migration of HEK Cells {#sec2dot1-cells-09-00275}
------------------------------------------------------------------------------------------------
To examine the role of IP~3~Rs in Ca^2+^ signalling, we made two different IP~3~R1-2-3 triple knockout HEK cell lines (IP~3~Rs-TKO) with CRISPR/Cas9 genomic editing technology using procedures similar to previous reports \[[@B17-cells-09-00275]\]. These IP~3~Rs-TKO cells were generated from two separate HEK cell lines stably expressing genetically encoded Ca^2+^ indicators (GECI): GCaMP6m \[[@B29-cells-09-00275]\], a cytosolic Ca^2+^ indicator; or R-CEPIA1er, an ER Ca^2+^ indicator \[[@B28-cells-09-00275]\]. Thus, they were named as GIPK (GCaMP6m cells with IP~3~Rs-TKO) or RIPK (R-CEPIA1er cells with IP~3~Rs-TKO). After confirming the effectiveness of knock-out with sequencing ([Table S1](#app1-cells-09-00275){ref-type="app"}), we tested their responses to 100 μM carbachol (CCh), an agonist for muscarinic acetylcholine receptor that could lead to IP~3~ release. In R-CEPIA1er cells transiently expressing GCaMP6m, CCh could induce Ca^2+^ release from ER, as indicated by a decrease in R-CEPIA1er signal (ER Ca^2+^ levels, red trace) and a transient increase in G-CaMP6m signal (cytosolic Ca^2+^ levels, green trace) ([Figure 1](#cells-09-00275-f001){ref-type="fig"}A, left panel). However, these two events were completely abolished in RIPK cells ([Figure 1](#cells-09-00275-f001){ref-type="fig"}A, right panel). Similarly, CCh also failed to increase cytosolic Ca^2+^ levels in GIPK cells, as indicated by no CCh-induced increase in GCaMP fluorescence ([Figure 1](#cells-09-00275-f001){ref-type="fig"}B). Together, these results reveal that all three IP~3~Rs were functionally knocked-out in both RIPK and GIPK cells.
To characterize the relative contribution of each IP~3~R isoform in IP~3~-induced Ca^2+^ release in HEK cells, we also generated double knockout cell lines (IP~3~Rs-DKO) using GCaMP6m cells ([Table S2](#app1-cells-09-00275){ref-type="app"}). The results showed that, similar to an earlier report \[[@B30-cells-09-00275]\], cells with endogenous IP~3~R2 could mediate CCh-induced Ca^2+^ release with amplitudes similar to those of control cells, while those bearing endogenous IP~3~R1 or IP~3~R3 had much smaller capability in releasing Ca^2+^ (71% and 15% relative to control, correspondingly) ([Figure 1](#cells-09-00275-f001){ref-type="fig"}C). To estimate Ca^2+^ release more accurately, 1-mM GdCl~3~ were used to block all Ca^2+^ transports across PM \[[@B31-cells-09-00275]\]. We then checked whether overexpressed of IP~3~Rs, especially IP~3~R1, could fully rescue CCh-induced Ca^2+^ releases in TKO cells. We used P2A system to co-express IP~3~Rs with different fluorescence proteins. These IP~3~R constructs expressed well as indicated by the presence of corresponding fluorescence. GIPK or RIPK cells transiently overexpressing any IP~3~R isoform could fully rescue the CCh-induced Ca^2+^ release ([Figure 1](#cells-09-00275-f001){ref-type="fig"}D and [Figure S1A,B](#app1-cells-09-00275){ref-type="app"}). As indicated by increases in GEM-GECO signals, RIPK cells with overexpressed IP~3~R1 had even larger CCh-induced Ca^2+^ releases than WT cells ([Figure 1](#cells-09-00275-f001){ref-type="fig"}D and [Figure S1B](#app1-cells-09-00275){ref-type="app"}), likely caused by higher protein levels of overexpressed-IP~3~R1. Together with one previous report showed that all three IP~3~Rs could generate Ca^2+^ puffs, the elementary Ca^2+^ release event, similarly \[[@B30-cells-09-00275]\], these results indicated that the smaller Ca^2+^ responses in DKO cells only expressing IP~3~R1/3 was caused by lower expression levels of corresponding IP~3~R isoforms. Indeed, quantitative RT-PCR results showed that mRNA levels of IP~3~R2 and IP~3~R3 is far higher than that of IP~3~R1 in WT cells. Moreover, the mRNA level of IP~3~R1 in IP~3~R2&3 DKO cells was further reduced by \~40% ([Figure 1](#cells-09-00275-f001){ref-type="fig"}E). Overall, these results indicate that HEK cells might predominantly express IP~3~R2, and IP~3~R1 only contribute marginally to CCh-induced Ca^2+^ release in HEK cells.
IP~3~Rs-mediated Ca^2+^ signalling is involved in diverse physiological processes, such as fertilization, metabolism, secretion apoptosis, and cell growth \[[@B32-cells-09-00275]\]. We thus examined the proliferation and migration rate of RIPK cells. Compared to WT cells, the proliferation rate of RIPK cells was significantly reduced (Left panel, [Figure 1](#cells-09-00275-f001){ref-type="fig"}F) and the migration rate was also minimally affected (Right panel [Figure 1](#cells-09-00275-f001){ref-type="fig"}F). Similar results were also obtained with GIPK cells (data not shown). These results thus showed that IP~3~Rs are involved in the regulation of growth and migration of HEK cells.
2.2. Overexpression of IP~3~R1 or IP~3~R3 Could Rescue the Impaired ER Ca^2+^ Leak and Refilling in IP~3~R-TKO Cells {#sec2dot2-cells-09-00275}
--------------------------------------------------------------------------------------------------------------------
To examine whether these changes in cell behavior were induced by alterations of ER Ca^2+^ homeostasis in IP~3~Rs-TKO cells, we first estimated the size of ER Ca^2+^ store in RIPK cells by comparing their total cytosolic Ca^2+^ releases from ER with those of R-CEPIA1er cells transiently expressing GEM-GECO1, a ratiometric cytosolic GECI \[[@B33-cells-09-00275]\]. To avoid artifacts caused by possible altered rate of Ca^2+^ extrusion via PM Ca^2+^ pumps or sodium/Ca^2+^ exchangers, 1 mM GdCl~3~ were included in extracellular solution to block all Ca^2+^ transports across PM \[[@B31-cells-09-00275]\]. The results showed that there is no detectible difference in thapsigargin (TG, 1 μM)-induced Ca^2+^ releases between WT and RIPK cells ([Figure 2](#cells-09-00275-f002){ref-type="fig"}A). Fura-2 measurements also failed to detect differences in TG-induced responses between these two types of cells ([Figure S2C](#app1-cells-09-00275){ref-type="app"}), indicating no changes in ER Ca^2+^ contents in RIPK cells. Consistent with these results, when directly measured with a ratiometric ER Ca^2+^ indicator GEM-CEPIA1er \[[@B28-cells-09-00275]\], RIPK cells also had similar GEM-CEPIA1er signals to those from R-CEPIA1er cells ([Figure S2D](#app1-cells-09-00275){ref-type="app"}). All these results showed that RIPK cells successfully maintain a normal ER Ca^2+^ homeostasis. Therefore, even though IP~3~Rs are major Ca^2+^ release channels on ER and were shown required for maintaining cellular energy homeostasis \[[@B34-cells-09-00275]\], they are not essential for maintaining ER Ca^2+^ homeostasis.
We then checked whether ER Ca^2+^ dynamics were affected by IP~3~Rs-TKO. We first examined the Ca^2+^ leak rate after blocking SERCA activity with 1 μM TG in cells expressing GEM-CEPIA1er. RIPK cells showed significantly slower decrease in GEM-CEPIA1er ratio ([Figure 2](#cells-09-00275-f002){ref-type="fig"}B), indicating a slower decay rate of ER Ca^2+^ levels. Similarly, the speed of decrease in R-CEPIA1er signals during passive ER store depletion induced by 10 mM EGTA in digitonin permeabilized RIPK cells were also significantly slower than R-CEPIA1er cells ([Figure 2](#cells-09-00275-f002){ref-type="fig"}C, left panel). Together, these results showed that RIPK cells had a slower ER Ca^2+^ leakage. Since ER Ca^2+^ levels are determined by the relative balance between Ca^2+^ leak by leak channels and refilling by SERCA, un-altered resting ER Ca^2+^ content together with a weaker ER Ca^2+^ leakage would imply a less SERCA pumping activity. We thus directly measured refilling of ER Ca^2+^ store with R-CEPIA1er and a protocol adapted from one previous report \[[@B35-cells-09-00275]\]. In permeabilized, ER Ca^2+^ store depleted RIPK cells, addition of 1 mM ATP and 100 nM Ca^2+^ would enable the refilling of Ca^2+^ into ER by SERCA, as indicated by increases in R-CEPIA1er signals ([Figure 2](#cells-09-00275-f002){ref-type="fig"}C). Compared to those of WT cells, the rate of increases in R-CEPIA1er signals in RIPK cells was slower, and the amplitude of increase is significantly smaller ([Figure 2](#cells-09-00275-f002){ref-type="fig"}D). This result demonstrated that the refilling of ER store is also slower in RIPK cells. Other factors may also affect ER Ca^2+^ dynamics directly or indirectly \[[@B36-cells-09-00275]\]. For example, Ca^2+^ buffering protein like calreticulin was shown to modulate IP~3~Rs-mediated Ca^2+^ releases and inhibit or delay corresponding SOCE \[[@B37-cells-09-00275],[@B38-cells-09-00275],[@B39-cells-09-00275],[@B40-cells-09-00275]\]. Further researches are needed to elucidate the effects of IP~3~Rs on them. Nevertheless, these results showed that ER Ca^2+^ dynamics, both leakage and refilling, is slower in IP~3~Rs-TKO cells.
In order to dissect out which isoform is responsible for the modulation of leakage and refilling of ER Ca^2+^ ions, we transiently transfected different IP~3~R isoforms and further examined the refilling and leak rate in permeabilized RIPK or R-CEPIA1er cells. Briefly, we used the resulting maximal increase in R-CEPIA1er signals after the addition of ATP and Ca^2+^ to demonstrate the efficacy of ER Ca^2+^ refilling via SERCA ([Figure 2](#cells-09-00275-f002){ref-type="fig"}E). After the ER Ca^2+^ levels reached a steady state, we then added TG to block SERCA, and used the rate of decreases R-CEPIA1er signals to indicate the release of Ca^2+^ through leak channels on ER ([Figure 2](#cells-09-00275-f002){ref-type="fig"}E). Compared to blank control, the maximal changes in amplitude of R-CEPIA1er signals was significantly larger in RIPK cells overexpressing IP~3~R3 ([Figure 2](#cells-09-00275-f002){ref-type="fig"}F, left panel), showing that the presence of IP~3~R3 may promote ER Ca^2+^ refilling. Consistent with one previous report showing functional couplings between IP~3~R1 and SERCA \[[@B41-cells-09-00275]\], overexpression of IP~3~R1 also significantly enhanced ER Ca^2+^ refilling ([Figure S1E](#app1-cells-09-00275){ref-type="app"}, left and middle panels). While overexpression of IP~3~R2 had no such effect ([Figure S1F](#app1-cells-09-00275){ref-type="app"}, left and middle panels). Together, these results indicate that, besides IP~3~R1, IP~3~R3 may also functionally couple with SERCA to regulate ER Ca^2+^ dynamics. Normally, there is a balance between SERCA activity and ER Ca^2+^ leakage \[[@B24-cells-09-00275]\], changes in SERCA behavior often lead to corresponding alterations in Ca^2+^ leakage. Indeed, IP~3~R1 and IP~3~R3 could significantly increase the decay rate of TG-induced decreases in R-CEPIA1er signals ([Figure 2](#cells-09-00275-f002){ref-type="fig"}F, right panel; [Figure S2E](#app1-cells-09-00275){ref-type="app"}, right panel), while IP~3~R2 had no such effect ([Supplementary Figure S1F](#app1-cells-09-00275){ref-type="app"}, right panel). Thus, IP~3~R1 and IP~3~R3 might also functionally couple with ER Ca^2+^ leak channels. Together, these results imply that IP~3~R1 and IP~3~R3 might be crucial players in ER Ca^2+^ signalling machinery, balancing the function of SERCA and Ca^2+^ leak channels on ER. We do not know how the expression of IP~3~Rs would affect the function of SERCA and Ca^2+^ leak channels. IP~3~Rs may exert their effects via direct physical interactions, or through IP~3~Rs-dependent transcriptional changes, or by some post-translational modifications. Further mechanistic studies are needed to get a better understanding on this.
We also examined the effect of IP~3~Rs-TKO on SOCE-mediated refilling of ER Ca^2+^ stores in intact cells. Cells were first bathed in Ca^2+^ free imaging solution containing 300 μM EGTA and 2.5 μM ionomycin for 8 min to get their ER Ca^2+^ stores depleted. Then ionomycin was washed out and the cells were kept in 0Ca^2+^ solution for 10 minutes. Ca^2+^ was then added into extracellular solutions to allow SOCE and corresponding refilling of Ca^2+^ store. Consistent with previous results showing that RIPK cells have normal basal ER Ca^2+^ levels ([Figure 2](#cells-09-00275-f002){ref-type="fig"}B and [Figure S1D](#app1-cells-09-00275){ref-type="app"}), SOCE responses of RIPK cells were still large enough to generate normal refilling when supplied with regular concentration of external Ca^2+^ (1 mM) (data not shown). However, when low external Ca^2+^ (300 μM) were applied, the increase rate of R-CEPIA1er signals was dramatically slower in RIPK cells than those of R-CEIPIA1er cells ([Figure 2](#cells-09-00275-f002){ref-type="fig"}G). Compared with results from permeabilized cells, SOCE-mediated ER Ca^2+^ refilling in intact RIPK cells were further slowed down (Right panel of [Figure 2](#cells-09-00275-f002){ref-type="fig"}G vs [Figure 2](#cells-09-00275-f002){ref-type="fig"}D), indicating a smaller SOCE in RIPK cells.
2.3. Expression of IP~3~R3 Restored the Reduced Orai1 Protein Level and SOCE in IP~3~Rs-TKO Cells {#sec2dot3-cells-09-00275}
-------------------------------------------------------------------------------------------------
We thus examined the effect of IP~3~Rs-KO on SOCE in HEK cells. Compared to corresponding WT cells, TG-induced increases in Fura-2 ratio or GCaMP signals were significantly reduced in GIPK and RIPK cells, correspondingly ([Figure 3](#cells-09-00275-f003){ref-type="fig"}A and [Supplementary Figure S2A](#app1-cells-09-00275){ref-type="app"}). These results showed that IP~3~Rs-KO reduced SOCE in HEK cells. This result is consistent with one previous finding obtained with IP~3~R mutants in *Drosophila* neurons \[[@B27-cells-09-00275]\], but different to an earlier finding in chicken B cells \[[@B42-cells-09-00275]\]. Thus, these observed correlation between expression of IP~3~Rs and SOCE amplitudes maybe specific to cell types. Nevertheless, IP~3~Rs-DKO cells expressing only endogenous IP~3~R1 showed \~50% reduction in SOCE compared to WT cells. While SOCE in another two DKO cells with endogenous IP~3~R2 or IP~3~R3 expressed was unaltered, indicating their critical roles in regulating SOCE ([Figure S2B](#app1-cells-09-00275){ref-type="app"}). This result also indicates IP~3~R2 and IP~3~R3 had some redundant roles in SOCE regulation.
We next examined the effect of overexpressing different IP~3~R isoforms on SOCE responses of GIPK or RIPK cells. Consistent with results from DKO cells ([Figure S2B](#app1-cells-09-00275){ref-type="app"}) and an earlier report \[[@B27-cells-09-00275]\], overexpression of IP~3~R1 had no effect on SOCE in RIPK cells ([Figure S2D](#app1-cells-09-00275){ref-type="app"}). Even though GCaMP cells with endogenous IP~3~R2 had normal SOCE, overexpression of IP~3~R2 also failed to rescue SOCE in GIPK cells ([Figure S2E](#app1-cells-09-00275){ref-type="app"}). This discrepancy is likely caused by some long-term adaptation in DKO cells. Nevertheless, only overexpression of IP~3~R3 could significantly increase SOCE in IP~3~Rs-TKO cells ([Figure 3](#cells-09-00275-f003){ref-type="fig"}B). Together, these results showed that the expression of IP~3~R3 also could regulate SOCE in HEK cells. It is known that SOCE-mediating CRAC channels, SERCA and IP~3~Rs are functionally coupled under physiological conditions \[[@B24-cells-09-00275],[@B25-cells-09-00275],[@B26-cells-09-00275],[@B43-cells-09-00275]\]. Our results thus added a new aspect of couplings among Ca^2+^ handling machineries.
We then set out to dissect how IP~3~Rs-TKO would affect SOCE. Inspired by a report showing that overexpression of STIM and Orai restored SOCE in *Drosophila* neurons with IP~3~R mutation \[[@B44-cells-09-00275]\], we compared the protein levels of STIM1 and Orai1in IP~3~Rs-TKO cells with those of WT cells. Western blotting results showed that the expression of STIM1 was similar, while the protein level of Orai1 was significantly lower in both IP~3~Rs-TKO cell lines ([Figure S2C](#app1-cells-09-00275){ref-type="app"}). Further results showed that overexpression of IP~3~R3 could significantly increase Orai1 protein levels in GIPK cells ([Figure 3](#cells-09-00275-f003){ref-type="fig"}C). These findings demonstrated that diminished SOCE in IP~3~Rs-TKO cells are caused by reduced expression of Orai1 protein, and that overexpression IP~3~R3 could largely restore both Orai1 levels and SOCE responses.
2.4. Overexpression of IP~3~R3 Restored SOCE by Inhibiting NEDD4L in GIPK Cells {#sec2dot4-cells-09-00275}
-------------------------------------------------------------------------------
To explore the underlying mechanism of reduced Orai1 protein level in IP~3~Rs-TKO cells, we first performed quantitative RT-PCR measurements. Results showed that mRNA levels of STIM1 and Orai1 in IP~3~Rs-TKO cells were similar to those in control cells ([Figure 3](#cells-09-00275-f003){ref-type="fig"}D), indicating no changes at transcription level. We then examined whether Orai1 expression could be altered differently after blockage of protein synthesis with 50 μg/mL cycloheximide (CHX). Results showed that CHX did not restore the decrease of Orai1 protein levels in IP~3~Rs-TKO cells ([Figure 3](#cells-09-00275-f003){ref-type="fig"}E and [Figure S3A](#app1-cells-09-00275){ref-type="app"}), indicating no significant changes in the synthesis rate of Orai1. These results thus suggested that post-translation mechanisms might get involved in down-regulation of Orai1 protein levels.
Previous reports showed that Orai1 protein levels could be regulated by ubiquitination \[[@B45-cells-09-00275]\], a common post-translational modification that induces proteasomal degradation of proteins \[[@B46-cells-09-00275]\]. We therefore examined the effects of proteasome inhibitor MG132 (10 μM) on Orai1 protein levels. Unlike those in WT cells, 12-h treatments of MG132 significantly increased Orai1 protein levels in both types of IP~3~Rs-TKO cells ([Figure 4](#cells-09-00275-f004){ref-type="fig"}A). These results suggested that IP~3~Rs-TKO could promote the degradation of Orai1 protein through the proteasomal pathway.
Lang's works showed that Orai1 could be degraded via an ubiquitin protein ligase, NEDD4L \[[@B45-cells-09-00275],[@B47-cells-09-00275]\]. We thus examined the effects of modulating NEDD4L activity on SOCE responses of GIPK cells. Compared to blank controls, GIPK cells expressing NL.1, a potent and specific intracellular activator of NEDD4L \[[@B48-cells-09-00275]\], had reduced SOCE. In contrast, the expression of a NEDD4L catalytic inhibitor NL.3 \[[@B48-cells-09-00275]\] had enhanced SOCE ([Figure 4](#cells-09-00275-f004){ref-type="fig"}B,C). We then examined the effects of overexpression of NL.3 and NL.1 on Orai1 protein levels in GIPK cells. Western-blotting results showed that inhibition of NEDD4L with NL.3 significantly increased the protein levels of Orai1, while activation of NEDD4L with NL.1 further decreased Orai1 protein levels ([Figure 4](#cells-09-00275-f004){ref-type="fig"}D). These results thus indicated that Orai1 levels and SOCE responses are modulated by NEDD4L-mediated ubiquitination. These results showed that the activity of NEDD4L E3 ligase could regulate Orai1 protein levels in GIPK cells.
To further understand the role of IP~3~R3 in the regulation of NEDD4L on Orai1 proteins and SOCE responses, we examined the effects of IP~3~R3 on mRNA levels of NEDD4L in GIPK cells using quantitative Real-time PCR. Compared to those in WT cells, the mRNA levels of NEDD4L were almost doubled in GIPK cells, suggesting possible presence of more NEDD4L protein in IP~3~Rs-TKO cells ([Figure 4](#cells-09-00275-f004){ref-type="fig"}E), indicating more degradation of Orai1 via enhanced NEDD4L E3 ligase activity in GIPK cells. Moreover, overexpression of IP~3~R3 could partially decrease NEDD4L mRNA levels in GIPK cells ([Figure 4](#cells-09-00275-f004){ref-type="fig"}E). This result indicated that IP~3~R3 could inhibit the activity of NEDD4L, restoring Orai1 protein levels and SOCE responses in GIPK cells ([Figure 3](#cells-09-00275-f003){ref-type="fig"}B,C). The mechanistic underpinning of the linked expression between IP~3~R3 and NEDD4L is yet to be established. It is likely that some IP~3~R-dependent transcriptional factors might get involved and further investigations are needed to elucidate this.
3. Methods {#sec3-cells-09-00275}
==========
3.1. Plasmids Construction {#sec3dot1-cells-09-00275}
--------------------------
CDS sequences of mTurquoise and BFP were synthesized by Qinglan Biotech and Syngen tech, correspondingly. mNeonGreen plasmid and a plasmid containing mScarlet sequence were gifts from Dr. Chen Liangyi, Peking University. pBactSTneoB-mIP~3~R1/2/3 plasmids were gifts from Dr. Mikoshiba. To generate mScarlet-P2A-NL.1/NL.3, the sequence of mScarlet and NL.1/3 were amplified by PCR from WZ\#325&649, WZ\#329&653 \[[@B48-cells-09-00275]\], and mScarlet, correspondingly. The sequence of P2A was contained in the primer to connect to the C terminal of mScarlet. They were then inserted into pCDNA3.1(+) between the KpnI and EcoRV sites with the help of multiple fragment homologous recombination kit (Cat\#: C113, Vazyme biotech, Nanjing, China). To generate mNeonGreen-P2A-mIP~3~R1, mTurquoise-P2A-mIP~3~R2 or BFP-P2A-mIP~3~R3, the corresponding sequences of fluorescent proteins and mIP~3~Rs were PCR-amplified, and then ligated into pcDNA3.1(+) with primers containing P2A sequence between NheI and NotI sites (Qinglan Biotech).
3.2. Cell Culture and Transfection {#sec3dot2-cells-09-00275}
----------------------------------
All the cells were grown in Dulbecco's modified Eagle's medium (DMEM, HyClone) containing 10% fetal bovine serum (Cleson Scientific), 1% penicillin and streptomycin (Thermo Scientific) and cells were maintained at 37 ℃ in a humidified atmosphere containing 5% CO~2~. Gene transfection was done with electroporation (Bio-Rad Gene Pulser Xcell system) using 4 mm cuvettes and OPTI-MEM medium. The electroporation protocol for HEK293 was a voltage step pulse (180V, 25 ms) \[[@B49-cells-09-00275]\].
3.3. Construction of Knockout Cell Lines {#sec3dot3-cells-09-00275}
----------------------------------------
CRISPR/Cas9 technology was used to generate knockout cell lines. Where CRISPR stands for "Clustered regularly interspaced short palindromic repeats", while Cas9 is the abbreviation for "CRISPR-associated protein 9". All knockout cell lines containing ITPR1-ITPR2-ITPR3 triple KO (*itpr1/itpr2/itpr3*), ITPR2-ITPR3 double KO (*itpr2/itpr3*), ITPR1-ITPR3 double KO (*itpr1/itpr3*), ITPR1-ITPR2 double KO (*itpr1/itpr2*) were generated by CPISPR/Cas9 gene editing technology \[[@B17-cells-09-00275],[@B50-cells-09-00275]\] from previously made GCaMP6m and R-CEPIA1er stable cells. sgRNA targeting sequences to these three genes were designed by SyngenTech company and inserted into lentiCRISPR v2 vector (Addgene plasmid no. 52961) by BsmBI site. The sequences of sgRNA set for three isoforms are shown in [Table 1](#cells-09-00275-t001){ref-type="table"}.
Protocols for the generation of KO cells were similar to those previously described \[[@B17-cells-09-00275]\]. Briefly, three sgRNA plasmids targeted ITPR1, ITPR2 and ITPR3 were transfected into cells by electroporation. Two days later, the transfected cells were selected with 2 μg/mL puromycin for 4 days. The survived multiple clonal cells were then seeded into 96 well dishes with a density of one cell/well. 10 days later, the resulting single clones were first functionally screened by calcium imaging and then the positive healthy clones were confirmed further with sequencing.
3.4. Single-Cell Intracellular Ca^2+^ Measurements {#sec3dot4-cells-09-00275}
--------------------------------------------------
All Ca^2+^-imaging assay was performed in the HEK293 cells as described \[[@B17-cells-09-00275]\]. Briefly, intracellular Ca^2+^ imaging was performed with a ZEISS oberserver-A1 microscope equipped with a 40× oil objective (NA = 1.30), Lambda DG4 light source, and the SlideBook6.0 software (Intelligent Imaging Innovations, Inc.). GFP-1828A-000 filter and FURA2-C-000 filter were used to measure cytosolic Ca^2+^ signals indicated by GCaMP6m \[[@B29-cells-09-00275]\] and Fura-2 respectively, and Endoplasmic Reticulum (ER) Ca^2+^ levels indicated by R-CEPlA1er \[[@B28-cells-09-00275]\]. Fluorescence were obtained with a TxRed-A-Basic-000 filter set. The fluorescence signals from different Ca^2+^ indicators were collected every 2 seconds. The Ca^2+^ imaging solution contains 7.2 mM KCl, 107 mM NaCl, 1.2 mM MgCl~2~, 11.5 mM glucose, 20 mM HEPES-NaOH (pH 7.2).
Protocols for Fura-2 loading were similar to those described before \[[@B51-cells-09-00275]\]. Cells were first incubated with imaging solution containing 2 μM Fura-2 AM and 1 mM CaCl~2~ for 40 min at room temperature in the dark and subsequently eliminated Fura-2 AM for another 30 min. Data were acquired from collection of emission fluorescence at 509 nm generated by 340 nm excitation light (F340) and 380 nm light (F380) at 2 s intervals. The intracellular cytosolic Ca^2+^ levels were indicated by F~340~/F~380~ ratio.
3.5. Real-Time PCR {#sec3dot5-cells-09-00275}
------------------
Total RNA was extracted from cells using TRIzol and the reverse transcription was performed with TransScript® First-Strand cDNA Synthesis SuperMix (Code\#: AT301-03, TracsGen Biotech) following the manufacturer's instructions later. Briefly, total RNA aliquots (3μg) were reverse transcribed at 65 ℃ for 5 min, ice-bath for 2 min, 42 ℃ for 30 min and 85 ℃ for 5 s. Reaction mixture aliquots (cDNA, 1 μL) were used as templates for two-step quantitative RT-PCR using TransStart® Top Green SuperMix (Code\#: AQ131-03, TransGen Biotech). PCR running condition was 94 ℃ for 30 s, followed by 40 cycles of 94 °C for 5 s and 60 °C for 30 s. Then, the melting curve program was run. Relative differences in numerous mRNA expression levels was calculated by comparative cycle threshold method \[[@B52-cells-09-00275]\]. The primer targets for genes are listed in [Table 2](#cells-09-00275-t002){ref-type="table"}.
3.6. Cell Proliferation and Migration Essay {#sec3dot6-cells-09-00275}
-------------------------------------------
Cell proliferation essay were performed by IncuCyte^®^ Live Cell Analysis System (ESSEN BioSCIENCE), following the standard protocols provided by the company. For migration assay, cells were cultured into 96-well with a density of 30,000 cell/well, and EssenR 96-well WoundMaker^TM^ was used to make scratch wound after 5 h, and then the dish was put into IncuCyte that recorded the migration every 2 h for three days. The data were analyzed with IncuCyte^®^ ZOOM Software.
3.7. Western Blotting {#sec3dot7-cells-09-00275}
---------------------
Total proteins were extracted by Total Protein Extraction Kit (BB18011; BestBio) following the manufacturer's instructions and concentration was measured by BCA protein assay kit (E162-01; GenStar). 6× loading buffer was added to protein extracts and boiled for 5 min at 60 ℃. 20 μg protein was separated on 8% SDS-PAGE and transferred onto PVDF membranes (Millipore). The resulting membranes were blocked for 1 h at 37 ℃ with TBST buffer (12 mM Tris-HCl, pH 7.5, 137 mM NaCl, 2.68 mM KCl, 0.1% Tween 20) containing 5% nonfat dried milk, then incubated with primary antibody overnight at 4℃. After washing three times (10 min each) in TBST buffer, the membranes were loaded with secondary antibody for 40 min. Detection was performed by ECL (GS009-4; Millipore) solution and imaged using Tanon5200 detection system finally. STIM1 and Orai1 was detected with anti-STIM1 antibody (5668S; CST) (1:1000 dilution) and anti-Orai1 antibody (o8264; Sigma) (1:300 dilution) followed by anti-rabbit-IgG (7074S; CST) (1:4000 dilution) respectively. Internal control GAPDH was detected with anti-GAPDH antibody (sc-32233; SANTA) (1:2000 dilution) and corresponding secondary antibody is anti-mouse-IgG (7076S; CST) (1:2000 dilution). The representative data shown were carried out from three independent experiments. The intensity of the images was quantified by Image J software, and the resulting data were plotted with prism7 software.
3.8. Proteasomal Inhibition Assays {#sec3dot8-cells-09-00275}
----------------------------------
Cells were seeded at 12-well plates, and were incubated for indicated hours respectively in DMEM high glucose with 10% FBS containing 50 μg/mL CHX or 10 μM MG132. We defined 0 h as 30 min-incubation. Incubated cells were washed in HBSS for second times and total proteins were extracted and immunoblotted as described above.
4. Conclusions {#sec4-cells-09-00275}
==============
Overall, our results demonstrated that the expression level of IP~3~R3 correlated with SERCA pump activity and ER Ca^2+^ leakage rate, as well as SOCE amplitudes in HEK cells. These results indicate that IP~3~R3 is a crucial coordinator of ER-mediated Ca^2+^ signalling. Further studies are needed to understand the exact mechanisms underlying this coupling among different ER-Ca^2+^ signalling machineries in HEK cells.
The following are available online at <https://www.mdpi.com/2073-4409/9/2/275/s1>, Figure S1: Characterizing ER-mediated Ca2+ responses in IP3Rs-TKO cells, Figure S2: Characterization of IP3Rs-TKO or IP3Rs-DKO cells, Table S1: Sequencing results of IP3Rs-TKO cells, Table S2: Sequencing results of IP3Rs-DKO cells.
######
Click here for additional data file.
S.H. and Y.W. conceived the project, designed the experiments. Y.W. and L.Y. wrote the manuscript. L.Y. generated all KO cells, performed Western blotting, some RT-PCR and calcium imaging assays. L.W. did most calcium imaging experiments and some RT-PCR measurements. Y.D. did NL.1/3 related experiments. W.Z., K.H., Y.M. and K.M. provided NL or IP3R plasmids, helped improve the manuscript. X.J. provided tech support for Western blotting analysis. Y.Z. and D.L.G. provided valuable ideal inputs and helped write the manuscript. All authors reviewed the results and approved the final version of the manuscript.
This research was funded by the Ministry of Science and Technology of China, grant number 2019YFA0802104, and the National Natural Science Foundation of China, grant number 31970723, 91954205 and 31671492, and the Fundamental Research Funds for the Central Universities, grant number 2017EYT21.
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.
{ref-type="app"} for typical curves). (**E**) Relative mRNA levels of three types of IP~3~Rs in HEK cells and IP~3~Rs-DKO. mRNA levels were first normalized against corresponding GAPDH levels, then normalized against corresponding IP~3~R1 levels of WT cells. Expression levels of IP~3~R1 in WT cells were set as 1 (mean ± SEM, \*\*\* *P* \< 0.0001, Student's *t*-test). (**F**) Typical proliferation (left panel) and migration (right panel) curves of RIPK and WT cells.](cells-09-00275-g001){#cells-09-00275-f001}
{#cells-09-00275-f002}
{#cells-09-00275-f003}
{#cells-09-00275-f004}
cells-09-00275-t001_Table 1
######
Sequences of sgRNAs and sequencing primers used for knocking out IP~3~Rs in HEK293 cells.
------------------------------------------------------------------------------
Target sgRNA Sequence (5′→3′) Target CDs Site Primers (5′→3′)
-------- ------------------------ ----------------- --------------------------
ITPR1 CAAAGACGACATATTAAAGG 687--706 F: TGCTGTGATTTTAGTGGCGT\
R: TCTCCACCCTACCCTTACCT
ITPR2 ACACGATGTCCCCTATGTAG 46--27 F: TCTTGGCCGCTGTAGTCC\
R: CCAGGGAAAACAAGCACACA
ITPR3 ACCCCCCTTCTCACGGAACG 1593--1574 F: GGGTTGTGGTCCAGCTTAGA\
R: CATCTAACCCAGTGCAAGGC
------------------------------------------------------------------------------
cells-09-00275-t002_Table 2
######
Primer sequences for quantitative real-time PCR.
Gene Forward Primer Reverse Primer NCBI Reference Sequence
-------- ---------------------- ------------------------ -------------------------
ITPR1 ATGACAGCTCTGAGGAGAA GCAGAAGAGACAGGAGATTTAG NM_001099952.2
ITPR2 CTGTGTGGAAGAGCAACTAC ATGGGTCTGTAGGTAGGAATAG NM_002223.3
ITPR3 CAAGCCCTCCAAAGATGAG CGAAGGTGTCGATGATTACC NM_002224.3
STIM1 GGATCTCAGAGGTTGGTAGA GATGGAAGAGGAGCAAGAAG NM_001277961.1
Orai1 GACGCTGACCACGACTA CCTTGACCGAGTTGAGATTG NM_032790.3
GAPDH AACTGCTTAGCACCCCTGGC ATGACCTTGCCCACAGCCTT NM_001256799.2
NEDD4L CCTAGATTGTCACGGTTTCC GACTAAACTCTCACCTCCTTTC NM_001144964.1
[^1]: These authors contributed equally to this work.
|
{
"pile_set_name": "PubMed Central"
}
|
Cellular immune responses of cattle to Cowdria ruminantium.
Peripheral blood mononuclear cells (PBM) from cattle immunised against Cowdria ruminantium infection (Heartwater), proliferated in vitro in the presence of either infected autologous endothelial cells pre-treated with T cell growth factors to induce MHC class II expression, or infected autologous monocytes. Proliferation was not observed in PBM cultured with a soluble extract of the agent, but PBM responded to two recombinant antigens of C. ruminantium, namely a 32 kDa (MAP1) and a 21 kDa antigen (MAP2). We hypothesize that infected endothelial cells and monocytes present Cowdria antigens to specific lymphocytes during infection and thereby play a role in the pathogenesis/immune response to the pathogen.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
.welcome-view {
&, body {
height: 100%;
}
body {
color: #666;
font: normal 16px/1.4 Helvetica,Arial,Sans-serif;
}
h1 {
color: #333;
font-size: 24px;
margin: 20px 0 10px 0;
padding: 0;
}
p {
margin: 0;
padding: 0;
}
.wrapper {
display: table;
height: 100%;
min-height: 100%;
width: 100%;
}
.wrapper--inner {
display: table-cell;
text-align: center;
vertical-align: middle;
}
.animation {
height: 70px;
position: relative;
}
.animation--stem {
position: absolute;
left: 50%;
top: 35px;
width: 2px;
height: 50px;
margin-left: -1px;
background: #348eda;
&:before {
content: "";
position: absolute;
left: 50%;
top: 0;
border-radius: 100%;
background: #348eda;
width: 6px;
height: 6px;
margin-left: -3px;
margin-top: -3px;
}
}
.animation--icon {
position: absolute;
left: 50%;
top: 50%;
width: 80px;
height: 80px;
margin-left: -40px;
margin-top: -40px;
span {
background: #348eda;
background: rgba(52, 142, 218, 0);
border-radius: 100%;
display: block;
width: 80px;
height: 80px;
transform: scale(.2);
animation-name: wilkomen;
animation-duration: 1.8s;
animation-iteration-count: infinite;
animation-timing-function: linear;
}
+ .animation--icon span {
animation-delay: .9s;
}
}
}
@keyframes wilkomen {
0% { }
25% {
background: rgba(52, 142, 218, 1);
}
50% {
background: transparent;
box-shadow: inset 0 0 2px rgba(52, 142, 218, 1);
}
100% {
transform: scale(1,1);
box-shadow: inset 0 0 2px rgba(52, 142, 218, 0);
}
}
|
{
"pile_set_name": "Github"
}
|
As Long as You're Near Me
As Long as You're Near Me () is a 1953 West German drama film directed by Harald Braun. It was entered into the 1954 Cannes Film Festival.
Cast
Brigitte Horney as Mona Arendt
O.W. Fischer as Frank Tornau
Maria Schell as Eva Berger
Walter Richter as Willi
Liesl Karlstadt as L'habilleuse
Mathias Wieman as Paul
Hardy Krüger as Stefan Berger
Paul Bildt as Bentz
Heini Göbel
Hans Henn
Michael Lenz
Wolfgang Molander
Gudrun Rabente
Rudolf Vogel
References
External links
Category:1953 films
Category:1950s drama films
Category:German drama films
Category:West German films
Category:German-language films
Category:German black-and-white films
Category:Films directed by Harald Braun
Category:Films about filmmaking
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Про це він заявив в інтерв’ю «Голосу України», коментуючи реформу прокуратури.
«Я обстоюватиму прокуратуру, яка би захищала, перш за все, людей. Люди, а потім держава, бо держава це і є люди. У тому числі й приватну власність людини. До речі, на мій погляд, вона має бути чітко виписана в законодавстві, а людина повинна мати можливість захищати її, у тому числі й зі зброєю в руках. На жаль, у нас це не зроблено», - зазначив Шокін.
Читайте такожШокін погрожує звільненням обласним прокурорам, які не розслідують злочини часів Революції гідності
За словами генпрокурора, він вніс до РНБО свої відповідні пропозиції, їх підтримали, в тому числі президент України.
«Ці пропозиції опрацьовуються, сподіваюся, що буде прийнято закон про зброю, реєстрацію, можливість її придбати і весь комплекс питань, що ними опікується МВС», - повідомляє прокурор.
Раніше УНІАН повідомляв, генпрокуратура встановила факти перешкоджання розслідуванню злочинів проти активістів Революції гідності з боку співробітників правоохоронних органів і відкрила кримінальне провадження.
Читайте останні новини України та світу на каналі УНІАН в Telegram
|
{
"pile_set_name": "OpenWebText2"
}
|
Numerous products today are shipped and sold in reclosable plastic bags. Examples of plastic bags having a reclose feature can be found in U.S. Pat. No. 4,617,683 issued to Christoff on Oct. 14, 1986, and U.S. Pat. No. 4,986,673 issued to Bell on Jan. 22, 1991. When using one of the bags disclosed in the above U.S. patents, the consumer first removes the sealed portion of the bag located directly above the reclosing feature. The consumer then pulls apart or separates the male and female securement members of the reclose feature. The consumer can now remove the desired portion of the product from the bag. The consumer then manually engages the male and female securement members of the reclose feature, thereby temporarily securing the bag and its contents until the next use. However, as is often the case with a bag having a reclose feature, there is no biasing means to keep the bag mouth open after the reclose feature has been pulled apart by the consumer. The lack of a biasing means to keep the package in a fully open condition can be cumbersome and annoying to a consumer who desires to have the bag mouth remain open after the reclose feature has been pulled apart, thereby facilitating access to the product within the bag. Improved access to the product is particularly desirable when the bag contains certain products, many of which are not entirely consumed upon initial opening of the pouch, e.g., coffee, salted snacks, fabric softeners, cleaning products, detergents, pharmaceuticals, etc. Improved access is also highly desirable when the pouch is to be used as a vessel for the reconstituting fluid, e.g., concentrated beverages which are to be diluted with water.
Commonly assigned U.S. Pat. No. 4,898,477 issued to Cox et al. on Feb. 6, 1990 and which is hereby incorporated herein by reference discloses a self-expanding flexible pouch that can be used as a measuring device for reconstituting a concentrated product contained therein. The flexible pouch includes an extensible stay located in the throat area of the pouch. When the pouch's top portion is removed, the extensible stay expands to a circular configuration opening the pouch's throat area. The consumer can then "milk" the concentrated product out of the pouch. The opening of the pouch's throat area allows the consumer then to reconstitute the fluid by adding water to the pouch and then pouring and mixing with the concentrated substance previously "milked" from the pouch. The Cox et al. pouch has the advantage of self-expansion, thereby improving access to the product within the pouch. However, this pouch would not be particularly well suited to package a product that required the pouch to be sealed between repeated multiple openings.
When using certain products that require repeated or intermittent use prior to exhaustion of the pouch's contents or when using products which are to be diluted using the empty pouch as a vessel for the reconstituting fluid, the consumer not only desires that the bag be reclosable between uses, but also that the bag mouth remain open. By way of example, liquid fabric softener, granular detergents, and salted snacks are products that the consumer would like to reclose between uses to prevent spilling or preserve freshness, and would also prefer to have a self-opening feature to improve or facilitate product access.
Accordingly, it is an object of the present invention to provide a flexible pouch having a self-opening and reclosing device suitable for use with a wide assortment of products;
It is further an object of the present invention to provide a flexible pouch that will assume a substantially predetermined configuration upon opening;
It is further an object of the present invention to provide a flexible pouch that will collapse when discarded and thereby reduce the volume of generated waste material;
It is further an object of the present invention to provide a flexible pouch that is sufficiently rigid and stable when opened to allow a consumer to firmly grasp and pour a fluid from the pouch with confidence that the pouch will not collapse;
It is further an object of the present invention to provide, in a particularly preferred embodiment, a flexible pouch that can be readily opened without having to use an implement such as a pair of scissors or a knife;
It is still another object of the present invention to provide a flexible pouch that accomplishes the aforementioned objectives at minimal cost.
|
{
"pile_set_name": "USPTO Backgrounds"
}
|
"You know, i've got to say it's great to finally meet you, rick." "Getting all those new mutants in the underground has made you quite the legend." "Well, i've been lucky." "What?" "Getting that shape shifter past the gsa blockade in akron, i think that took a little bit more than luck." "Well, hopefully we can keep these two under their radar." "Yeah." "What's their story?" "They seem okay." "Something's a little off, if you know what i mean." "Huh." "What about their mutancy?" "Hey, todd." "Whoa, superspeed." "Yeah." "That's going to come in handy." "Yeah." "He's a show-off." "Oh, my wife, jerri." "Of course, you've got plenty to show off there yourself, big guy." "She's got x-ray vision." "Oh, hey... , brennan." "Rick?" "Stay back." "Are you okay?" "I can't stop." "Adam, we've got a problem." "That should contain his powers." "Even while his vital signs are weakening." "I thought you had him stabilized." "I've done the best i can." "But i'll be honest with you:" "He's barely holding on." "He's not going to die, is he?" "I don't know." "All i know for sure is he's completely lost control of his new mutant abilities." "Better run a battery of tests to help me figure out what's causing this." "Well, once you've got that figured out you'll be able to treat him, right?" "Well, diagnosis is the first step." "And what if we figure out whatever he has is contagious?" "Well, rick might just be the tip of the iceberg." "A new mutant was brought in this morning." "It's the second case in two days we've discovered." "And that's the operative word here, isn't it, mr marlowe?" "Sir?" ""Discovered"." "There's no telling how many of these aberrant loose cannons are wreaking havoc out there." "Well, this loose cannon turned the entire city aquarium into one ice block." "Killed all the fish inside." "Ichthyology has never been a particular interest of mine." "Well, along with the diver who was feeding them." "What about the other one?" "Her new mutant power of suggestion caused a thirty-two-car pile-up on the trans-island highway." "Any human fatalities there?" "Two, sir." "That's tragic." "But a credible excuse to initiate a clampdown." "What have your interrogations revealed?" "Since being apprehended, they have fallen extremely ill." "High fevers, deteriorating vital signs." "And you still have no idea what's causing this?" "No." "But our medical staff are running batteries of tests and we just need more time." "Well, you don't have more time." "Solving this problem is priority one." "I want all squads on full alert to locate and incarcerate these out-of-control new mutants." "Yes, sir." "Brennan checked in and asked me for a progress report." "Alright, anyway, tell him that he and the wileys should avoid using their powers." "Using them accelerates the spread of the infection." "I think he's more concerned about how long he's going to have to stay cooped up with them." "Well, you tell them that he's not to leave the safehouse under any circumstances." "How come?" "Because any new mutant who's been exposed to this thing needs to remain quarantined." "But i thought you said there was a good chance rick didn't enter the infectious stage until after he left the safehouse." "Yeah, but until we're certain of that isolation is the only way to keep this thing from spreading." "Got any idea what we're dealing with?" "Yeah." "Off." "Cladosporium." "Here." "There it is." "It's simple, common, airborne spore." "Completely harmless to most human beings, but in the case of a rare genetic aberration, which occurs in less than 1% of all new mutants, exposure to the spore causes uncontrolled mutation and eventually death." "Less than 1%?" "Well, not to sound heartless, but i mean if we're only talking less than 1% than what's the big deal?" "It's what happens after the exposure that's the big deal." "Once the person shows symptoms they become highly infectious and the disease is passed on." "New mutant to new mutant." "Sorry, guys, but we've got a little setback here." "Setback?" "What kind of setback?" "Well, we're all under quarantine." "Quarantine?" "Huh, for how long?" "Until they figure out what infected rick." "Wait a minute, you mean we've been exposed to something?" "We could get sick?" "Don't worry." "I'm sure we'll all be fine." "You know, it's just a precaution." "It's only going to be delayed a little while, right?" "I think i've got it." "It starts in the throat, right?" "Come on, jer, you never get sick." "So what are we looking at here, an hour or two?" "Huh?" "I can't answer that." "And until we do figure out what we're dealing with, adam thinks we should stop using our mutant abilities." "It could trigger an episode." "Fine." "Todd?" "Please." "Right." "Right." "It's in my throat." "No doubt about it." "It's going to be a long day." "So what's this fire we're investigating?" "Alice robbins." "She has the power to throw off thermal energy." "I remember." "We slipped her into the underground last year." "Sounds like she's becoming the latest victim of this infection." "Her powers are spinning completely out of control." "Yeah, i would have thought in a situation like this the they would have been, but they got called off by higher authority." "Jesse told me." "Better activate your hazmat suit." "Good." "How come you're not wearing one?" "Well, not being a new mutant i'm immune." "So i'm the only one that gets to make a hazmat fashion statement?" "But you just wait, it'll be the rage of the spring collection." "I'm sure." "It's locked." "It's hot." "Alice?" "Alice?" "You think you can take this down?" "I guess so." "Well, if you're afraid of catching something, don't be - the suit will protect you." "No, it's not that." "It's... you know how i am around fire." "Shalimar, the fear of fire is perfectly naturally for a feral." "You and i both know that there's only one way for you to get over it." "Well, you better stand over there, then." "God, she... she's burning up." "I mean, literally." "Adam." "Is there anything we can do?" "Too late." "Just go." "How long until her next flare?" "Not long." "I'm afraid the next one might kill her." "So, it looks like the news is out." "Well now i know who called off the authorities." "Back off or he's dead." "Like i give a damn." "Well working for eckhart has brought out your sensitive side, huh marlowe?" "See, now that's not fair." "I was bitter long before i took this job." "Yeah, well we'll deal with your psyche some other time." "Miss robbins here is about to reach her ignition point." "There's not a lot we can do." "I suppose that would mean a swift death for her." "Yeah, and for anyone else in the vicinity." "Don't stay on my account." "Alice, what can i do?" "Just go, please." "Just go." "We've got to get out of here." "Marlowe." "Tell me." "Look out!" "Well, i gather you've met my new lieutenant." "This infection, it just killed this woman." "It could wipe out every new mutant in a week." "Now, i might be able to beat it, but i don't stand a chance if i have to fight you at the same time." "What are you suggesting?" "A truce to accomplish a common goal." "What?" "Do you honestly believe i'd shed a single tear if your freakish monstrosities were wiped out?" "If the new mutants disappear, so does the gsa, mason." "Overnight you're obsolete." "Obsolescence in the wake of total victory is a mixed blessing, but a blessing nonetheless." "Yes, but what about you?" "You lose your bete noir, your reason for living." "I'm deeply touched by your concern for my welfare, adam." "For a man who's life is completely defined by that which he hates the most, this could be your undoing." "Give me a moment." "So who's the new lieutenant?" "Charles marlowe." "I lost track of him since he left genomex." "I think it's a huge mistake to trust adam with the gsa technologies and secrets." "While gaining access to his genius and expertise." "Bottom line is our priorities and his remain fundamentally opposed." "Well, under normal circumstances you'd be absolutely correct." "But right now we find ourselves in the rare position of having our agendas intersect." "At least until he discovers a remedy." "He really hates you, you know?" "Eckhart's always hated me." "I'm talking about marlowe." "Hm." "Tell me what you need." "You made a deal with the devil." "I can't deny that, emma." "Look, what have they got that we don't have here?" "Highly trained scientists and medical staff that works around the clock with unlimited funds." "If rick stays here, there's no reason to risk moving him." "But eckhart is going to try... i know, i can handle eckhart, all right?" "The real problem here is the speed of this infection." "Now, we have approximately twenty hours before the majority of new mutants are cut down." "Hey, buddy." "I've been thinking, we've been here a few hours, right?" "If we were going to get sick, chances are it would have happened by now." "Follow me?" "Yeah, i think i follow you." "So in that case, i figure we may as well get the hell out of here." "No." "Aw, come on." "Hey, look, until i get the world we stay put." "And that's that." "Hey, speedy, was i talking to myself or did you not hear me say no using our abilities, huh?" "What are you, fast and deaf?" "So you think we'll be okay?" "Yeah, we'll be fine." "What?" "Oh, hey, come on you're not using your x-ray vision, are you?" "No, don't be silly." "Yeah." "It looks pretty good from this angle too." "Please." "God, brennan's got to be going nuts trapped in the safehouse with that couple." "Yeah, not knowing." "Yeah." "Not knowing." "That's the part that would drive me crazy." "You don't think he's got it, do you?" "I wouldn't know, jesse." "Well, can't you do one of your tele-empathic tune-ins and see if he's infected?" "It doesn't work like that." "Hm." "I've got to go check in on rick." "You've got to stay in there, rick." "Just until we find a way to make you better." "Adam, i'm your liaison." "I don't recall having a liaison looking over my shoulder as part of the deal." "Not my call." "Alright, so what is your mandate, mr marlowe?" "Report your progress on mr eckhart." "Just the thought of working with him again." "Stop kidding yourself, adam, you two are joined at the hip." "That's an ugly thought." "You need eckhart as much as eckhart needs you." "Oh, i thought i left all that behind." "Well, the past has this funny way of sneaking up on you, doesn't it?" "Yes, and rearing it's ugly head." "It's a tough thing to escape." "So, what, you're his eyes and ears around here, mr marlowe?" "Yes." "As well as making sure that gsa technology is not used for detrimental purposes." "Detrimental to what, the imprisonment and the murder of new mutants?" "Hey, i just work here." "Yeah." "And that doesn't give you pangs of guilt?" "No." "It's a free marketplace, adam." "My services are available to the highest bidder." "Which, by the way, if you want to make a bid that beats out genomex." "What, so... so you're a mercenary now?" "Yes, i am." "And you better get used to me being on this side of the fence." "Besides, my retinal scan is the only way to get access to that computer." "Limited access, i'm sure." "Of course." "Of course." "How are you and sunshine getting along?" "After he left genomex he must have suffered a major psychological trauma." "Or maybe he's just a creep." "No, people don't have such a change in personality for no reason... his anger's got to have a source." "Think he's having a hard time dealing with his new mutant abilities?" "Well, that's definitely part of it, yeah." "Didn't turn the rest of us into creeps." "Well, none of you had to deal with his invulnerability." "Are you kidding me?" "I would love to have his power of invulnerability." "No, no, listen, think about this now." "All human behaviour and animal behaviour, for that matter, is predicated on one simple motivation - survival, right?" "Yeah." "Alright, so what happens when survival isn't an issue?" "When the motivation is taken away?" "I guess there's no telling how someone would react." "Exactly." "Unpredictability." "Which makes marlowe more dangerous then eckhart." "Tell me." "Tell me!" "Hey!" "Hey!" "What was that all about?" "What was that all about?" "I asked him a question." "I got no answers for you." "Alright, then you're banned from this ward!" "You are a guest here!" "You get out of here!" "This is not your turf." "You remember that, adam." "Come on, rick, just hang in there." "Come on, rick." "Emma?" "Emma, are you okay?" "Why did you do that, jesse?" "You were in trouble." "Well we've both been exposed." "Oh, man, you don't give up, do you?" "What you don't understand is that i need to get out of here, okay?" "And now." "What's your rush, todd?" "We've got no place to go." "I'm not in a rush, jerri." "Okay?" "I just want out." "Excuse me." "Sorry about that, but i don't think you're going to get very far without this." "What the hell is that in there?" "In where?" "There's a picture in your wallet." "I got lots of pictures in my wallet." "Oh, you mean this picture?" "Who the hell is this?" "A friend." "She teaches a spin class at my gym." "Yeah." "I bet that's not all she teaches." "Now, that was really swift." "See, now it doesn't take a genius to know that a married man does not carry a picture of his girlfriend around in his wallet." "Adam?" "I just came through the ward." "Marlowe is at it again." "Alright." "Thanks, take over." "Take another reading, quick." "Come on, you can do it." "What are you, the angel of death?" "No, i'm just asking those on the edge a few questions." "Alright, there's a lot of people on the edge around here." "We're trying to save their lives." "Well, you don't seem to be doing a very good job, do you?" "Yeah, and you're not helping." "I'm not here to help." "Alright, there's no place for voyeurism." "I want you to keep your distance." "Well, we can't all have what we want, can we now?" "You know, i can't believe that you've ended up like this." "A man with your incredible gift." "Let me tell you about the precious gift you gave me, doctor." "Standing in the wreckage of a car crash you caused." "Seeing your wife and child, breathing their last terrified breaths." "And you're just going to walk away." "Knowing that you're never going to see them again." "Not in heaven, not in hell." "And for that i thank you, good doctor." "Thank you." "Thank you." "How are you feeling, jess?" "Never better." "You?" "Fine." "Perfectly normal." "As normal as we ever get, huh?" "I feel bad about rick." "Yeah, me too." "You think he suffered?" "I don't know." "It didn't like the most peaceful way to go." "Hard to think that's what we've all got coming, huh?" "I can't think that way, jesse, and neither should you." "Well you better face reality, emma." "What if adam can't come up with a cure?" "He will." "Besides, we both feel fine, right?" "Right." "So far so good." "Maybe we got lucky, huh?" "Maybe we did." "Alright." "Okay." "Emma." "Get it off." "Emma, are you all right?" "Say something." "Open your eyes." "What happened?" "Come on, we've got to get you to bed." "You've got it too." "Adam." "Jesse." "Emma's sick." "I think i've got it too." "Alright, jess, you be careful." "If you phase for more than thirty seconds you could dissipate for good." "Yeah, tell me something i don't know." "Shalimar, jesse and emma have come down with it." "Oh, no." "Alright, now i need you to get them over here as quickly as possible." "What happened?" "Apparently rick went into arrest, emma tried to revive him and jesse jumped in." "Okay, i'm on my way." "Go." "Now." "Hurry." "This partnership with adam seems to be as big a misfire as the one you forged twenty years ago." "You're giving him everything that he asks for, aren't you?" "Yes." "He still can't get a handle on the cure." "I have to say, mr marlowe, i find your attitude rather singular." "You certainly wouldn't be walking around today if it wasn't for adam's efforts." "I'm supposed to thank him for that?" "Most other new mutants seem to." "Yeah, well most other new mutants aren't blessed with my particular powers." "Powers, i must say, i wouldn't mind possessing were i one of you." "Well, then you should have gotten adam to twist up your dna when you were still on good terms with him." "That's not what genomex was about back then." "Yeah, well that was then this is now." "I can't believe you're putting your trust in him." "You haven't had anything to do with adam in years." "I've had everything to do with adam." "I know the man." "With something so dear to him at stake... if there's a cure to be found, he'll find it." "That sounds like wishful thinking." "It's better than then your perpetual negative thinking, mr marlowe." "Negative, positive, what's the difference?" "In the end, the result's all the same to me." "Just don't let your personal despair impede the successful completion of this mission." "I told you, he's getting everything he wants." "Because whoever controls this disease and its cure controls new mutants." "You're talking about wresting it away from him, aren't you?" "Of course." "And with you heading the opposition, it won't be much of a fight." "No, it won't." "Multiple dvcs." "..., he's crashing." "Three... of o-negative, let's go." "O- negative, let's go, let's go." "We've got a... he's still tachying." "She's not breathing." "Paddles, come on." "Give me a cbc on him." "Eckhart has more confidence in you than i do." "But he's a born optimist." "Scientists worked for years trying to cure the symbolist disease." "How can i be expected to solve this one in two days?" "One of the creators desperately struggles to try and save his precious creations." "That's right." "In a way, i think of you as my children." "A man shouldn't outlive his children, should he?" "I'm sorry for your loss." "I am." "I can't imagine the pain you've been through." "Well, i appreciate your expressions of sympathy, doctor." "But if it wasn't for you and your genetic tampering i'd be where i should be." "Next to my wife and child." "You also never would have seen your twelfth birthday or ever had a wife or a child, okay?" "The work that i did at genomex saved not only your life, but the life of every other new mutant." "Yeah, and you go take a look." "Take a look at the legacy, huh?" "Yeah, but you know there are a lot of other people who should have died who are out there leading productive lives." "They're using their abilities to help people." "I don't need you to judge me, adam." "I'm not judging you." "I'm telling you that there is a different way." "There is a way to give you a little peace in your life." "That is not going to happen." "Not for me." "And that is thanks to your experimenting." "All right, look, it was a new frontier." "It was genetics." "We were the pioneers." "Oh, so i'm a guinea pig now." "I saved your life." "Your parents brought you to me." "And you had no idea what was held in store for me." "No, not at the time." "Eckhart told me that i should kill you the second you found the cure." "But i'm not going to wait that long." "Of course." "What?" "I thought that you possessed invulnerability." "What you really got is a capacity for cellular generation which is off the charts." "Same difference." "No, no, no, not in my world it isn't." "Marlowe, you are the key to the cure." "Emma?" "Shalimar." "Jesse, you're sick." "You shouldn't be phasing in and out like that." "I can't help it." "How long has she been like this?" "Twenty minutes." "Can you walk?" "Yeah." "I have to get you guys to triage." "I need to isolate the dna strand that's in your system to beat this infection." "Now, the instruments to do that don't exist, okay?" "So i have to invent a whole new paradigm." "And i have to ask you to come back to my lab with me." "Because the lives of a lot of people that i care about are at stake here." "So what's in it for me?" "I don't know." "Jesse and emma are checked in." "Good." "Life support will keep them going for three to four hours." "Looks like you two are losing teammates." "I need your help, charles." "Please." "Okay, if i can identify and extract the proper genetic strand then i can execute a massive replication." "Yeah, but how are you going to get the antidote to the rest of the new mutants in time?" "Theoretically, if i can transform the binary components of your genetic code into a microwave form then i can uplink it into the gsa communications satellite and then disperse it across the continent in a biotransmission." "Kind of like a blanket inoculation." "That's the idea." "All right, charles, i'm going to commence extraction." "Alright, the column on the right is cladosporium, the bacterial strain." "On the left here is the microwave form of the antidote." "Now, we're onto something if we can keep the bacterial level under ten." "It's happening." "Look at that." "It's happening." "That's it." "Bleeding." "I'm bleeding." "There's no sign of instant cellular regeneration." "I'll get the first-aid kit." "Yeah." "The extraction must have had a counter-mutating effect on your dna." "Is it permanent?" "I have no idea." "But right now there's nothing indestructible about you." "Hey, when you looked at that photo it looked to me like you recognized this friend of his." "You know what?" "I really don't want to get involved in your domestic problems." "And you wouldn't be if you would have let me get away when i had the chance." "Hey, todd, like i said i really, really don't want to get involved in your domestic problems." "Well, you know what?" "Forget that, honey." "You're stuck here with us." "Which means you're involved." "Hey, did you hear that?" "I didn't hear anything." "Yeah, you just got an e-mail." "No, i didn't." "Hey." "Hey, what did i say about using your powers, speedy?" "Huh." "It's from stanley gambert." "Hey, that's private." ""Missing you terribly, my sweet." "I count the minutes until we're together." "Love, stanley. "" "Stanley... stanley gambert?" "Okay, so we're both cheats." "You happy?" "You know, you're right." "You shouldn't have got involved." "Whenever you two are done tippy toeing around each other's backs, i think there's something you should know." "See, now i wasn't entirely sure when i saw that picture of your, well, girlfriend." "But when you received that e- mail from stanley gambert, all the pieces are fitting together." "Check it out." "One of the things that we do is keep a running... gallery of all the gsa agents and their informants." "And you guys hit the jackpot." "Hoo." "Talk about picking winners." "So ain't this a pickle?" "See, now here you both are trying to use the new mutant underground as a way to ditch each other and, well, run off with your new red hot lovers." "Whose idea, by the way, of a romantic getaway is a statis pod." "So congratulations." "You've been had by the gsa." "Get word to adam." "I want marlowe here with the cure in half an hour or i start removing patients from life support." "Beginning with these two." "The bio-transmission code's here in my handheld." "I'll sneak in and hook it up." "All right, good, i'm going to check on jesse and emma and make sure they're alright." "Hey, wait." "I need you to go with shalimar." "Nah, i did my part." "No, no, no, your retinal scan is the only way to access the gsa communications satellite." "Why should i risk my neck for you?" "Because i just made it possible for you to understand what risk means again." "Okay." "Okay." "You ready?" "I was born ready." "Got it?" "We're in." "Great." "Can i assume you've completed?" "Call in a full retreat, mr eckhart." "What good is your gun against him?" "I must say, i'm very disappointed in you, mr marlowe." "You call in that retreat or you'll never get a chance to get over your disappointment." "This is mr eckhart." "Pull back all units at once." "We're almost there." "You've got these guys well-trained, don't you?" "Shoot him." "Everything's uplinked and ready to go." "I'll call a doctor." "No." "Adam, no." "This is what i want." "He's gone." "Let's do what we came to do." "Now." "That's step one." "Is it working?" "Well, if it is we should see immediate results." "I'll go check the ward." "How are you guys doing?" "Starting to feel better." "Good." "And you." "You are not supposed to be out of bed yet." "He's still got it." "Oh, no." "Oh, no." "That one?" "It was just for fun." "What is wrong with you?" "Hey, welcome." "You guys get through it okay?" "Yeah, twelve hours later and it was like we were never sick." "How about you?" "Yeah, well i survived." "Look, brennan, both jerri and i want to apologize for everything." "And thank you too for saving our lives." "Well, it's no big deal." "So what are your guy's big plans?" "Oh, we're going to give it another shot." "Really?" "Yeah, we both gave it some thought." "And we both realized there's still something there." "Well, the good news is the new mutant underground is a great place for a second chance." "So good luck." "Jerri?" "What the...?" "Oh that, my friend is x-ray vision." "Hey." "Still upset about marlowe?" "Yeah." "Part of me feels responsible for his pain." "You know every new mutant alive today owes their life to you" " for what you did at genomex." "I lost too many to this disease." "Then there's the ones like marlowe that choose to serve eckhart and the gsa." "I just don't understand that." "Someone once told me you can't force right or wrong on anyone." "All you can do is lead by example and hope they lean toward the light." "I'm glad you remembered." "Yeah." "Thanks."
|
{
"pile_set_name": "OpenSubtitles"
}
|
Specificities of human TAP alleles for HLA-B27 binding peptides.
Allelic TAP polymorphism has been linked to susceptibility to Reiter's syndrome and was suggested to influence disease phenotype in HLA-B27 positive patients with ankylosing spondylitis. In the present study, we examined whether the human TAP alleles functionally differ in their translocation specificity for HLA-B27-binding nonamers. TAP translocation of a panel of HLA-B27-binding peptides was measured with a labeled reporter peptide containing an N-linked glycosylation acceptor site in streptolysin O-permeabilized cells with different TAP alleles. The different human TAP alleles tested did not measurably differ in their peptide specificity. The polymorphism of human TAP does not affect the translocated repertoire of HLA-B27 ligands and is therefore unlikely to play a decisive role in the development of HLA-B27-associated disease.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Happy Thanksgiving to All!
If you've ever wondered how Buckley faculty and staff celebrate Thanksgiving, clickhere and you will find out. Thank you, Mr. Topel, for capturing the spirit of this holiday in our community.
Thank you to all of our upper school students, our kindergärtners, and their music/dance teachers for preparing a beautiful Thanksgiving concert for all of us to enjoy yesterday: Mr. Aly, Mrs. Baritot, Mrs.Chun, Ms. Coon, Mrs. Demopoulos, Dr. Glennon, Mr. Mulz, Mrs. Stadlander,
Thank you to Mr. Persaud for organizing our Thanksgiving food and coat drive and to all of you for your generous donations of food, money, and winter coats. Sixty dinner boxes and coats were delivered to Hillside House, a shelter for women and their children in Jamaica, Queens. In addition, boxes of food were delivered to the Interfaith Nutrition Network in Hempstead. Please clickhereto see how this all happened.
From our family to yours, I would like to wish all the members of the Buckley community a very, very happy Thanksgiving.
|
{
"pile_set_name": "Pile-CC"
}
|
#
Thinkstock
#
Copyright © 2014 by Julie Ganz
All rights reserved. No part of this book may be reproduced in any manner without the express written consent of the publisher, except in the case of brief excerpts in critical reviews or articles. All inquiries should be addressed to Skyhorse Publishing, 307 West 36th Street, 11th Floor, New York, NY 10018.
Skyhorse Publishing books may be purchased in bulk at special discounts for sales promotion, corporate gifts, fund-raising, or educational purposes. Special editions can also be created to specifications. For details, contact the Special Sales Department, Skyhorse Publishing, 307 West 36th Street, 11th Floor, New York, NY 10018 or info@skyhorsepublishing.com.
Skyhorse® and Skyhorse Publishing® are registered trademarks of Skyhorse Publishing, Inc.®, a Delaware corporation.
Visit our website at www.skyhorsepublishing.com.
10 9 8 7 6 5 4 3 2 1
Library of Congress Cataloging-in-Publication Data is available on file.
Cover design by Danielle Ceccolini
Cover photo credit Thinkstock
ISBN: 978-1-62914-472-6
Ebook ISBN: 978-1-63220-030-3
Printed in China
To my travel-loving family, who has encouraged me to travel the world.
And to Joe—our journey is just beginning.
Acknowledgments
I'd like to thank my family members, both immediate and extended, for providing many of the color photographs that appear in this book. I hope seeing them here makes your travel experiences come alive again.
Contents
Introduction
Part One They Came, They Saw, They Conquered
Part Two Hot 'N' Cold
Part Three Exploration
Part Four Trains, Planes, and Automobiles
Part Five Accommodations
Part Six Traveling the Continents
Index of those quoted
INTRODUCTION
Traveling cross-town in Manhattan every morning to get to work is one of my favorite times of the day. No, it's not because I'm a "morning person;" nor is it because I enjoy being herded like cattle onto the subway. Rather, it is for the journey, albeit brief, that I take to get there and for the people whom I meet (and sometimes interact with) along the way. Rarely does this journey seem routine, and by the time I get to the office, I have already learned one new thing, been fascinated by another person's story, or observed something that I had never before seen.
On arguably a smaller scale, this "trip" to work contains the facets of travel about which I am most passionate. I can admit that I much prefer to be the tourist rather than to dodge the wide-eyed New York City tourists standing in the middle of the sidewalk gazing up at the sky (or the ceiling). Nevertheless, as soon as they open their mouths, they enable me to take a brief journey of my own. Some may call it daydreaming, but I prefer to call it reminiscing . . .
I pass by a few Spanish speakers, and in doing so I travel back to Spain, where my host family is fervently discussing a TV program as I try with all my brain power to catch each and every word that seems to be flying off their tongues.
I assist a couple lost on the subway, and in the process I recall the time that, as a teenager who knew not a word of French, I wandered around Paris by myself. I had lost my friend on the Champs de Elyssees, at which point panic began to set in, before I eventually found her where I had started.
I grin as a family engages a NYC tour bus employee, as I recall riding the double-decker bus in London with my own family members.
I admire the way the visitors view the old cathedrals, synagogues, and various other places of worship, as I remember walking from cathedral to cathedral in Italy and admiring the exquisite synagogues in Israel.
I long to bask in the sun with the couple sitting on the bench, or the teenagers at the park, as I think fondly of leisurely days spent in the Aruban sun.
I try to put myself in the shoes of the young woman who takes in a seemingly commonplace Manhattan street scene, and I recall walking the streets of Morocco and marveling at its unique beauty.
Naturally, most travel requires time and resources. But those feelings that we undergo doing something—anything—new, even on our walk to work, can be re-experienced over and over again.
Technology often plays a part, for better or for worse, in this ability to experience new things over and over again, via a camera lens, a computer, or a movie screen. Nevertheless, it is important that, when time and money allow, we open ourselves to new experiences, which enable us to climb out of our comfort zone and to see things for the first time. There's nothing like seeing the world's wonders in person.
I could go on forever singing the praises of travel, but the fact of the matter is that the quotes in this book do a much better job of this. And for that reason, I will leave you with some very brief words of wisdom (if I can be so bold as to proclaim them as such): remember that experience is not defined by how many days it's been since you left home or how many miles you've traveled. Whether you're crossing the bridge to the neighboring city or making your way around the world, it's usually more enjoyable if you focus on how you feel and what you take away from the journey.
That's what you'll remember when you arrive home.
—Julie Ganz,
Spring 2014
Thinkstock
PART ONE
They Came, They Saw, They Conquered
The real voyage of discovery consists not in seeking new landscapes, but in having new eyes.
—MARCEL PROUST
• • •
Two roads diverged in a wood and I—I took the one less traveled by.
—ROBERT FROST
• • •
For my part, I travel not to go anywhere, but to go. I travel for travel's sake. The great affair is to move.
—ROBERT LOUIS STEVENSON,
Travels with a Donkey in the Cevennes
• • •
I was a traveler! The word never had tasted so good in my mouth before. I had an exultant sense of being bound for mysterious lands and distant climes which I never have felt in so uplifting a degree since.
—MARK TWAIN,
Life on the Mississippi
• • •
Travel brings power and love back into your life.
—RUMI
• • •
A man of ordinary talent will always be ordinary, whether he travels or not; but a man of superior talent (which I cannot deny myself to be without being impious) will go to pieces if he remains forever in the same place . . .
—WOLFGANG AMADEUS MOZART
• • •
I think that travel comes from some deep urge to see the world, like the urge that brings up a worm in an Irish bog to see the moon when it is full.
—LORD DUNSANY
• • •
I beg young people to travel. If you don't have a passport, get one. Take a summer, get a backpack and go to Delhi, go to Saigon, go to Bangkok, go to Kenya. Have your mind blown, eat interesting food, dig some interesting people, have an adventure, be careful.
—HENRY ROLLINS
• • •
I am not the same having seen the moon shine on the other side of the world.
—MARY ANNE RADMACHER
• • •
My pictures are evidence that I was there, that I cared enough to pay attention, that I noticed, and honored, those tiny miracles of life we are all given, along whatever path we have chosen to travel. Now, when I travel, I feel a simultaneous quickening of desire, and a thickening of time.
—DOMINIQUE BROWNING,
The New York Times
• • •
I like animals. I like natural history. The travel bit is not the important bit. The travel bit is what you have to do in order to go and look at animals.
—DAVID ATTENBOROUGH
• • •
Delight in travel has for long been an English characteristic, but among the young men of the 'twenties the cult became an obsession.
—CHRISTOPHER SYKES
• • •
Tourists and travelers aren't so different. They might characterize each other in stereotypical fashion but essentially they both want to enjoy being somewhere else.
—DICCON BEWES
• • •
Travel penetrates your consciousness, but not in a rational way.
—MARTIN GLASER
• • •
I knew that travel would also reinforce those shared traits that make us all members of one human family: the laughter, the tears, the need for community, the love of children, and that special tingle of pleasure we all feel when we touch each other's hearts.
—RITA GOLDEN GELMAN,
Female Nomad and Friends
• • •
Travel is "Maybe I don't have to do it that way when I get back home."
—NICK MILLER,
Isn't It Pretty to Think So?
• • •
Travel is a beautiful thing but not because it is a competition. It really doesn't matter who has visited more countries or who has done more things. It is the experience. This is the main reason why I have tried to immerse myself in cultures rather than traveling from hotel to hotel.
—MARCELLO ARRAMBIDE
• • •
When we travel, we learn as much about ourselves as we do about a place.
—DOMINIQUE BROWNING,
The New York Times
• • •
Being a local is often a really great way to ignore your surroundings. And if that's true, why the hell are travelers asking locals for advice? Surely those people have the most blinkered and jaded opinions in town?
—MIKE SOWDEN
• • •
A truly global pilgrimage would provide a deeper understanding of how individuals live in, and change, the world. As a committed traveler, I knew that diving into other cultures and gathering the fruits of wanderlust enabled one to see the world with a wider, more original lens.
—HOLLY MORRIS,
Adventure Divas
• • •
Perhaps travel cannot prevent bigotry, but by demonstrating that all peoples cry, laugh, eat, worry, and die, it can introduce the idea that if we try and understand each other, we may even become friends.
—MAYA ANGELOU
• • •
Even the confident, resourceful, and independent traveler has much to gain from the solitary journey—
whether it's perspective on a life left behind, unexpected connections with locals, insights into how individual relationships are separate from the realm of global politics, or the simple and pure joy of a few anonymous days of not having to organize our hours around the needs of anyone else.
—FAITH CONLON AND INGRID EMERICK,
Go Your Own Way
• • •
The tourist is the other fellow.
—EVELYN WAUGH
• • •
The difference between being a tourist and a traveler is that a traveler is open to unplanned experience and doesn't have her nose stuck in a guidebook, tracking down famous sites. She ventures out from behind glass windows (in hotels and touring buses) and meets people. She connects.
—JUDITH FEIN,
Life is a Trip
• • •
It only makes sense to get to know who your neighbors are, to see their heritage, history, and the beautiful places they come from.
—ANTHONY MELCHIORRI,
"Hotel Impossible" Host
• • •
Later, I'll look back and recognize this rush of excitement as my first glimpse of what exactly it means to travel alone in a faraway land: I can go anywhere. I can do anything. And the all-important: I can be anyone.
—RACHEL FRIEDMAN,
The Good Girl's Guide to Getting Lost
• • •
I have numerous plans rolling around upstairs: cycling around the world, taking a year out to explore the canals of Great Britain on a boat, restoring a campervan and travelling around Europe for the summer or . . . walking a stupid distance through a part of the world that beckons to me.
—KEITH FOSKETT,
The Last Englishman
• • •
I never think, "Oh, I've done all that, that's it.
That's over." I know until the end of my life, they'll always be somewhere I want to go, that I've not been to. And that's what sort of keeps you young.
—MICHAEL PALIN
• • •
Stop worrying about the potholes in the road and enjoy the journey.
—BABS HOFFMAN
• • •
Travel makes one modest, you see what a tiny place you occupy in the world.
—GUSTAVE FALUBERT
• • •
People travel to faraway places to watch, in fascination, the kind of people they ignore at home.
—DAGOBERT D. RUNES
• • •
My personal survival method? Escape. Even now, dripping with sweat and frantically racing to make it across country lines, I felt that familiar burst of exhilaration that flooded me every time I booked an international flight or added a new stamp to my passport.
—JENNIFER BAGGETT,
The Lost Girls
• • •
It's not just taking pictures of the Eiffel Tower, or even a wonderful volcano or Victoria Falls. I get the most out of it when I meet people who live in conditions very different to me, cultures very different to me, eat different food, have different religions . . .
—MICHAEL PALIN
• • •
Roslyn and Aaron Ganz
It occurred to me that if travel broadened the mind it could also be a very leveling experience.
—MERRIDY EASTMAN,
How Now Brown Frau
• • •
You lose sight of things . . . and when you travel, everything balances out.
—DARANNA GIDEL
• • •
The first condition of understanding a foreign country is to smell it.
—RUDYARD KIPLING
• • •
Traveling tends to magnify all human emotions.
—PETER HOEG
• • •
People who are more interested in others than they are in themselves are, well, more interesting. We are better as people and as a culture when we learn more about other people and cultures, and the best way to do that is to go meet them and experience them.
—SPUD HILTON
• • •
No one realizes how beautiful it is to travel until he comes home and rests his head on his old, familiar pillow.
—LIN YUTANG
• • •
Most of my treasured memories of travel are recollections of sitting.
—ROBERT THOMAS ALLEN
• • •
Any successful journey begins by packing your luggage full of imagination.
—KATHRINE PALMER PETERSON
• • •
Twenty years from now you will be more disappointed by the things you didn't do than by the ones you did do.
So throw off the bowlines, sail away from the safe harbor.
Catch the trade winds in your sails. Explore. Dream. Discover.
—MARK TWAIN
• • •
As the traveler who has once been from home is wiser than he who has never left his own doorstep, so a knowledge of one other culture should sharpen our ability to scrutinize more steadily, to appreciate lovingly, our own.
—MARGARET MEAD
• • •
Travel is the frivolous part of serious lives, and the serious part of frivolous ones.
—ANNE SOPHIE SWETCHINE
• • •
If a common theme was emerging to these crusades, then it was the search for something beyond the typical tourist routine, an antidote to the vacuous consumerism of contemporary travel. A trip to the moral high ground—
I hear the view's excellent from up there.
—TIM MOORE,
Travels with My Donkey
• • •
Travel, in the younger sort, is a part of education; in the elder, a part of experience.
—FRANCIS BACON
• • •
Travel is so much more than getting on a plane and jetting off to some foreign locale; it's about a journey of self-discovery and fulfillment that can positively impact the world while you are doing it.
—MARK MURPHY
• • •
It can hardly be a coincidence that no language on Earth has ever produced the phrase, "as pretty as an airport."
Airports are ugly. Some are very ugly. Some attain a degree of ugliness that can only be the result of a special effort.
—DOUGLAS ADAMS,
The Long Dark Tea-Time of the Soul
• • •
Half the fun of the travel is the esthetic of lostness.
—RAY BRADBURY
• • •
Once a year, go someplace you've never been before.
—DALAI LAMA
• • •
Every traveler has two or three or even a hundred of them: moments on a journey when you taste something and you're forever changed.
—JAMES OSELAND,
A Fork in the Road
• • •
And that's the wonderful thing about family travel:
it provides you with experiences that will remain locked forever in the scar tissue of your mind.
—DAVE BARRY
• • •
I was a vagabond, a young man on the greatest adventure of his life, and vagabonds do not call their parents for extra calling card minutes. In fact, I wasn't sure vagabonds even had parents.
—CHRISTOPHER COFFMAN
• • •
Bizarre travel plans are dancing lessons from God.
—KURT VONNEGUT
• • •
He who would travel happily must travel light.
—ANTOINE DE SAINT-EXUPERY
• • •
I haven't been everywhere, but it's on my list.
—SUSAN SONTAG
• • •
Travel changes you. As you move through this life and this world you change things slightly, you leave marks behind, however small. And in return, life—and travel—leaves marks on you. Most of the time, those marks—on your body or on your heart—are beautiful. Often, though, they hurt.
—ANTHONY BOURDAIN
• • •
But we are all travelers in what John Bunyan calls the wilderness of this world—all, too, travelers with a donkey: and the best that we find in our travels is an honest friend.
—ROBERT LOUIS STEVENSON,
Travels with a Donkey in the Cevennes
• • •
The only aspect of our travels that is guaranteed to hold an audience is disaster . . . The fact is, we cherish our disasters and here we are one up on the great travellers who have every impressive qualification for the job but lack jokes.
—MARTHA GELLHORN,
Travels with Myself and Another
• • •
Travelers never think that they are the foreigners.
—MASON COOLEY
• • •
Roslyn and Aaron Ganz
You'll never understand people if you live in one place.
If you really want to understand people you're going to have to travel.
—MATTHEW KEPNES
• • •
Why do you go away? So that you can come back.
So that you can see the place you came from with new eyes and extra colors. And the people there see you differently, too. Coming back to where you started is not the same as never leaving.
—TERRY PRATCHETT
• • •
Traveling is a brutality. It forces you to trust strangers and to lose sight of all that familiar comfort of home and friends. You are constantly off balance. Nothing is yours except the essential things—air, sleep, dreams, the sea, the sky—all things tending towards the eternal or what we imagine of it.
—CESARE PAVESE
• • •
Travel is not really about leaving our homes, but leaving our habits.
—PICO IYER
• • •
The farther you go, however, the harder it is to return. The world has many edges, and it's easy to fall off.
—ANDERSON COOPER
• • •
Make voyages! Attempt them . . . there's nothing else.
—TENNESSEE WILLIAMS
• • •
As one expert put it, "Travel is good medicine.
Because it challenges the brain with new and different experiences and environments, it is an important behavior that promotes brain health and builds brain resilience across the lifespan."
—SAMI HASSANYEH,
chief digital officer, AARP
• • •
To travel. An intransitive verb. A state of being, not a journey to a destination.
—JONATHAN RABAN,
Coasting
• • •
There is no moment of delight in any pilgrimage like the beginning of it.
—CHARLES DUDLEY WARNER
• • •
There were so many places I wanted to go, but not just to go. I wanted to experience living in these places, not only visiting for a week or two.
—LYNNE MARTIN,
Home Sweet Anywhere
• • •
I have found out that there ain't no surer way to find out whether you like people or hate them than to travel with them.
—MARK TWAIN
• • •
When you travel you see what the world is actually like.
You find out that a vast majority of people simply want peace, they want enough food to eat, they want their children to have a better life than they do.
—PAULINE FROMMER
• • •
What the traveler cannot see he imagines, another motive for his travels.
—JOSE SARAMAGO,
Journey to Portugal
• • •
At first I believed in pure travel, and that it was necessary never to have a purpose . . . Still later I began to weave the background and the incidents of travel into my novels, and now, as I observe the change that has taken place over the years, I wonder if I am any longer capable of enjoying travel for its own sake.
—NORMAN LEWIS,
View of the World
• • •
A traveler without observation is a bird without wings.
—MOSLIH EDDIN SAADI
• • •
I said, Mary Frances, now is the time to go to new places and find different views of things . . .
It was a vivid period of slow wandering, very rich, like a carpet I had often trod before I realized that it was there.
—M. F. K. FISHER,
As They Were
• • •
"Travel" is the name of a modern disease which became rampant in the mid-fifties and is still spreading.
The disease—its scientific name is travelitis furiosus—is carried by a germ called prosperity.
—GEORGE MIKES
• • •
Travel empties out everything you've put into the box called your life, all the things you accumulate to tell yourself who you are.
—CLAIRE FONTAINE AND MIA FONTAINE,
Have Mother, Will Travel
• • •
There is nothing like returning to a place that remains unchanged to find the ways in which you yourself have altered.
—NELSON MANDELA
• • •
True and sincere travelling is no pastime but it is as serious as the grave, or any part of the human journey, and it requires a long probation to be broken into it.
—HENRY DAVID THOREAU,
A Week on the Concord and Merrimack Rivers
• • •
The tourist thrives on the uncanny, moving happily through a phenomenal world of effects without causes.
—JONATHAN RABAN,
Driving Home
• • •
All journeys have secret destinations of which the traveler is unaware.
—MARTIN BUBER
• • •
Traveling is not just seeing the new; it is also leaving behind. Not just opening doors; also closing them behind you, never to return. But the place you have left forever is always there for you to see whenever you shut your eyes.
—JAN MYRDAL
• • •
It's hard to make travel arrangements to visit a dream.
The voyage I was planning was on a river which existed only in my head. The real Mississippi was an abstraction.
—JONATHAN RABAN,
Old Glory
• • •
The best litmus test for measuring your vagabonding gumption is found not in travel but in the process of earning your freedom to travel.
—ROLF POTTS,
Vagabonding: An Uncommon Guide to the Art of Long-Term World Travel
• • •
Do not follow where the path may lead. Go instead where there is no path and leave a trail.
—RALPH WALDO EMERSON
• • •
Not all those who wander are lost.
—J. R. R. TOLKIEN
• • •
Learning how to interact with people from other countries and cultures equips future leaders in all sectors to address urgent issues—from curing diseases and finding energy solutions, to fighting terrorism and hunger—shared across borders.
—ALLAN E. GOODMAN AND STACIE NEVADOMSKI BERDAN,
The New York Times
• • •
I conceived these trips as vacations—as respites from my ongoing life—but that wasn't how they turned out.
—MICHAEL CRICHTON,
Travels
• • •
It is not down in any map; true places never are.
—HERMAN MELVILLE,
Moby Dick
• • •
The best thing about a vacation is planning it.
—ANDY ROONEY,
Years of Minutes
• • •
It's a dangerous business, Frodo, going out your door.
You step onto the road, and if you don't keep your feet, there's no knowing where you might be swept off to.
—J. R. R. TOLKIEN
• • •
I have just been all round the world and have formed a very poor opinion of it.
—SIR THOMAS BEECHAM
• • •
Travel becomes a political act only if you actually do something with your broadened perspective once you return home.
—RICK STEVES
• • •
Travelling is like flirting with life. It's like saying,
"I would stay and love you, but I have to go; this is my station."
—LISA ST. AUBIN DE TERAN
• • •
You can fall in love at first sight with a place as well as a person.
—ALEC WAUGH
• • •
Travelers learn to celebrate, rather than fear, the diversity on our planet. Learning in a different culture and place allows us to see our own challenges in sharp contrast, and with more clarity, as we observe smart people in other lands dealing with similar issues.
—RICK STEVES
• • •
I travel a lot; I hate having my life disrupted by routine.
—CASKIE STINNETT
• • •
All the escapism found in a good novel can be found first-hand in travel.
—CHRIS ERSKINE,
LA Times
• • •
Though we travel the world over to find the beautiful, we must carry it with us or we find it not.
—RALPH WALDO EMERSON
• • •
Lisa Ganz
Before I traveled around the world at age twenty-nine, I had never gone to a movie or out to eat alone.
On airplanes I was so scared I took sleeping pills.
Now, after years of travel and risk taking, I savor the moments I have to myself.
—MARYBETH BOND,
Gutsy Women
• • •
When you're traveling, you are what you are right there and then. People don't have your past to hold against you. No yesterdays on the road.
—WILLIAM LEAST HEAT-MOON
• • •
A journey is best measured in friends, not in miles.
—TIM CAHILL
• • •
Traveling is almost like talking with men of other centuries.
—RENÉ DESCARTES
• • •
Travel can introduce kids to the world's real-life wonders, changing their perspective on topics they may have only read about in books.
It can literally change their lives.
—KATIA HETTER,
CNN
• • •
You got to be careful if you don't know where you're going, because you might not get there.
—YOGI BERRA
• • •
Good company in a journey makes the way seem shorter.
—IZAAK WALTON
• • •
One's destination is never a place, but rather a new way of looking at things.
—HENRY MILLER
• • •
Through travel I first became aware of the outside world;
it was through travel that I found my own introspective way into becoming a part of it.
—EUDORA WELTY
• • •
A good vacation is over when you begin to yearn for your work.
—MORRIS FISHBEIN
• • •
This vacation tip: If you see a gas station with a sign that says "Clean Rest Rooms," it means one thing—
the door is locked and no one can find the key.
—PAT WILLIAMS,
Winning with One-Liners
• • •
I have wandered all my life, and I have traveled;
the difference between the two is this—we wander for distraction, but we travel for fulfillment.
—HILAIRE BELLOC
• • •
A travel adventure has no substitute. It is the ultimate experience, your one big opportunity for flair.
—ROSALIND MASSOW
• • •
Something as simple as going out for a cup of coffee can have an impact—no matter how large or small—on your life and the lives of those around you . . . If a trip to the coffee shop can do this much, imagine what a trip around the world can do!
—AMBER ISRAELSEN
• • •
Travel is more than the seeing of sights; it is a change that goes on, deep and permanent, in the ideas of living.
—MIRIAM BEARD
• • •
Comfort has its place, but it seems rude to visit another country dressed as if you've come to mow its lawns.
—DAVID SEDARIS
• • •
Like all great travelers, I have seen more than I remember, and remember more than I have seen.
—BENJAMIN DISRAELI
• • •
Thanks to the Interstate Highway System, it is now possible to travel across the country from coast to coast without seeing anything.
—CHARLES KURALT,
On the Road With Charles Kuralt
• • •
Travel is like love, mostly because it's a heightened state of awareness, in which we are mindful, receptive, undimmed by familiarity and ready to be transformed.
That is why the best trips, like the best love affairs, never really end.
—PICO IYER
• • •
We travel, initially, to lose ourselves; and we travel, next, to find ourselves.
—PICO IYER
• • •
Travel can be one of the most rewarding forms of introspection.
—LAWRENCE DURRELL
• • •
We all know travel elitists: They will actually pick a dinner party fight about whether tourists are real travelers. Or whether a vacation is really a trip.
They scoff at the mere suggestion of mass tourism.
They have canned, over-intellectualized, one-size-fits-all answers as to the "right way" to travel . . . that is usually their way of traveling.
—WILLIAM D. CHALMERS,
The Huffington Post
• • •
Is there a better icon of family life than an SUV loaded with tents, beach balls and canoe paddles?
Many of our fondest memories are made while on vacation, as children or as parents.
—CHRIS ERSKINE,
LA Times
• • •
I find it fascinating that most people plan their vacations with better care than they plan their lives.
Perhaps that's because escape is easier than change.
—JIM ROHN
• • •
A vacation is what you take when you can no longer take what you've been taking.
—EARL WILSON
• • •
We are living proof that one extremely bad trip doesn't need to decide the rest of your lives.
Without doubt, travel will have its ups and downs, but we are prepared to face those downs with grace and move on.
—DAVE BOUSKILL AND DEB CORBEIL
• • •
For the born traveler, traveling is a besetting vice.
Like other vices, it is imperious, demanding its victim's time, money, energy and the sacrifice of comfort.
—ALDOUS HUXLEY
• • •
To travel is to discover that everyone is wrong about other countries.
—ALDOUS HUXLEY
• • •
Some travel surprises are good: discovering a hidden gem of a trattoria in Florence where you are greeted with hugs and an open bottle of Chianti.
Or finding out that the Louvre is free the one day you are in Paris. But then there are the bad surprises: getting the final bill on your cruise and seeing hundreds of dollars tacked on for gratuities.
—TERRY WARD, BUDGET TRAVEL
• • •
Once you have traveled, the voyage never ends, but is played out over and over again in the quietest chambers. The mind can never break off from the journey.
—PAT CONROY
• • •
A good traveler has no fixed plans and is not intent on arriving.
—LAO TZU
• • •
Traveling will change you like little else can.
It will put you in places that will force you to care for issues that are bigger than you.
—JEFF GOINS
• • •
Really. Is there anything nice to be said about other people's vacations?
—AMOR TOWLES,
Rules of Civility
• • •
Babies don't need a vacation but I still see them at the beach. I'll go over to them and say, "What are you doing here, you've never worked a day in your life!"
—STEVEN WRIGHT
• • •
As is often the case when I travel, my vulnerability—
like not knowing what the hell I'm going to do upon arrival—makes me more open to outside interactions than I might be when I'm at home and think I know best what needs to be done. On the road, serendipity is given space to enter my life.
—ANDREW MCCARTHY,
The Longest Way Home: One Man's Quest for the Courage to Settle Down
• • •
I believe travel is a necessity, and that no matter our age, we are all on a journey to understand ourselves, our world and who we are in it. Nothing gives us that opportunity like travel.
—SAMANTHA BROWN, AARP
• • •
I'm a big believer that you're never going to find perfect city travel experience or the perfect meal without a constant willingness to experience a bad one.
Letting the happy accident happen is what a lot of vacation itineraries miss, I think, and I'm always trying to push people to allow those things to happen rather than stick to some rigid itinerary.
—ANTHONY BOURDAIN
• • •
Voyage, travel, and change of place impart vigor.
—SENECA
• • •
The world is a book, and those who do not travel read only a page.
—SAINT AUGUSTINE
• • •
A tourist is a fellow who drives thousands of miles so he can be photographed standing in front of his car.
—EMILE GANEST
• • •
Like writing in a journal—another great form of stress-relief—traveling alone allows you to witness how you react to different situations.
—CAROLYN GREGOIRE,
The Huffington Post
• • •
People don't take trips—trips take people.
—JOHN STEINBECK
• • •
Don't tell me how educated you are, tell me how much you have traveled.
—MOHAMMED
• • •
People travel to wonder at the height of mountains, at the huge waves of the sea, at the long courses of rivers, at the vast compass of the ocean, at the circular motion of the stars; and they pass by themselves without wondering.
—SAINT AUGUSTINE
• • •
Michael Ganz
In both business and personal life, I've always found that travel inspires me more than anything else I do.
Evidence of the languages, cultures, scenery, food, and design sensibilities that I discover all over the world can be found in every piece of my jewelry.
—IVANKA TRUMP
• • •
The medieval pilgrimage routes . . . are the beginnings of the modern tourism industry. Which is to say that you can draw a more or less straight line from a Dark Ages peasant blistering his feet trudging to a church displaying the Virgin Mary's dried-up breast milk to me vomiting into a barf bag on a sightseeing boat . . .
—SARAH VOWELL,
Assassination Vacation
• • •
The words "travel" and "vacation" are often used interchangeably, but careful examination indicates subtle differences between the two. Vacation implies an escape, while travel may offer the opportunity for total immersion in a different culture.
—LISA MERCER,
Demand Media
• • •
Travel broadens the mind but also the hips unfortunately, (because who can resist sampling new cuisine?).
—KATHY LETTE
• • •
Being a tourist can give you access to experiences you wouldn't have otherwise—experiences that aren't so much exclusive as inclusive, drawing their appeal from the company of other people.
—PETER JON LINDBERG
• • •
"Nothing," Homer sings, "is harder on mortal man than wandering." That's why the words travel and travail have a common origin.
—WILLIAM LEAST HEAT MOON,
Blue Highways
• • •
While you're still young, you should get cultured.
Get to know the world and the magnificent people that fill it. There's nothing quite like walking alongside the Colosseum or seeing Michelangelo's David in person.
—JEFF GOINS
• • •
The world . . . is no longer large, and to ignore it likely requires more effort than simply to take notice.
—TOM BISELL,
Chasing the Sea
• • •
Don't be a tourist. Plan less. Go slowly. I traveled in the most inefficient way possible and it took me exactly where I wanted to go.
—ANDREW EVANS,
National Geographic
• • •
Because we have traveled a lot, we're kind of lucky in that we have friends around the world.
—ANGELINA JOLIE
• • •
While facing the future without a plan might not be the best way to live, I'm convinced it's the best way to travel.
—ERIC RAFF,
No Sense of Direction
• • •
Sometimes the best thing you can do for yourself is jump on a damn plane and go somewhere.
—LEANNE ADAMS,
Pretty Woman Spitting
• • •
I'm the last person in the world—admittedly with the possible exceptions of Britain's Queen and Mariah Carey—who should go backpacking.
While all sorts of people go travelling, only a few rare kinds throw a pack on their back, fly to a faraway land and hit the Not Completely Beaten Track.
—CATHERINE RYAN HOWARD,
Backpacked
• • •
When I first started traveling years ago, and went to a few countries, I was very emotional about what I saw and it changed me as a person and as a mother.
—ANGELINA JOLIE
• • •
I never travel without my diary. One should always have something sensational to read in the train.
—OSCAR WILDE
• • •
There are two kinds of vagabonds: those who make room in their backpacks for an inflatable clothes hanger, and those who don't.
—MIKE MCINTYRE,
The Wander Year
• • •
It's become a little bit of an obsession of mine to try to find items to take with me the next time I travel. I'll look at a shirt and think, "Hmm, that's a shirt I'll be able to wear all day on a trip . . ."
—JOHN CHO
• • •
Often I feel I go to some distant region of the world to be reminded of who I really am.
—MICHAEL CRICHTON,
Travels
• • •
I met a lot of people in Europe. I even encountered myself.
—JAMES BALDWIN
• • •
After ten years as a travel writer in the UK, I was looking for something different, so moving abroad and learning a new language seemed like a good idea—until I actually started having to grapple with German grammar.
—DICCON BEWES
• • •
A man who leaves home to mend himself and others is a philosopher; but he who goes from country to country, guided by a blind impulse of curiosity, is a vagabond.
—OLIVER GOLDSMITH,
The Citizen of the World
• • •
They say travel broadens the mind; but you must have the mind.
—G. K. CHESTERTON,
The Shadow of the Shark
• • •
To travel is worth any cost or sacrifice.
—ELIZABETH GILBERT,
Eat, Pray, Love
• • •
That's not a question you ask a willing nomad!
Everywhere has something beautiful to it. The end!
—ERIN O'CONNOR,
when asked about her least favorite travel spot
• • •
If there isn't anyone else around to talk to, I could probably make friends with a four-foot-tall pile of sheetrock. That is why I'm not afraid to travel to the most remote places in the world, not if there are human beings there to meet.
—ELIZABETH GILBERT,
Eat, Pray, Love
• • •
I daydreamed about having the freedom to travel wherever chance or fancy took me, unencumbered by schedules and obligations and too many preplanned destinations.
—ALICE STEINBACH,
Without Reservations
• • •
The traveler sees what he sees, the tourist sees what he has come to see.
—G. K. CHESTERTON
• • •
Land was stationary and always belonged to somebody.
Water, on the other hand, was free. It moved and shifted and traveled.
—ROSEMARY MAHONEY,
Down the Nile
• • •
I was halfway across America, at the dividing line between the East of my youth and the West of my future.
—JACK KEROUAC,
On the Road
• • •
Blame it on Daphne du Maurier's Frenchman's Creek or Johnny Depp as Captain Jack Sparrow in the Pirates of
the Caribbean films . . . if a holiday destination comes with a tale involving a skull and crossbones, I'm up for it in spades.
—PENNY SMITH,
The Daily Mail
• • •
I was far away from home, haunted and tired with travel . . . and I looked at the cracked high ceiling and didn't really know who I was for about fifteen strange seconds.
—JACK KEROUAC,
On the Road
• • •
The joy of life comes from our encounters with new experiences, and hence there is no greater joy than to have an endlessly changing horizon, for each day to have a new and different sun . . . Once you become accustomed to such a life you will see its full meaning and its incredible beauty.
—JON KRAKAUER,
Into the Wild
• • •
I love to travel, my thing is I just love going away.
—ALAN CARR
• • •
The world is a never-ending rabbit hole of majesty when it comes to nature. When you think you've seen the most beautiful landscape that's ever existed, there's another just around the corner waiting to take your breath away.
—GABRIELA KRUSCHEWSKY,
The Huffington Post
• • •
He who does not travel does not know the value of men.
—MOORISH PROVERB
• • •
Life is not measured by the number of breaths we take but by the places and moments that take our breath away.
—PATRICIA SCHULTZ
• • •
The past makes its presence known no matter where we travel.
—ALICE STEINBACH,
Without Reservations
• • •
What gives value to travel is fear.
—ALBERT CAMUS
• • •
Please be a traveler, not a tourist. Try new things, meet new people, and look beyond what's right in front of you. Those are the keys to understanding this amazing world we live in.
—ANDREW ZIMMERN
• • •
Some people acquire the travel bug. Others are born with it.
—ERIC WEINER,
The Geography of Bliss
• • •
You can never be too ready for what might happen during holiday travels.
—KIMBER CRANDALL,
Fox News
• • •
Three words should describe your travels as a twenty-something: beautiful, rough, messes.
—GABRIELA KRUSCHEWSKY,
The Huffington Post
• • •
For me, a place unvisited is like an unrequited love.
A dull ache that—try as you might to think it away, to convince yourself that she really wasn't the right country for you—just won't leave you in peace.
—ERIC WEINER,
The Geography of Bliss
• • •
There is one voyage, the first, the last, the only one.
—THOMAS WOLFE
• • •
Not I, nor anyone else, can travel that road for you. You must travel it for yourself.
—WALT WHITMAN
• • •
A journey is like marriage. The certain way to be wrong is to think you control it.
—JOHN STEINBECK,
Travels with Charley
• • •
A trip, a safari, an exploration, is an entity, different from all other journeys. It has personality, temperament, individuality, uniqueness. A journey is a person in itself; no two are alike.
—JOHN STEINBECK,
Travels with Charley
• • •
To young men contemplating a voyage I would say go.
—JOSHUA SLOCUM
• • •
All I really want to do on holiday is sit by a pool and enjoy wonderful food.
—CHEF RAYMOND BLANC
• • •
Raise your fork if you travel for the food! The best part of exploring a nation through its culinary tradition is that—by taking a class or two—you can bring the journey home to share with family and friends.
—STEPHANIE ELIZONDO GRIEST,
100 Places Every Woman Should Go
• • •
The use of traveling is to regulate imagination by reality, and instead of thinking how things may be, to see them as they are.
—SAMUEL JOHNSON
• • •
When you travel, remember that a foreign country is not designed to make you comfortable. It is designed to make its own people comfortable.
—CLIFTON FADIMAN
• • •
All travel has its advantages. If the passenger visits better countries, he may learn to improve his own. And if fortune carries him to worse, he may learn to enjoy it.
—SAMUEL JOHNSON
• • •
Honestly, I prefer to go to places where no one else goes that haven't been ruined by the touristy thing.
—CLAY AIKEN
• • •
Sometimes the road less traveled is less traveled for a reason.
—JERRY SEINFELD
• • •
The more I travel, the more I read, and the more people I talk to, the more I am inspired by the great potential of each human being.
—JANE GOODALL WITH MARC BEKOFF,
The Ten Trusts
• • •
I looked north, in its direction—the very thought of that bridge a beacon to me. I looked south, to where I'd been, to the wild land that'd schooled and scorched me, and considered my options. There was only one, I knew.
There was always only one. To keep walking.
—CHERYL STRAYED,
Wild
• • •
There's no such thing as a bad trip, just good travel stories to tell back home.
—PATRICIA SCHULTZ,
1,000 Places to See Before You Die
• • •
No one travelling on a business trip would be missed if he failed to arrive.
—THORSTEIN VEBLEN
• • •
If you're waiting for a special occasion to make your next trip happen, then consider this: The day you get off the couch and head for the airport, that's the special occasion.
—PATRICIA SCHULTZ,
1,000 Places to See Before You Die, second edition
• • •
Travel, we agreed, was a litmus test: if we could make the best of the chaos and serendipity that we'd inevitably meet in transit, then we'd surely be able to sail through the rest of life together just fine.
—JULIA CHILD,
My Life in France
• • •
To constantly look for yourself even when you encounter the foreign is never to travel; while we know that it's in observing and overhearing others as we journey that we often discover unexpected dimensions of ourselves.
—AMIT CHAUDHURI
• • •
In a profounder sense the best travel writers are not really writing about travel at all. They are recording the effects of places or movements upon their own particular temperaments—recording the experience rather than the event, as they might make literary use of a love affair, an enigma or a tragedy.
—JAN MORRIS
• • •
PART TWO
Hot 'N' Cold
I first visited Martinique at age 15, on a tropical interlude with my Swiss parents, and was instantly engulfed in the Caribbean breath.
—SYLVIE BIGAR
• • •
I looked forward to dusk as I had become fascinated by the ruby-colored soldier crabs that climbed from earthen holes to begin their nightly scavenge for food. In Brooklyn, stars were scarce, but in Barbados, the night sky was littered with them.
—BERNICE L. MCFADDEN
• • •
What was incredible about the Maldives was that the entire island we were on consisted of sand. There didn't seem to be any dirt. You could walk around for hours barefoot with your white trousers skimming the ground, and they'd still be pristine white.
—KT TUNSTALL
• • •
A tropical vacation is a place where fantasy and reality can meet.
—LAURA SUTHERLAND,
Tropical Family Vacations
• • •
Since I was 19, I've always gone where there was a reason to be. Maybe I'll be lucky and there'll be a reason to go somewhere tropical for a while.
—FEIST
• • •
A real friend is someone who takes a winter vacation on a sun-drenched beach and does not send a card.
—FARMER'S ALMANAC
• • •
On the beach, you can live in bliss.
—DENNIS WILSON, BEACH BOYS
• • •
The beach is in our blood. Everyone in our family returns to the beach instinctively, just like the sea turtles.
—SANDY ARCHIBALD
• • •
I love going to the beach in the tropics and doing whatever I do—surfing, swimming or being—and the glow when I get a tan that deepens. I walk around with red and gold in my skin and feel like the most beautiful thing on the planet!
—ANIKA NONI ROSE
• • •
My love for traveling to islands amounts to a pathological condition known as nesomania, an obsession with islands.
This craze seems reasonable to me, because islands are small self-contained worlds that can help us understand larger ones.
—PAUL THEROUX
• • •
I find being cut off in a ski resort away from the demands of normal living to be remarkably conducive to writing, so I always do some work on my current book.
—PETER JAMES
• • •
My family never went skiing. My dad was afraid of heights and my mom felt that a vacation was only a vacation if it involved reading at least two books on the beach.
—LISA GREENWALD,
Sweet Treats & Secret Crushes
• • •
They're such odd places, ski resorts—they tend to have more in common with each other than they do with the countries they're in.
—TIM DOWLING
• • •
The climate of Barrow is Arctic. Temperatures range from cold as shit to fucking freezing.
—STEVE NILES,
30 Days of Night
• • •
PART THREE
Exploration
Exploration is really the essence of the human spirit.
—FRANK BORMAN
• • •
Being an explorer is more than just geography. It's also an attitude. It's assuming that if something looks "boring," you're not looking hard enough. It's trying to see what's really in front of you, not what the world generally agrees upon.
—MIKE SOWDEN
• • •
For a journey by land, no less than by sea, was subject to the proper omens. To dream of quail presaged being tricked or meeting bandits en route, owls meant meetings storms or bandits, wild boars meant meeting storms. A gazelle foretold an easy or hard trip, depending on its physical condition.
—LIONEL CASSON,
Travel in the Ancient World
• • •
It suddenly struck me that that tiny pea, pretty and blue, was the Earth. I put up my thumb and shut one eye, and my thumb blotted out the planet Earth. I didn't feel like a giant. I felt very, very small.
—NEIL ARMSTRONG
• • •
I'm not trivializing Star Trek, because my hopes for the future is based on USS Enterprise, to boldly go where man's not been before.
—BUZZ ALDRIN
• • •
What everyone in the astronaut corps shares in common is not gender or ethnic background, but motivation, perseverance, and desire—the desire to participate in a voyage of discovery.
—ELLEN OCHOA
• • •
What mattered to the man [Douglas Mawson] instead—and what drove the vast ambitions of the AAE—was the urge to explore land that had never before been seen by human eyes, and to bring back from the southern continent the best science that men in the field might be capable of.
—DAVID ROBERTS,
Alone on the Ice
• • •
Nobody can discover the world for somebody else.
Only when we discover it for ourselves does it become common ground and a common bond and we cease to be alone.
—WENDELL BERRY
• • •
Roslyn and Aaron Ganz
Escort vessels with dignitaries, press, and Kane's own father had shadowed his departure up the East River and twenty miles out to sea. He was a new American explorer on a noble, world-renowned rescue mission.
—TODD BALF,
Farthest North
• • •
I feel that another expedition—unless it crosses the continent—is not much.
—ERNEST SHACKLETON
• • •
Michael Ganz
Our real discoveries come from chaos, from going to the place that looks wrong and stupid and foolish.
—CHUCK PALAHNIUK,
Invisible Monsters
• • •
He was, above all, an explorer in the classic mold—utterly self-reliant, romantic, and just a little swashbuckling.
—ALFRED LANSING,
Endurance
• • •
What an admirable position of the New World, that man has yet no enemies but himself.
—ALEXIS DE TOCQUEVILLE,
Democracy in America
• • •
What can we gain by sailing to the moon if we are not able to cross the abyss that separates us from ourselves? This is the most important of all voyages of discovery, and without it, all the rest are not only useless, but disastrous . . .
—THOMAS MERTON
• • •
The United States is a nation of explorers. America is the spirit of human exploration distilled.
—ELON MUSK
• • •
Polar exploration is at once the cleanest and most isolated way of having a bad time which has been devised. It is the only form of adventure in which you put on your clothes at Michaelmas and keep them on until Christmas, and, save for a layer of the natural grease of the body, find them as clean as though they were new.
—APSLEY CHERRY-GARRARD,
The Worst Journey in the World
• • •
When Christopher Columbus spotted manatees on his voyage of discovery in 1493, he thought these gentle marine giants were mermaids, though he reported they were "not half as beautiful as they are painted."
—JAY CLARKE,
Miami Herald
• • •
Coming back to your native land after an absence of many years is a surprisingly unsettling business, a little like waking from a long coma.
—BILL BRYSON,
I'm a Stranger Here Myself
• • •
By choosing to visit the wrong place at the wrong time, not only do I lose the crowds; I almost always experience my surroundings, especially familiar ones, in an unexpected way.
—DAISANN MCLANE,
National Geographic Traveler
• • •
Man cannot discover new oceans unless he has the courage to lose sight of the shore.
—ANDRE GIDE
• • •
We must go beyond textbooks, go out into the bypaths and untrodden depths of the wilderness and travel and explore and tell the world the glories of our journey.
—JOHN HOPE FRANKLIN
• • •
Michael Ganz
The sea is dangerous and its storms terrible, but these obstacles have never been sufficient reason to remain ashore . . . unlike the mediocre, intrepid spirits seek victory over those things that seem impossible . . . it is with an iron will that they embark on the most daring of all endeavors . . .
to meet the shadowy future without fear and conquer the unknown.
—FERDINAND MAGELLAN
• • •
The most spectacular places to view from space are all of the Hawaiian Islands, the mountain ranges and the Kamchatka peninsula in Russia. When you fly over it at night, it's like the Fourth of July because of all of the volcanoes erupting simultaneously.
—JON MCBRIDE, NASA
• • •
PART FOUR
Trains, Planes, and Automobiles
The act of flying itself is great, because you get time to unwind and get into a new frame of mind. You're not bothered, you're sitting there, they're feeding you, you have daylight all the way. You're going to a different world—emotionally, spiritually, and physically. If you work it out right, it can be just marvelous.
—ALEX TREBEK
• • •
Sometimes, flying feels too God-like to be attained by man. Sometimes, the world from above seems too beautiful, too wonderful, too distant for human eyes to see.
—CHARLES A. LINDBERGH,
The Spirit of St. Louis
• • •
It's a powerful thing: the airline window seat. Kids clamor for them. Shutterbugs are drawn to them. And some of us can't not look out the window.
—THOM PATTERSON, CNN
• • •
When I was a kid, and into early adulthood, I loved being on planes. I loved the sense that time was suspended, that I was in between, and that I would land somewhere and find everything strange—roads and faucets reconceived, language a jumble of sounds, the time of day out of whack.
—ELISABETH EAVES,
Wanderlust
• • •
Noting that, according to my ticket, we will land in San Diego barely two hours after we leave Seattle, I share with Mary the remarkable fact that it will take us five months to cover the same ground on our return hike north on the Pacific Crest Trail (PCT).
—LIEF CARLSON,
A Gang of One
• • •
Never be late for a plane with a girl. 'Cuz a girl runs like a girl—with the little steps and the arms flailing out . . . You wanna make this plane, you've gotta run like a man! Get your knees up!
—JERRY SEINFELD
• • •
I have a fantasy of air travel, and it goes like this: I cab to the airport with my pre-printed boarding pass in hand, go through an efficiently moving security line and get to the gate just as the flight is starting to board.
Clarification: I get to the gate just as the flight is starting to board my particular zone.
—JOE YONAN,
The Washington Post
• • •
If you are a parent who lets your kids scream and go nuts on a plane—congratulations—you top the list of most annoying etiquette violators in the air.
—HUGO MARTIN (based on a survey by expedia.com)
• • •
It was exciting to be off on a journey she had looked forward to for months. Oddly, the billowing diesel fumes of the airport did not smell like suffocating effluence, it assumed a peculiar pungent scent that morning, like the beginning of a new adventure, if an adventure could exude a fragrance.
—E. A. BUCCHIANERI,
Brushstrokes of a Gadfly
• • •
I will always be awed by the physics that gets a fat metal bird into the upper troposphere.
—JONAH LEHRER,
The Observer
• • •
I've seen passengers get naked, attempt to open an emergency door to get off the "bus," reach inside a first-class meal cart and eat leftover food from a dirty plate, and get hit on the head by luggage—then threaten to sue the airline because the injury had affected their psychic abilities.
—HEATHER POOLE,
Cruising Altitude
• • •
Travel, in other words, is a basic human desire.
We're a migratory species, even if our migrations are powered by jet fuel and Chicken McNuggets.
—JONAH LEHRER,
The Observer
• • •
Fly enough, and you learn to go brain-dead when you have to. It's sort of like time travel. One minute you're bending to unlace your shoes, and the next thing you know you're paying fourteen dollars for a fruit cup, wondering, How did I get here?
—DAVID SEDARIS,
Let's Explore Diabetes with Owls
• • •
Like a growing number of middle-aged American travelers, they were dressed in brightly colored sweat suits, brand new Nike athletic shoes and fanny packs that hung from their waists like decorative sashes. It wasn't clear whether they were preparing to fly to Paris or work as road monitors at the New York Marathon.
—ELLIOTT HESTER,
Plane Insanity
• • •
Jackals, neelgais [antelopes], monitor lizards, peacocks, porcupines, snakes, monkeys, foxes, dogs and birds of prey—the diversity of wildlife at Indira Gandhi International Airport can give any small town zoo a run for its money.
—HINDUSTAN TIMES
• • •
Our family has two vegetarians, one picky eater, and one person who gets hives thinking about airport food prices . . . Once, due to a late arrival, we had a tight connection in Dallas and had to eat Tic Tacs for dinner. Never again!
—JAMIE PEARSON,
TravelSavvyMom.com
• • •
You haven't seen a tree until you've seen its shadow from the sky.
—AMELIA EARHART
• • •
There is no sport equal to that which aviators enjoy while being carried through the air on great white wings.
—WILBUR WRIGHT
• • •
There are only two emotions in a plane: boredom and terror.
—ORSON WELLES
• • •
Airplane travel is nature's way of making you look like your passport photo.
—AL GORE,
The World According to Gore
• • •
I have, however, been in Sane. They don't have an airport; you have to be driven there. I have made several trips there, thanks to my children, friends, family, and work.
—GEORGE CARLIN
• • •
A ship in harbor is safe, but that is not what ships are built for.
—JOHN A. SHEDD
• • •
I am standing at a railing, looking out to sea with nothing but water and froth from me to the horizon . . . I feel as if I am on the very edge of the dangers of nature's wrath, should I be foolish enough to test them.
—DAVID MOLYNEAUX,
Miami Herald
• • •
Now that I am about to board the Rickmers Seoul freighter
(Chinese-built, German-managed, Marshall Islands-registered), being a passenger on a cargo ship seems a lot like being an inmate in a prison, except that on a ship you can't tunnel yourself out.
—PATRICIA MARX,
The New Yorker
• • •
"I love all this!" said Luke. "I really do. Can't help it. I love the machinery itself, you know—the inventiveness of it, the way it all differs from boat to boat. So much more fun than writing."
—REDMOND O'HANLON,
Trawler
• • •
The great difference between voyages rests not with the ships, but with the people you meet on them.
—AMELIA E. BARR
• • •
Land was created to provide a place for boats to visit.
—BROOKS ATKINSON
• • •
The goal is not to sail the boat, but rather to help the boat sail herself.
—JOHN ROUSMANIERE
• • •
Idyllic as it was to float the Colorado River between massive tapestried walls, I felt vaguely imprisoned by our rubber raft. More vivid were the times we worked our way up side canyons, or clambered slickrock to the rim. In retrospect, I realize I was far happier on rock than on water.
—MICHAEL ENGELHARD,
Hell's Half Mile
• • •
A sailor's joys are as simple as a child's.
—BERNARD MOITESSIER
• • •
At sea, I learned how little a person needs, not how much.
—ROBIN LEE GRAHAM
• • •
The lovely thing about cruising is that planning usually turns out to be of little use.
—DOM DEGNON
• • •
I quickly realized that everything about the bus was designed for discomfort. I was sitting beside the heater, so that while chill draughts teased my upper extremities, my left leg grew so hot that I could hear the hairs on it crackle. The seats were designed by a dwarf seeking revenge on full-sized people.
—BILL BRYSON,
Neither Here, Nor There
• • •
This hiss of metal on metal, the very slight swaying of the carriage, the feeling of being securely enclosed in a comfortable, well-lighted space while the world is flung by in glossy darkness outside, all this puts me in a mood to read, as if the material world had been suspended and I were entirely in the realm of the mind.
—TIM PARKS,
Italian Ways
• • •
I have seldom heard a train go by and not wished I was on it.
Those whistles sing bewitchment: railways are irresistible bazaars, snaking along perfectly level no matter what the landscape, improving your mood with speed, and never upsetting your drink.
—PAUL THEROUX,
The Great Railway Bazaar
• • •
That three-day ride atop the train, baking in the fierce sun, clinging for dear life as the hot iron chugged across the plains of India, was the unpromising start of my family's journey.
—RICHARD C. MORAIS,
The Hundred-Foot Journey
• • •
I was looking forward to one of those quiet, soothing journeys that only trains can provide. So it was with some dismay that I discovered that the seat behind me was occupied by Cellphone Man.
—BILL BRYSON,
Notes from a Small Island
• • •
Trains tap into some deep American collective memory.
—DANA FRANK
• • •
I like trains. I like their rhythm, and I like the freedom of being suspended between two places, all anxieties of purpose taken care of: for this moment I know where I am going.
—ANNA FUNDER,
Stasiland: Stories from Behind the Berlin Wall
• • •
It's not getting from A to B. It's not the beginning or the destination that counts. It's the ride in between . . . This train is alive with things that should be seen and heard. It's a living, breathing something—you just have to want to learn its rhythm.
—DAVID BALDACCI,
The Christmas Train
• • •
Bib Block was sure that in any part of the country at all, whenever the name of this road was mentioned, people's hearts pivoted like Moslems to the east and flopped over. Sooner or later, he believed, at one stage of the journey or another, all roads led to the New Jersey Turnpike.
—KATHRYN KRAMER,
A Handbook For Visitors From Outer Space
• • •
There's nothing quite like putting the whole family into the car and hitting the open road, leaving your worries behind, driving mile after carefree mile, sometimes getting as many as three carefree miles before everybody in the car hates everybody else and gunfire breaks out in the back seat.
—DAVE BARRY
• • •
Now, on this road trip, my mind seemed to uncrinkle, to breathe, to present to itself a cure for a disease it had not, until now, known it had.
—ELIZABETH BERG,
The Year of Pleasures
• • •
Seeing the world from a bike, you have a whole new appreciation than when you're being powered by something gasoline.
—KELLY SCHWAN
• • •
PART FIVE
Accommodations
The great advantage of a hotel is that it is a refuge from home life.
—GEORGE BERNARD SHAW
• • •
My definition of a good hotel is a place I'd stay at.
—ROBERT DE NIRO
• • •
Los Angeles is a large city-like area surrounding the Beverly Hills Hotel.
—FRAN LEBOWITZ
• • •
I've never stolen anything from a hotel room, I strictly observe the environmental towel signals, and I even leave the complimentary shampoo and conditioner where I find them unless I actually use them. I stop short of feeling guilty about leaving the bed unmade, but other than that I'd be proud to have me as a guest.
—PAM JANIS,
The Washington Post
• • •
I stayed in a really old hotel last night.
They sent me a wake-up letter.
—STEPHEN WRIGHT
• • •
I work in a hotel. I know what you're probably thinking, and no, I am not a hooker. Not unless you're not a cop.
—JAROD KINTZ,
If you bring the booze and food,
I'll bring the thirst and hunger
• • •
To put it rather bluntly, I am not the type who wants to go back to the land; I am the type who wants to go back to the hotel.
—FRAN LEBOWITZ
• • •
I do not think that when in a hotel you have to feel
"at home", on the contrary, you have to get the feeling that you are definitely elsewhere . . .
—AURELIO VAZQUEZ DURAN
• • •
I would rather sleep in a bathroom than in another hotel.
—BILLY WILDER
• • •
The past is a hotel. You can visit any time; enjoy the view. But you can't live there. The cost is too high.
—B.J. NEBLETT,
Elysian Dreams: Where the Past Meets the Present
• • •
I've always thought a hotel ought to offer optional small animals. I mean a cat to sleep on your bed at night, or a dog of some kind to act pleased when you come in. You ever notice how a hotel room feels so lifeless?
—ANNE TYLER
• • •
When I'm paying, I'll still take a $3 Snickers bar and charge it to the room . . . But if Paramount Pictures is paying, then I'll take three Snickers.
—CHRIS ROCK
• • •
A hotel room all to myself is my idea of a good time.
—CHELSEA HANDLER,
My Horizontal Life: A Collection of One-Night Stands
• • •
A hotel isn't like a home, but it's better than being a house guest.
—WILLIAM FEATHER
• • •
Room service? Send up a larger room.
—GROUCHO MARX,
A Night at the Opera
• • •
PART SIX
Traveling the Continents
EUROPE:
There's nowhere else like London. Nothing at all, anywhere.
—VIVIENNE WESTWOOD
• • •
JESSICA: When did you last go to London?
HUGH: When I read a Dickens novel. That was when I resolved never to travel there.
—BAUVARD,
The Prince Of Plungers
• • •
The emerald glimmers of the sea and the orchards;
the amber butterflies; the pearl and topaz gelati;
the timeworn cameos of the paint-box houses—all the inexhaustible treasures of the Cinque Terre.
—LIESL SCHILLINGER
• • •
When you're meant to be somewhere, everything in between feels like you're treading water, just waiting for that wave to lift you and carry you onto the shore of your new land. My new land was Italy.
—JENNIFER CRISWELL,
At Least You're in Tuscany
• • •
Most people in Europe know Malta. After all, it's a popular holiday destination boasting 7,000 years of history. But when I travel to the United States and tell people where I'm from, they sometimes say: "Is that Malta, Ohio?" There is a Malta in Ohio, by the way, but it's not as good as my one.
—JOSEPH CALLEJA
• • •
The way to Santiago de Compostela is one of the three great Christian pilgrims' trails—the others are the Via Francigena from Canterbury to Rome and the pilgrimage to Jerusalem from anywhere.
—HAPE KERKELING,
I'm Off Then
• • •
Crossing the Sierra was not just a stage on my journey, in spite of the physical barrier. It was also one of those sudden, jerky advances in life, which once made closes the past forever.
It was a frontier for me in more ways than one, and not till I'd passed it did I feel really involved in Spain.
—LAURIE LEE,
As I Walked Out One Midsummer Morning
• • •
Rarely in traveling do expectations and reality intersect, but they certainly do in Donostia [Spain].
—MICHAEL YAEGER
• • •
I had seen Spain's most technically proficient performers in concert in America. But the "duende" I yearned for was conspicuously absent.
—DIANE CALDWELL,
The Quest for Duende
• • •
You don't have to spend long in Seville to see why so many operas have been set there. A sense of drama pervades the Andalusian capital, from its Moorish royal palaces and extravagant Catholic festivals to the way the strum of a guitar tends to send a whole room into syncopated clapping.
—CHARLY WILDER,
The New York Times
• • •
After a lifetime of travel in Europe, Asia, and America, I am convinced that to be respectable is one of the pinnacles of universal human desire, felt as strongly in the heart of the Persian nomad as it is in New York, Chicago, Valparaiso, or Tooting Bec. And that London has very often known the art of concocting this subtle elixir.
—V. S. PRITCHETT,
London Perceived
• • •
Italy always has had a magnetic north pull on my psyche . . . My idea of heaven still is to drive the gravel farm roads of Umbria and Tuscany, very pleasantly lost.
—FRANCES MAYES,
Under the Tuscan Sun
• • •
I grew up watching vintage Italian movies, but it wasn't until my forties that I finally got to Rome. It was worth the wait.
—SIMON DOONAN
• • •
I'm not one for sightseeing, but wherever you are in Rome, there's something interesting around the corner. It's an amazing place, with little bars everywhere and coffee shops. It's a brilliant place to go with your wife.
—AUSTIN HEALY,
The Independent
• • •
I could find nothing to prevent my taking a sightseeing trip. I therefore motor-cycled up to the hill village of Capaccio . . . presiding with cool if distant charm over the raucous confusion below and representing for me all that was most romantic in the landscape of Southern Italy.
—NORMAN LEWIS,
Naples '44
• • •
There is, I think, something that sets France apart from many other parts of the world. I know of no other country that is so fascinating yet frustrating, so aware of the world and its place within it but at the same time utterly insular.
—SARAH TURNBULL,
Almost French
• • •
Paris is always a good idea.
—AUDREY HEPBURN
• • •
This is Paris—Paris! A place where in a setting of incredible beauty, romance simply cannot not happen. Take Psyche Revived by Cupid's Kiss, Antonio Canova's spectacular winged sculpture in the Louvre. Or any of Renoir's knockout nudes, plump and luscious, displayed in the Musée de l'Orangerie.
—COLETTE O'CONNOR,
Paris, With Parents
• • •
I had died and been delivered to a place of golden coins [Paris] that minted themselves by the tens of thousands from gods' mouths in fountains. All of the talk I heard, though I understood none of it, was wise and mythical and rare.
—RAY BRADBURY,
My Favorite Place on Earth
• • •
The beauty of Prague, where I was born, can be found in its ancient cobbled streets, towering cathedral spires, haunting statues, and architecture that ranges from Byzantine to bizarre. I love being an American, but each time I go back, I find that a piece of my heart has never left.
—MADELEINE ALBRIGHT
• • •
Bosnia had emerged from the 1990's war and rebuilt its infrastructure, but tourists had yet to discover the country.
It was time to visit.
—JAN BURAK SCHWERT,
The Tiny Red Kettle
• • •
LATIN AMERICA
The most important thing Paris gave me was a perspective on Latin America. It taught me the differences between Latin America and Europe and among the Latin American countries themselves through the Latins I met there.
—GABRIEL GARCIA MARQUEZ
• • •
I decided to go to Latin America because many of my students in Washington emigrated from this region and inspired me to learn more about their home countries.
—JENNA BUSH
• • •
The remote Osa Peninsula, which juts into the Pacific Ocean from Costa Rica's southwestern corner, seemed to hold an increasingly rare chance to observe the rain forest in all its fecund, carbon-storing, oxygen-producing glory, without quite so much human company.
—AMY HARMON,
The New York Times
• • •
ASIA:
The first word I heard when I stepped off the plane at Ataturk Airport in Istanbul1 was buyurun—welcome—a word often spoken in my home in New York City where it was always open house, people dropping in day or night for a meal. Looking at these Turkish faces, I feel as though I'm surrounded by relatives.
—GLORIA KIRCHHEIMER,
Out of Smyrna
• • •
For me the East had begun weeks before. Even in Greece I had felt Europe falling away. There was the East in the food, the emphasis on sweets . . . in the posters for Indian films with the actress Nargis, a favourite, I was told, of Greek audiences; in the instantaneous friendships, the invitations to meals and homes.
—V. S. NAIPAUL
• • •
I'm just over a week into my physical travels across northern India, but a dozen weeks into the personal journey that began with giving notice on a 20-year career, and planning my solo wander across a land that has existed in my mind as a magical and challenging destination, no more or less real than Narnia or Brigadoon.
—BILL GIEBLER,
The Tea in Me
• • •
Not many people can say they really visited what Guangzhou, China, has to offer without drinking an elixir of snake's blood. Sure, it may be unpleasant, but you'll impress the locals and will likely never forget the experience.
—HARRY MARKS
• • •
When I first visited Beijing, I enjoyed it so much that I remained for five months, exploring the old and new parts by foot, bicycle, and rickshaw. The Chinese capital stimulates the brain and enlarges the mind.
—GAY TALESE
• • •
The transition from countryside to city was abrupt: one moment we saw rice paddies, vegetable patches and fishponds, and several hundred yards later the train came to a halt in Changsa, a city of more than one million people and the capital of Hunan Province.
—MARK SALZMAN,
Iron and Silk
• • •
The most gleamingly modern airports I've been to are in China. The coolest and most technologically challenging hotel I've ever stayed in is in China. The most tricked-out hair salon I've been to is in China. The worst pollution I've endured is in China.
—AMY TAN
• • •
One might as well go blindfold as travel through a country like Asia Minor without preparing oneself with some part of this rewarding drudgery to make the landscape quicken in one's sight.
—FREYA STARK,
The Lycian Shore
• • •
My next expedition, crossing Delhi's main Janpath Road, may not seem epic, but as I stood on its curb with traffic flowing in six lanes like the Ganges in flood, the other side seemed unreachable.
—KATE CRAWFORD
• • •
I could've spent years at the Grand Bazaar, in Istanbul, sifting through rugs and kilims, the amazing baubles at Sehrazat, and the all-natural bath products at Abdulla. The trip informed my entire Resort 2013 collection. Everything I saw—the jewel-like colors; the architecture of the Old City; the handcrafted decorative details—added up to a dazzling inspiration board.
—MICHAEL KORS
• • •
MIDDLE EAST:
Their migrations were not always dictated by the need to find water or better grazing for their herds; sometimes they struck their tents and began to move for no discernible reason . . . they were animated by an impulse, perhaps lodged in their nomadic genes, to get going, it didn't matter where.
—PHILIP CAPUTO,
on the Bedouin tribesmen, The Longest Road
• • •
NORTH AMERICA:
The U.S. is a microcosm of nearly every world culture, climate, landscape and category of wildlife. (And whatever doesn't occur naturally gets recreated at Disney.)
—JORDAN BURCHETTE,
CNN
• • •
There's nothing American tourists like more than the things they can get at home.
—STEPHEN COLBERT
• • •
Our holidays are generally here in the United States because there are so many places to go and see in our own country that we can't see much reason to go anywhere else.
—JON MCBRIDE, NASA
• • •
Michael Ganz
Overlooked during the westward expansion of the American frontier in the 1800s, Yellowstone was made the world's first national park the way you might give the last kid picked for kickball the top spot in the order.
—JORDAN BURCHETTE,
CNN
• • •
My heart lives in New York, where I was born and raised. My body lives in Los Angeles. But my soul lives in Montana.
—HENRY WINKLER
• • •
Michael Ganz
I inhale the mixture of humidity and magnolias, and it's like Proust biting into the madeleine.
—WALTER ISAACSON,
on New Orleans
• • •
Despite the dangers, for a man of Powell's character the temptation was irresistible. Perhaps nowhere on earth were science and adventure as intertwined as in the American Southwest.
—EDWARD DOLNICK,
Down the Great Unknown
• • •
To arrive in the Rocky Mountains by plane would be to see them in one kind of context, as pretty scenery. But to arrive after days of hard travel across the prairies would be to see them in another way, as a goal, a promised land.
—ROBERT M. PIRSIG,
Zen and the Art of Motorcycle Maintenance
• • •
I've been to a lot of great places, but New York is the best.
—CHRIS ROCK
• • •
When I'm in Canada, I feel this is what the world should be like.
—JANE FONDA
• • •
I read and learned and fretted more about Canada after I left than I ever did while I was home. I absorbed anything I could on topics that ranged from folklore to history to political manifestos . . . I ranted and raved and seethed about things beyond my control. In short I acted like a Canadian.
—WILL FERGUSON
• • •
Canada is an interesting place; the rest of the world thinks so, even if Canadians don't.
—TERENCE M. GREEN
• • •
Mazatlán [Mexico] was a promising paradise: soft sand, warm water and a sea rich with shrimp and tuna. Turn around, though, and there was the graffiti-tagged house, formerly owned by drug cartel capos, followed by others.
—DAMIEN CAVE,
The New York Times
• • •
We stood on the border of Mexico, one foot in one country, one in the other. I felt both the fear of the unknown and the lure of inner knowing.
—GAIL D. STOREY,
I Promise Not to Suffer
• • •
My father would have been thrilled that I've come to Mexico to mourn him; he loved Latin American and Spanish culture. He collected Day of the Dead statues, Tarahumara pottery, and Mexican postcards of 1930s film stars; he devoured everything he could find to read about pre-Colombian history, the Mayans, the mummies of Guanajuato. But mostly he loved the music.
—LAVINIA SPALDING,
The Ghosts of Alamos
• • •
AUSTRALIA:
Each time you fly from North America to Australia, and without anyone asking how you feel about it, a day is taken away from you when you cross the international date line.
—BILL BRYSON,
In a Sunburned Country
• • •
Of everywhere I've been on my bike, the one I'd suggest to anyone looking for the ideal place to go cycling would be Melbourne. I adore Australia. I love hanging out on a beach holiday for a week or two—but when you want things to do and see, it's hard to beat Melbourne for having pretty much everything.
—SIR CHRIS HOY
• • •
I always had a wanderlust for traveling and I wanted to take a year off to go take an adventure, and it was.
—MATTHEW MCCONAUGHEY,
on his exchange program in Australia
• • •
Australia has always been a big melting pot of cultures and I love that about it. Sydney is like the pretty older sister that everyone looks at and goes, "Wow!" But Melbourne is amazing. It's more of an art city and has an almost ethnic feel.
—RICK SPRINGFIELD
• • •
AFRICA:
The more travelled roads were good and often paved for a short distance, but once the pavement ended, an aeroplane, if one were at hand, could save hours of weary toil behind the wheel of a lurching car—provided the driver were skilful enough to keep it lurching at all.
—BERYL MARKHAM,
West with the Night
• • •
Climbing Kilimanjaro is one of the simplest and most difficult things you can do with five days.
—JOSH NOEL,
Chicago Tribune
• • •
Cape Town is the only place I know where I can indulge almost all of my interests simultaneously. It has beaches on both the Atlantic and Indian oceans, several indigenous cuisines, and a major wine region inside the city limits: Constantia.
—JAY MCINERNEY
• • •
The window of Africa, like the window on a train rushing through the night, is a distorting mirror that partly reflects the viewer's own face.
—PAUL THEROUX,
The Last Train to Zona Verde
• • •
I like places that are dusty, far-flung and mysterious. I went to Nepal to hike and meditate. Nepal was a little easier for me to get around because it isn't as high up as Tibet and it is more touristy.
—MARGARET CHO
• • •
To be alone in nearly 300,000 acres of pristine wilderness with young Masai warriors guiding you is really a peak experience.
—EDWARD NORTON
• • •
CARIBBEAN
In many ways, Jacmel—along with the rest of Haiti—is a lot like that less flashy gentleman caller, who demands a closer look but is guaranteed to steal your heart.
—EDWIDGE DANTICAT
• • •
Each time I travel up that miraculous road [in Trinidad], I press my face into the wind, I drink in the salt-filled air, I take snapshots in my mind's eye of the trees, the flowers, the roaring ocean, and I am prepared once more for wintry days in the land I now call home.
—ELIZABETH NUNEZ
• • •
Nose, eyes, ears, tongue, fingers, toes—an appreciation of the Caribbean involves all your sensory equipment. It also demands something of you: the willingness to see, hear, feel, touch and smell the unpretty things. Only then can you fully embrace the islands' beauty.
—JERRY V. HAINES
• • •
1 Istanbul is part of Asia and Europe.
INDEX
A
Douglas Adams
Leanne Adams
Clay Aiken
Madeleine Albright
Buzz Aldrin
Robert Thomas Allen
Maya Angelou
Sandy Archibald
Neil Armstrong
Marcello Arrambide
Brooks Atkinson
David Attenborough
Saint Augustine ,
B
Francis Bacon
Jennifer Baggett
David Baldacci
James Baldwin
Todd Balf
Amelia E. Barr
Dave Barry ,
Bauvard
Miriam Beard
Sir Thomas Beecham
Hilaire Belloc
Marc Bekoff
Stacie Nevadomski Berdan
Elizabeth Berg
Yogi Berra
Wendell Berry
Diccon Bewes ,
Sylvie Bigar
Tom Bisell
Raymond Blanc
Marybeth Bond
Frank Borman
Anthony Bourdain ,
Dave Bouskill
Ray Bradbury ,
Dominique Browning ,
Samantha Brown
Bill Bryson , , ,
Martin Buber
E. A. Bucchianeri
Jordan Burchette ,
Jenna Bush
C
Tim Cahill
Diane Caldwell
Joseph Calleja
Albert Camus
Philip Caputo
George Carlin
Lief Carlson
Alan Carr
Lionel Casson
Damien Cave
William D. Chalmers
Amit Chaudhuri
Apsley Cherry-Garrard
G. K. Chesterton ,
Julia Child
John Cho
Margaret Cho
Jay Clarke
Christopher Coffman
Stephen Colbert
Faith Conlon
Pat Conroy
Mason Cooley
Anderson Cooper
Deb Corbeil
Kimber Crandall
Kate Crawford
Michael Crichton ,
Jennifer Criswell
D
Edwidge Danticat
Dom Degnon
Robert De Niro
Antoine de Saint-Exupery
René Descartes
Alexis de Tocqueville
Benjamin Disraeli
Edward Dolnick
Simon Doonan
Tim Dowling
Aurelio Vazquez Duran
Lord Dunsany
Lawrence Durrell
E
Amelia Earhart
Merridy Eastman
Elisabeth Eaves
Ingrid Emerick
Ralph Waldo Emerson ,
Michael Engelhard
Chris Erskine ,
Andrew Evans
F
Clifton Fadiman
Gustave Falubert
Farmer's Almanac
William Feather
Judith Fein
Feist
Will Ferguson
Morris Fishbein
M. F. K. Fisher
Claire Fontaine and Mia Fontaine
Jane Fonda
Keith Foskett
Dana Frank
John Hope Franklin
Rachel Friedman
Pauline Frommer
Robert Frost
Anna Funder
G
Emile Ganest
Martha Gellhorn
Rita Golden Gelman
Andre Gide
Daranna Gidel
Bill Giebler
Elizabeth Gilbert ,
Martin Glaser
Jeff Goins ,
Oliver Goldsmith
Jane Goodall
Allan E. Goodman
Al Gore
Robin Lee Graham
Terence M. Green
Lisa Greenwald
Carolyn Gregoire
Stephanie Elizondo Griest
H
Jerry V. Haines
Chelsea Handler
Amy Harmon
Sami Hassanyeh
Austin Healy
William Least Heat-Moon ,
Audrey Hepburn
Elliott Hester
Katia Hetter
Spud Hilton
Peter Hoeg
Babs Hoffman
Catherine Ryan Howard
Aldous Huxley ,
I
Walter Isaacson
Amber Israelsen
Pico Iyer ,
J
Peter James
Pam Janis
Angelina Jolie ,
Samuel Johnson ,
K
Matthew Kepnes
Hape Kerkeling
Jack Kerouac ,
Jarod Kintz
Rudyard Kipling
Gloria Kirchheimer
Michael Kors
Kathryn Kramer
Jon Krakauer
Gabriela Kruschewsky ,
Charles Kuralt
L
Dalai Lama
Alfred Lansing
Fran Lebowitz ,
Laurie Lee
Jonah Lehrer ,
Kathy Lette
Norman Lewis ,
Charles A. Lindbergh
Peter Jon Lindberg
M
Ferdinand Magellan
Rosemary Mahoney
Nelson Mandela
Beryl Markham
Harry Marks
Gabriel Garcia Marquez
Hugo Martin
Lynne Martin
Groucho Marx
Patricia Marx
Rosalind Massow
Frances Mayes
Jon McBride ,
Andrew McCarthy
Matthew McConaughey
Bernice L. McFadden
Jay McInerney
Mike McIntyre
Daisann McLane
Margaret Mead
Anthony Melchiorri
Herman Melville
Lisa Mercer
Thomas Merton
George Mikes
Henry Miller
Nick Miller
Mohammed
Bernard Moitessier
David Molyneaux
Tim Moore
Moorish proverb
Richard C. Morais
Holly Morris
Jan Morris
Wolfgang Amadeus Mozart
Mark Murphy
Elon Musk
Jan Myrdal
N
V. S. Naipaul
B. J. Neblett
Steve Niles
Josh Noel
Edward Norton
Elizabeth Nunez
O
Ellen Ochoa
Colette O'Connor
Erin O'Connor
Redmond O'Hanlon
James Oseland
P
Chuck Palahniuk
Michael Palin ,
Tim Parks
Thom Patterson
Cesare Pavese
Jamie Pearson
Kathrine Palmer Peterson
Robert M. Pirsig
Heather Poole
Rolf Potts
Terry Pratchett
V. S. Pritchett
Marcel Proust
R
Jonathan Raban , ,
Mary Anne Radmacher
Eric Raff
David Roberts
Chris Rock ,
Jim Rohn
Henry Rollins
Andy Rooney
Anika Noni Rose
John Rousmaniere
Rumi
Dagobert D. Runes
S
Moslih Eddin Saadi
Mark Salzman
Jose Saramago
Liesl Schillinger
Patricia Schultz , ,
Kelly Schwan
Jan Burak Schwert
David Sedaris
Jerry Seinfeld ,
Ernest Shackleton
George Bernard Shaw
John A. Shedd
Joshua Slocum
Penny Smith
Susan Sontag
Mike Sowden ,
Lavinia Spalding
Rick Springfield
Freya Stark
Lisa St. Aubin de Teran
Alice Steinbach ,
John Steinbeck ,
Robert Louis Stevenson ,
Rick Steves ,
Caskie Stinnett
Gail D. Storey
Cheryl Strayed
Laura Sutherland
Anne Sophie Swetchine
Christopher Sykes
T
Gay Talese
Amy Tan
Paul Theroux , ,
Henry David Thoreau
J. R. R. Tolkien ,
Amor Towles
Alex Trebek
Ivanka Trump
KT Tunstall
Sarah Turnbull
Mark Twain , ,
Anne Tyler
Lao Tzu
V
Thorstein Veblen
Kurt Vonnegut
Sarah Vowell
W
Izaak Walton
Terry Ward
Charles Dudley Warner
Alec Waugh
Evelyn Waugh
Eric Weiner ,
Orson Welles
Eudora Welty
Vivienne Westwood
Walt Whitman
Oscar Wilde
Billy Wilder
Charly Wilder
Pat Williams
Tennessee Williams
Dennis Wilson
Earl Wilson
Henry Winkler
Thomas Wolfe
Stephen Wright
Steven Wright
Wilbur Wright
Y
Michael Yaeger
Joe Yonan
Lin Yutang
Z
Andrew Zimmern
|
{
"pile_set_name": "Books3"
}
|
Tim Burtonâs 3D film Alice In Wonderland has been banned by Odeon in a dispute over the length of its cinema run. Here’s our pick of other movies sent to the cinematic naughty step, from Avatar in China to The Simpsons in Burmaâ¦
Avatar: 2D version was banned in China Avatar – China
The 2D version of James Cameronâs sci-fi epic Avatar was axed in cinemas across China after authorities claimed that the plot mirrors forced land evictions in the country.
The box office hit is set around a ruthless corporation forcing the native Na’vi from their homes in order to mine natural resources from the planet.
A columnist in the China Daily newspaper wrote: “All the forced removal of old neighbourhoods in China makes us the only earthlings today who can really feel the pain of the Na’vi.”
The Simpsons Movie – Burma The Simpsons Movie: Too yellow for Burma
Matt Groeningâs Simpsons Movie never really stood a chance in Burma, because their notoriously intransigent Motion Picture & Video Censor Board issued the bizarre edict that the colours yellow and red are banned in movies.
And thatâs a clear problem for the popular Springfield family.
Zargana, a 40-something year-old comedian-turned-director from Burma, said the nutty censorship board might issue one ruling in the morning, then completely change its mind and revoke it in the afternoon. “They never explain why,” he said.
Borat: Cultural Learnings of America for Make Benefit Glorious Nation of Kazakhstan – Russia
First of all, whatâs amazing is that Sacha Baron Cohen âs send-up of life in Kazakhstan didnât annoy their film censors at all. But it was a different matter in Russia. They slammed it as âoffensiveâ.
Yuri Vasyuchkov, head of Moscowâs film licensing department, said: âThere are moments in the film that could offend some viewersâ religious or national sensibilities.â
Monty Pythonâs The Life Of Brian – Wales
Monty Python’s Life of Brian: Aberystwyth council thought the Pythons were very naughty boys
Remarkably, the Python crewâs uproariously funny parody of organised religion was banned by Aberystwyth council in Wales until 2009 â until cast member Sue Jones-Davies, who plays Judith, became mayor!
It was also met with the cold shoulder in snowy Norway, who banned it for one year. The Swedes, amusingly, then used that fact as a marketing ploy, saying âitâs the film that is so funny that it was banned in Norway!â
All Quiet On The Western Front – Germany
This 1930s classic may have won an Academy Award, but its anti-war theme didnât go down well at all with Germanyâs Nazi party, who released rats into cinemas to scare off movie-goers.
Pineapple Express – Malaysia
The comedy starring Seth Rogen as a marijuana-smoking court clerk sparked a stern response from the ultra-strict Film Censorship Board of Malaysia. Itâs a devout Islamic nation and so the movieâs reference to drug use as âGodâs vaginaâ raised their hackles somewhat. So this Golden Globe-nominated stoner flick was quickly stubbed out.
Zoolander – Iran Ben Stiller’s Zoolander: Too camp for Iran
Any film that is seen to promote homosexuality, gay rights, or anti-islamic attitudes is banned outright in Iran. So Ben Stillerâs 2001 comedy Zoolander bit the dust. There are no outright gay jokes in it, but it was just too camp for the Ayatollah.
Indiana Jones and the Temple of Doom – India
They went ape over the blockbuster Steven Spielberg movie in India, because Indiana Jones and his co-horts are served monkey brains in one scene. And Hindus, of which there are many in the highly religious country, consider the creatures to be sacred. The ban was temporary, though.
The Matrix Reloaded – Egypt
Egyptâs censorship body didnât have any issues with the first Matrix film, but decided that the 2003 sequel wasnât really its cup of tea, owing to violent scenes that had the potential to “harm social peaceâ.
Its religious themes riled them, too. In a statement, the men in charge of banning stuff said: “Despite the high technology and fabulous effects of the movie, it explicitly handles the issue of existence and creation, which are related to the three divine religions, which we all respect and believe in.”
Catch 22 – Portugal
The Portuguese movie authorities prevented this one hitting the silver screen for four years because of a scene showing Alan Arkin, who played Capt. John Yossarian, sitting naked in a tree.
Theyâre barking, surely.
|
{
"pile_set_name": "OpenWebText2"
}
|
Follow @nbcbayarea !function(d,s,id){var js,fjs=d.getElementsByTagName(s)[0],p=/^http:/.test(d.location)?'http':'https';if(!d.getElementById(id)){js=d.createElement(s);js.id=id;js.src=p+'://platform.twitter.com/widgets.js';fjs.parentNode.insertBefore(js,fjs);}}(document, 'script', 'twitter-wjs');
What's cooler than a Girl Scout selling cookies?
A business-savvy Girl Scout who sets up shop outside a cannabis clinic to rev up her sales skills.
Thirteen-year-old Girl Scout Danielle Lei did brisk business last Monday selling Dulce de Leches and other flavors outside The Green Cross medical marijuana clinic in San Francisco, Mashable reported.
Danielle was able to sell 117 boxes within two hours outside the cannabis clinic -- 37 more than what she sold within the same time frame at a local Safeway the next day, according to Mashable.
"It's no secret that cannabis is an appetite stimulant -- so it's not shocking that a lot of our patients came and purchased cannabis, and then saw the cookies and purchased them," said Holli Bert, a spokesperson for The Green Cross. "But it wasn't just patients, staff members and neighbors also bought the cookies. I personally bought five boxes. It turned out to be a big success."
Bert said that Danielle's mother had contacted The Green Cross to set up the sale outside the store.
"We were happy to have her (Danielle) come -- she is extremely business savvy," Bert said.
Danielle's mother Carol takes her two daughters to different places around San Francisco to sell cookies so that "they can learn about different environments," Mashable reported.
Carol told NBC Bay Area that the family is not doing any interviews.
Green Cross is so happy with the outcome that they have invited Danielle back to their store on Saturday, Feb. 22, from 4 to 6 p.m. for a repeat performance.
The Green Cross posted about Danielle's visit on its Facebook page, including a link to an external poll, which asks people whether Girl Scouts should sell cookies outside legal marijuana shops.
The Girl Scouts of Northern California don't have a problem with this business strategy.
Dana Allen, the organization's marketing and communications director, told Mashable that "the mom decided this was a place she was comfortable with her daughter being at."
"We're not telling people where they can and can't go if it's a legitimate business," she said.
|
{
"pile_set_name": "OpenWebText2"
}
|
/*
* Copyright (C) 2012 The Guava Authors
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package com.google.common.io;
import static com.google.common.base.Preconditions.checkArgument;
import static com.google.common.io.TestOption.AVAILABLE_ALWAYS_ZERO;
import static com.google.common.io.TestOption.CLOSE_THROWS;
import static com.google.common.io.TestOption.OPEN_THROWS;
import static com.google.common.io.TestOption.READ_THROWS;
import static com.google.common.io.TestOption.SKIP_THROWS;
import static com.google.common.io.TestOption.WRITE_THROWS;
import static org.junit.Assert.assertArrayEquals;
import com.google.common.base.Charsets;
import com.google.common.collect.ImmutableList;
import com.google.common.collect.ImmutableSet;
import com.google.common.collect.Iterables;
import com.google.common.hash.Hashing;
import com.google.common.primitives.UnsignedBytes;
import com.google.common.testing.TestLogHandler;
import java.io.ByteArrayOutputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.OutputStream;
import java.util.Arrays;
import java.util.EnumSet;
import junit.framework.TestSuite;
/**
* Tests for the default implementations of {@code ByteSource} methods.
*
* @author Colin Decker
*/
public class ByteSourceTest extends IoTestCase {
@AndroidIncompatible // Android doesn't understand suites whose tests lack default constructors.
public static TestSuite suite() {
TestSuite suite = new TestSuite();
for (boolean asCharSource : new boolean[] {false, true}) {
suite.addTest(
ByteSourceTester.tests(
"ByteSource.wrap[byte[]]",
SourceSinkFactories.byteArraySourceFactory(),
asCharSource));
suite.addTest(
ByteSourceTester.tests(
"ByteSource.empty[]", SourceSinkFactories.emptyByteSourceFactory(), asCharSource));
}
suite.addTestSuite(ByteSourceTest.class);
return suite;
}
private static final byte[] bytes = newPreFilledByteArray(10000);
private TestByteSource source;
@Override
protected void setUp() throws Exception {
source = new TestByteSource(bytes);
}
public void testOpenBufferedStream() throws IOException {
InputStream in = source.openBufferedStream();
assertTrue(source.wasStreamOpened());
assertFalse(source.wasStreamClosed());
ByteArrayOutputStream out = new ByteArrayOutputStream();
ByteStreams.copy(in, out);
in.close();
out.close();
assertTrue(source.wasStreamClosed());
assertArrayEquals(bytes, out.toByteArray());
}
public void testSize() throws IOException {
assertEquals(bytes.length, source.size());
assertTrue(source.wasStreamOpened() && source.wasStreamClosed());
// test that we can get the size even if skip() isn't supported
assertEquals(bytes.length, new TestByteSource(bytes, SKIP_THROWS).size());
// test that we can get the size even if available() always returns zero
assertEquals(bytes.length, new TestByteSource(bytes, AVAILABLE_ALWAYS_ZERO).size());
}
public void testCopyTo_outputStream() throws IOException {
ByteArrayOutputStream out = new ByteArrayOutputStream();
assertEquals(bytes.length, source.copyTo(out));
assertTrue(source.wasStreamOpened() && source.wasStreamClosed());
assertArrayEquals(bytes, out.toByteArray());
}
public void testCopyTo_byteSink() throws IOException {
TestByteSink sink = new TestByteSink();
assertFalse(sink.wasStreamOpened() || sink.wasStreamClosed());
assertEquals(bytes.length, source.copyTo(sink));
assertTrue(source.wasStreamOpened() && source.wasStreamClosed());
assertTrue(sink.wasStreamOpened() && sink.wasStreamClosed());
assertArrayEquals(bytes, sink.getBytes());
}
public void testRead_toArray() throws IOException {
assertArrayEquals(bytes, source.read());
assertTrue(source.wasStreamOpened() && source.wasStreamClosed());
}
public void testRead_withProcessor() throws IOException {
final byte[] processedBytes = new byte[bytes.length];
ByteProcessor<byte[]> processor =
new ByteProcessor<byte[]>() {
int pos;
@Override
public boolean processBytes(byte[] buf, int off, int len) throws IOException {
System.arraycopy(buf, off, processedBytes, pos, len);
pos += len;
return true;
}
@Override
public byte[] getResult() {
return processedBytes;
}
};
source.read(processor);
assertTrue(source.wasStreamOpened() && source.wasStreamClosed());
assertArrayEquals(bytes, processedBytes);
}
public void testRead_withProcessor_stopsOnFalse() throws IOException {
ByteProcessor<Void> processor =
new ByteProcessor<Void>() {
boolean firstCall = true;
@Override
public boolean processBytes(byte[] buf, int off, int len) throws IOException {
assertTrue("consume() called twice", firstCall);
firstCall = false;
return false;
}
@Override
public Void getResult() {
return null;
}
};
source.read(processor);
assertTrue(source.wasStreamOpened() && source.wasStreamClosed());
}
public void testHash() throws IOException {
ByteSource byteSource = new TestByteSource("hamburger\n".getBytes(Charsets.US_ASCII));
// Pasted this expected string from `echo hamburger | md5sum`
assertEquals("cfa0c5002275c90508338a5cdb2a9781", byteSource.hash(Hashing.md5()).toString());
}
public void testContentEquals() throws IOException {
assertTrue(source.contentEquals(source));
assertTrue(source.wasStreamOpened() && source.wasStreamClosed());
ByteSource equalSource = new TestByteSource(bytes);
assertTrue(source.contentEquals(equalSource));
assertTrue(new TestByteSource(bytes).contentEquals(source));
ByteSource fewerBytes = new TestByteSource(newPreFilledByteArray(bytes.length / 2));
assertFalse(source.contentEquals(fewerBytes));
byte[] copy = bytes.clone();
copy[9876] = 1;
ByteSource oneByteOff = new TestByteSource(copy);
assertFalse(source.contentEquals(oneByteOff));
}
public void testSlice() throws IOException {
// Test preconditions
try {
source.slice(-1, 10);
fail();
} catch (IllegalArgumentException expected) {
}
try {
source.slice(0, -1);
fail();
} catch (IllegalArgumentException expected) {
}
assertCorrectSlice(0, 0, 0, 0);
assertCorrectSlice(0, 0, 1, 0);
assertCorrectSlice(100, 0, 10, 10);
assertCorrectSlice(100, 0, 100, 100);
assertCorrectSlice(100, 5, 10, 10);
assertCorrectSlice(100, 5, 100, 95);
assertCorrectSlice(100, 100, 0, 0);
assertCorrectSlice(100, 100, 10, 0);
assertCorrectSlice(100, 101, 10, 0);
}
/**
* Tests that the default slice() behavior is correct when the source is sliced starting at an
* offset that is greater than the current length of the source, a stream is then opened to that
* source, and finally additional bytes are appended to the source before the stream is read.
*
* <p>Without special handling, it's possible to have reads of the open stream start <i>before</i>
* the offset at which the slice is supposed to start.
*/
// TODO(cgdecker): Maybe add a test for this to ByteSourceTester
public void testSlice_appendingAfterSlicing() throws IOException {
// Source of length 5
AppendableByteSource source = new AppendableByteSource(newPreFilledByteArray(5));
// Slice it starting at offset 10.
ByteSource slice = source.slice(10, 5);
// Open a stream to the slice.
InputStream in = slice.openStream();
// Append 10 more bytes to the source.
source.append(newPreFilledByteArray(5, 10));
// The stream reports no bytes... importantly, it doesn't read the byte at index 5 when it
// should be reading the byte at index 10.
// We could use a custom InputStream instead to make the read start at index 10, but since this
// is a racy situation anyway, this behavior seems reasonable.
assertEquals(-1, in.read());
}
private static class AppendableByteSource extends ByteSource {
private byte[] bytes;
public AppendableByteSource(byte[] initialBytes) {
this.bytes = initialBytes.clone();
}
@Override
public InputStream openStream() {
return new In();
}
public void append(byte[] b) {
byte[] newBytes = Arrays.copyOf(bytes, bytes.length + b.length);
System.arraycopy(b, 0, newBytes, bytes.length, b.length);
bytes = newBytes;
}
private class In extends InputStream {
private int pos;
@Override
public int read() throws IOException {
byte[] b = new byte[1];
return read(b) == -1 ? -1 : UnsignedBytes.toInt(b[0]);
}
@Override
public int read(byte[] b, int off, int len) {
if (pos >= bytes.length) {
return -1;
}
int lenToRead = Math.min(len, bytes.length - pos);
System.arraycopy(bytes, pos, b, off, lenToRead);
pos += lenToRead;
return lenToRead;
}
}
}
/**
* @param input the size of the input source
* @param offset the first argument to {@link ByteSource#slice}
* @param length the second argument to {@link ByteSource#slice}
* @param expectRead the number of bytes we expect to read
*/
private static void assertCorrectSlice(int input, int offset, long length, int expectRead)
throws IOException {
checkArgument(expectRead == (int) Math.max(0, Math.min(input, offset + length) - offset));
byte[] expected = newPreFilledByteArray(offset, expectRead);
ByteSource source = new TestByteSource(newPreFilledByteArray(input));
ByteSource slice = source.slice(offset, length);
assertArrayEquals(expected, slice.read());
}
public void testCopyToStream_doesNotCloseThatStream() throws IOException {
TestOutputStream out = new TestOutputStream(ByteStreams.nullOutputStream());
assertFalse(out.closed());
source.copyTo(out);
assertFalse(out.closed());
}
public void testClosesOnErrors_copyingToByteSinkThatThrows() {
for (TestOption option : EnumSet.of(OPEN_THROWS, WRITE_THROWS, CLOSE_THROWS)) {
TestByteSource okSource = new TestByteSource(bytes);
try {
okSource.copyTo(new TestByteSink(option));
fail();
} catch (IOException expected) {
}
// ensure stream was closed IF it was opened (depends on implementation whether or not it's
// opened at all if sink.newOutputStream() throws).
assertTrue(
"stream not closed when copying to sink with option: " + option,
!okSource.wasStreamOpened() || okSource.wasStreamClosed());
}
}
public void testClosesOnErrors_whenReadThrows() {
TestByteSource failSource = new TestByteSource(bytes, READ_THROWS);
try {
failSource.copyTo(new TestByteSink());
fail();
} catch (IOException expected) {
}
assertTrue(failSource.wasStreamClosed());
}
public void testClosesOnErrors_copyingToOutputStreamThatThrows() {
TestByteSource okSource = new TestByteSource(bytes);
try {
OutputStream out = new TestOutputStream(ByteStreams.nullOutputStream(), WRITE_THROWS);
okSource.copyTo(out);
fail();
} catch (IOException expected) {
}
assertTrue(okSource.wasStreamClosed());
}
public void testConcat() throws IOException {
ByteSource b1 = ByteSource.wrap(new byte[] {0, 1, 2, 3});
ByteSource b2 = ByteSource.wrap(new byte[0]);
ByteSource b3 = ByteSource.wrap(new byte[] {4, 5});
byte[] expected = {0, 1, 2, 3, 4, 5};
assertArrayEquals(expected, ByteSource.concat(ImmutableList.of(b1, b2, b3)).read());
assertArrayEquals(expected, ByteSource.concat(b1, b2, b3).read());
assertArrayEquals(expected, ByteSource.concat(ImmutableList.of(b1, b2, b3).iterator()).read());
assertEquals(expected.length, ByteSource.concat(b1, b2, b3).size());
assertFalse(ByteSource.concat(b1, b2, b3).isEmpty());
ByteSource emptyConcat = ByteSource.concat(ByteSource.empty(), ByteSource.empty());
assertTrue(emptyConcat.isEmpty());
assertEquals(0, emptyConcat.size());
}
public void testConcat_infiniteIterable() throws IOException {
ByteSource source = ByteSource.wrap(new byte[] {0, 1, 2, 3});
Iterable<ByteSource> cycle = Iterables.cycle(ImmutableList.of(source));
ByteSource concatenated = ByteSource.concat(cycle);
byte[] expected = {0, 1, 2, 3, 0, 1, 2, 3};
assertArrayEquals(expected, concatenated.slice(0, 8).read());
}
private static final ByteSource BROKEN_CLOSE_SOURCE =
new TestByteSource(new byte[10], CLOSE_THROWS);
private static final ByteSource BROKEN_OPEN_SOURCE =
new TestByteSource(new byte[10], OPEN_THROWS);
private static final ByteSource BROKEN_READ_SOURCE =
new TestByteSource(new byte[10], READ_THROWS);
private static final ByteSink BROKEN_CLOSE_SINK = new TestByteSink(CLOSE_THROWS);
private static final ByteSink BROKEN_OPEN_SINK = new TestByteSink(OPEN_THROWS);
private static final ByteSink BROKEN_WRITE_SINK = new TestByteSink(WRITE_THROWS);
private static final ImmutableSet<ByteSource> BROKEN_SOURCES =
ImmutableSet.of(BROKEN_CLOSE_SOURCE, BROKEN_OPEN_SOURCE, BROKEN_READ_SOURCE);
private static final ImmutableSet<ByteSink> BROKEN_SINKS =
ImmutableSet.of(BROKEN_CLOSE_SINK, BROKEN_OPEN_SINK, BROKEN_WRITE_SINK);
public void testCopyExceptions() {
if (!Closer.SuppressingSuppressor.isAvailable()) {
// test that exceptions are logged
TestLogHandler logHandler = new TestLogHandler();
Closeables.logger.addHandler(logHandler);
try {
for (ByteSource in : BROKEN_SOURCES) {
runFailureTest(in, newNormalByteSink());
assertTrue(logHandler.getStoredLogRecords().isEmpty());
runFailureTest(in, BROKEN_CLOSE_SINK);
assertEquals((in == BROKEN_OPEN_SOURCE) ? 0 : 1, getAndResetRecords(logHandler));
}
for (ByteSink out : BROKEN_SINKS) {
runFailureTest(newNormalByteSource(), out);
assertTrue(logHandler.getStoredLogRecords().isEmpty());
runFailureTest(BROKEN_CLOSE_SOURCE, out);
assertEquals(1, getAndResetRecords(logHandler));
}
for (ByteSource in : BROKEN_SOURCES) {
for (ByteSink out : BROKEN_SINKS) {
runFailureTest(in, out);
assertTrue(getAndResetRecords(logHandler) <= 1);
}
}
} finally {
Closeables.logger.removeHandler(logHandler);
}
} else {
// test that exceptions are suppressed
for (ByteSource in : BROKEN_SOURCES) {
int suppressed = runSuppressionFailureTest(in, newNormalByteSink());
assertEquals(0, suppressed);
suppressed = runSuppressionFailureTest(in, BROKEN_CLOSE_SINK);
assertEquals((in == BROKEN_OPEN_SOURCE) ? 0 : 1, suppressed);
}
for (ByteSink out : BROKEN_SINKS) {
int suppressed = runSuppressionFailureTest(newNormalByteSource(), out);
assertEquals(0, suppressed);
suppressed = runSuppressionFailureTest(BROKEN_CLOSE_SOURCE, out);
assertEquals(1, suppressed);
}
for (ByteSource in : BROKEN_SOURCES) {
for (ByteSink out : BROKEN_SINKS) {
int suppressed = runSuppressionFailureTest(in, out);
assertTrue(suppressed <= 1);
}
}
}
}
public void testSlice_returnEmptySource() {
assertEquals(ByteSource.empty(), source.slice(0, 3).slice(4, 3));
}
private static int getAndResetRecords(TestLogHandler logHandler) {
int records = logHandler.getStoredLogRecords().size();
logHandler.clear();
return records;
}
private static void runFailureTest(ByteSource in, ByteSink out) {
try {
in.copyTo(out);
fail();
} catch (IOException expected) {
}
}
/** @return the number of exceptions that were suppressed on the expected thrown exception */
private static int runSuppressionFailureTest(ByteSource in, ByteSink out) {
try {
in.copyTo(out);
fail();
} catch (IOException expected) {
return CloserTest.getSuppressed(expected).length;
}
throw new AssertionError(); // can't happen
}
private static ByteSource newNormalByteSource() {
return ByteSource.wrap(new byte[10]);
}
private static ByteSink newNormalByteSink() {
return new ByteSink() {
@Override
public OutputStream openStream() {
return new ByteArrayOutputStream();
}
};
}
}
|
{
"pile_set_name": "Github"
}
|
2012 Superlatives: The Best Of The Best
It's been an amazing year for ultimate. We will have a lot more about the biggest stories of the year, but we wanted to get things started with a list of the biggest and best from 2012. From Best Game to Biggest Spike, we've got you covered for the top moments from the year.
It’s been an amazing year for ultimate. We will have a lot more about the biggest stories of the year, but we wanted to get things started with a list of the biggest and best from 2012. Did we miss something? Should we give out awards in other categories? Let us know in the comments.
Disclaimer: Ultiworld started in July 2012, so really this is the best of the best since then. We didn’t follow the college season closely, so we leave that to commenters!
Best Game: Seattle Sockeye v. Austin Doublewide at the Emerald City Classic.
Best Play: You already know…it’s Dylan Freechild’s universe point layout D in the Portland Rhino v. Doublewide game at the Labor Day Championships. It’s the number one play in our Labor Day Top 10.
Best Achievement: TIE. Doublewide’s first national championship, marking the first non-coastal winner in the open division since 1986 — AND — San Francisco Fury’s seventh straight title, giving them the longest win streak in any division in US history.
Biggest Upset: Toronto’s GOAT over Atlanta’s Chain Lightning at the Chesapeake Invite.
Biggest News: The explosion of professional ultimate, pioneered by Josh Moore and the American Ultimate Disc League. Now expanding further, with Major League Ultimate and, potentially, NexGen.
Most Memorable Line: “You’re not on the FIELD!” — Steve Slocum, Tournament Director, Mid-Atlantic Regionals, after Philadelphia Southpaw players were outraged due to a rules infraction late in their finals matchup against Washington, D.C.’s Truck Stop.
Biggest Scandal: Canada’s play against Japan at Worlds.
Biggest Apology: Team Canada to Japan.
Biggest Surprise: The AUDL lawsuit.
Best Breakout Rookie: Doublewide’s Dalton Smith.
Best Breakout Team: Madison’s Heist, who finished 11th at Nationals after not losing a single game all season.
Best Spirit: We didn’t see this ourselves, but this letter from a spectator at Canadian Nationals shows how powerful great spirit can be.
Most Viral Video: Brodie Smith’s Top 21 Trick Shots.
Top Reddit Post: The official candy of Ultimate (comments section).
Biggest Longshot: Ultimate making it into the Olympics.
Most-Read Ultiworld Article: Exclusive Early Details About Major League Ultimate
Biggest Spike: Medicine Men’s Rico Johnson (#10 in our Mid-Atlantic Regionals Top 10).
Biggest Fan: Sherri Rose.
Best Highlight Reel: World Juniors Ultimate Championships.
Best Play By A Dog: Davy Whippet and Rob McLeod, World Record, Distance to Canine Catch.
|
{
"pile_set_name": "OpenWebText2"
}
|
Izu Kokubun-ji
is a Buddhist temple located in what is now the city of Mishima, Shizuoka, Japan. It is the modern successor of one of the provincial temples established by Emperor Shōmu during the Nara period (710 – 794) for the purpose of promoting Buddhism as the national religion of Japan and standardising control of Yamato rule over the provinces. The foundation stones for the seven-story pagoda of original temple was designated as a National Historic Site by the Japanese government in 1956.
History
Izu Kokubun-ji was founded in 741 as the provincial temple of Izu Province. It was converted at some uncertain time to the Shingon sect, and was burned down repeatedly in the incessant battles between the forces the Takeda clan and the Odawara Hōjō clan during the Sengoku period. In the early Edo period, it converted to the Nichiren sect and was completely rebuilt; however, these buildings were all destroyed by the 1855 Ansei Edo earthquake and the site was abandoned.
In 1923, the temple was rebuilt again as , and was renamed to Izu Kokubun-ji in 1954. Archaeological investigations were conducted in 1956. The foundations of the South Gate, Central Gate, Kon-dō and Kōdō (Lecture Hall) were discovered, as wells as the foundations of the surrounding cloister. The layout of the buildings was in accordance with the standardized "Shichidō garan" formation, similar to Tōdai-ji in Nara, upon which the kokubunji temples were based.
Outside of the main complex, eight foundation stones of the original Nara-period Kokubun-ji’s pagoda were uncovered, immediately behind the present temple’s Hondō. The foundation stones were made of tuff, and duding from the size and layout of the foundation, the pagoda was seven-stories high with a height of 60-meters.
The temple is located approximately one kilometer south of modern Mishima Station on the Tōkaidō Main Line railway.
See also
List of Historic Sites of Japan (Shizuoka)
provincial temple
References
External links
Mishima City home page
Shizuoka Prefecture official site
Category:Historic Sites of Japan
Category:Mishima, Shizuoka
Category:Izu Province
Category:8th-century establishments in Japan
Category:Nara period
Category:Archaeological sites in Japan
Category:Nichiren-shū temples
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Karnitsky
Karnitsky () is a surname. Notable people with the surname include:
Alyaksandr Karnitsky (born 1989), Belarusian footballer
Valery Karnitsky (born 1991), Belarusian footballer
A Polish form of the name is Karnicki
Borys Karnicki
Category:Belarusian-language surnames
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Callipara queketti
Callipara queketti is a species of sea snail, a marine gastropod mollusk in the family Volutidae, the volutes.
Description
Distribution
References
External links
Category:Volutidae
Category:Gastropods described in 1901
|
{
"pile_set_name": "Wikipedia (en)"
}
|
CONFIG_SMP=y
CONFIG_NR_CPUS=4
CONFIG_HOTPLUG_CPU=y
CONFIG_PREEMPT_NONE=n
CONFIG_PREEMPT_VOLUNTARY=n
CONFIG_PREEMPT=y
|
{
"pile_set_name": "Github"
}
|
名前：名無しさん投稿日：2015年09月11日
名前：名無しさん投稿日：2015年09月11日
一人一シリアルだってよ
名前：名無しさん投稿日：2015年09月11日
本人確認をさせて頂く場合がございます(するとは言ってない)
名前：名無しさん投稿日：2015年09月11日
じゃ家族の名前で申し込んで最悪身分証かりるわ
名前：名無しさん投稿日：2015年09月11日
1人1シリアルと本確は朗報
名前：名無しさん投稿日：2015年09月11日
1人1シリアル？
つまりBD８枚も予約した俺死亡なの？
名前：名無しさん投稿日：2015年09月11日
1シリアルで連番無効ってことなら運営GJすぎる
名前：名無しさん投稿日：2015年09月11日
発表遅すぎね？すでに複数予約したやつどうすんの？
名前：名無しさん投稿日：2015年09月11日
なんだよこの色々憶測する発表はよ
名前：名無しさん投稿日：2015年09月11日
複垢でいけば問題ない
名前：名無しさん投稿日：2015年09月11日
複垢同住所はアウト？
名前：名無しさん投稿日：2015年09月11日
複数応募は弾くって書きなよ
名前：名無しさん投稿日：2015年09月11日
あ、違うな
連番無効じゃなくて、一人１シリアルってだけか。
名前：名無しさん投稿日：2015年09月11日
尼なら余裕でキャンセルできるぞ
名前：名無しさん投稿日：2015年09月11日
家族の名前借りれば良いじゃん
名前：名無しさん投稿日：2015年09月11日
これで1尻4限とかだったら笑う
名前：名無しさん投稿日：2015年09月11日
住所同じだと弾かれる可能性
別居ならいいんだけど
名前：名無しさん投稿日：2015年09月11日
あー1シリアルだから1dayだな
名前：名無しさん投稿日：2015年09月11日
同一住所名前は弾くぐらいはっきり書いて欲しい
名前：名無しさん投稿日：2015年09月11日
1人1シリアルって、前回までみたいに1つのシリアルで2人分予約できないわけか？
名前：名無しさん投稿日：2015年09月11日
まあ今後も改訂はあるでしょ
騒ぐことじゃない
名前：名無しさん投稿日：2015年09月11日
運営としては1人１シリアルとか制限しないでどんどん積ませた方が儲かるんじゃないの？
名前：名無しさん投稿日：2015年09月11日
わざわざ書くってことはやっぱりキャパ変わらず改修後のSSAかね会場は
名前：名無しさん投稿日：2015年09月11日
友達に申込み頼むかな
友達少ないけど、まあ10シリアル分くらいならアテが・・・無ぇよ
名前：名無しさん投稿日：2015年09月11日
ということはだ、2daysだった場合普通に申し込んだらどちらか片方しか行けないな
まあ、２次受付とか一般販売とかあるんだろうけど
名前：名無しさん投稿日：2015年09月11日
あれ…尼はキャンセルすりゃあいいけどタペ3種ネットで予約しちまったよ
4シリ分の住所用意しないとダメかな
名前：名無しさん投稿日：2015年09月11日
もう無理や・・・
名前：名無しさん投稿日：2015年09月11日
キッズをふるいにかけられない・・・
金の力でブイブイ言わせられない・・・
名前：名無しさん投稿日：2015年09月11日
一人1尻って事は先行外れたらあとは一般発売しかないって事だろうな
名前：名無しさん投稿日：2015年09月11日
これ倍率変わるんかなぁ絶対行きたいんなら
結局友人に頼んで複垢するしかないんじゃね
名前：名無しさん投稿日：2015年09月11日
今回は金の力で積みまくって行こうとしたのになー
名前：名無しさん投稿日：2015年09月11日
複数応募が完全に抽選から除外されるなら詰み
名前：名無しさん投稿日：2015年09月11日
実家の名前と住所借りて応募すればいいってことかね
てか同一住所弾く可能性あんのね
てか一人１シリって転売にしか旨味ない気がすんだけど？
名前：名無しさん投稿日：2015年09月11日
5thの時はシリアルいっぱい入れたらいっぱい当たることもあるって答えが返ってきた
けど実際は1人1シリしか有効でなく連日当選すらなかった
複アカでシリアル散らした奴が勝ち組になってた
今回は初めから1人1シリしか有効じゃないて言ってる分まだ良心的
名前：名無しさん投稿日：2015年09月11日
どうせタペは揃えるつもりだったしいいや、もうAmazon以外キャンセル出来ねぇ
名前：名無しさん投稿日：2015年09月11日
こうどなじょうほうせん
確実にいえることはキャンセルする奴は後悔することになる
名前：名無しさん投稿日：2015年09月11日
20枚積む予定のガチの友人死亡ってことか…
名前：名無しさん投稿日：2015年09月11日
※チケットは一般販売でもご購入可能です
この１文に可能性感じた
名前：名無しさん投稿日：2015年09月11日
住所一緒でも名義別なら大丈夫でしょ
兄弟5人ラブライバーなら殺し合いになるよ
名前：名無しさん投稿日：2015年09月11日
4枚積んだけどこれ以上増量は出来なくなったのは確か
売り上げ的にはこれでいいんかなぁ…
名前：名無しさん投稿日：2015年09月11日
テンバイヤーは適当な住所で済ませられるのが現実
名前：名無しさん投稿日：2015年09月11日
お一人様それほど買わなくていいですよってことか
もうこれは相当デカい会場なんじゃね？
名前：名無しさん投稿日：2015年09月11日
高校生とか結構な人数参加してきてライブが荒れそう
まあ本確は文面見る限り今回もやる気なさそうだし友達に応募頼むか
名前：名無しさん投稿日：2015年09月11日
けどこれは公式が積んでも無駄よ宣言したようなもんだな
名前：名無しさん投稿日：2015年09月11日
連番なら良いけど単発の場合名前一緒だと死ぬんじゃね？
いや同姓同名ですーでセーフか
.＿__
.／ ノヽ＼
;| （○）（○|: はああああああああああ？
:|ヽ (_人_)/; １人１シリアルってことは、複数購入意味ないってことじゃん
. :| |. ⌒ .|; まぁ家族の名前とか使えばいけそうな気もしないけど・・・
:h ／;
: /; ’
/ く､ ＼
;| ＼＼_ ＼
;|ミ |`ー＝っ ＼
.＿__
.／ノ ヽヽ:
;| （0):（0)|: でもこれもし外れたら・・・ショックってレベルじゃないんだけど・・・
:|ヽ (人)/:
. :| | ^ |.
:h ノ::
:| /:
/ く､ ＼
;| ､`､_､ ＼
;|ミ| ￣＾＼ ＼
.＿_
./ ハヽ
;| 0:0| ６個くらい購入するつもりだったのに・・・
:|ヽω/ いやだ・・･堕ちたくない・・・落ちたくないよおおお
. :|| ‘^ |.
:h ノ
:|/:
/ |､ ＼
;| ､､_､ ＼
;|ﾐ| ￣＾＼ ＼
|
{
"pile_set_name": "OpenWebText2"
}
|
Sample preparation of carbadox, furazolidone, nitrofurazone, and ethopabate in medicated feeds for high pressure liquid chromatography.
Medicated feeds (pelleted or mash) containing guarantees of carbadox, furazolidone, nitrofurazone, and ethopabate are pretreated with water, extracted with 95% dimethylformamide overnight at room temperature, cleaned up on a column of alumina, and injected into a high pressure liquid chromatograph for quantitative measurement. Carbadox, nitrofurazone, and furazolidone can be separated; chromatograms show excellent baseline resolution, and results are in good agreement with colorimetric methods. The same extraction and cleanup can be used to improve colorimetric methods for furazolidone and nitrofurazone.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
At the end of last year, mayor of Nelson Mandela Bay Athol Trollip made a commitment that the city would eradicate bucket toilets in informal settlements by the end of 20171. It’s not surprising that the pledge was made: the municipality has the highest number of bucket toilets in the country, according to data released from Stats SA’s latest Non-financial census of municipalities report.
Despite Nelson Mandela Bay making headway in reducing the number of consumer units2 using bucket toilets from a high of 30 202 in 2013 to 16 317 in 2016, the system remains an ugly reality for many residents. In May 2016, frustrations over bucket toilets erupted into a violent service delivery protest within the city3.
The discontent is understandable: of the 68 028 consumer units in South Africa using a bucket toilet in 2016, 24% were located in Nelson Mandela Bay.
Politicians and municipal administrators speak out against bucket toilets while residents find the system deplorable. Nonetheless, 45 of South Africa’s 278 municipalities still provide communities with this form of sanitation, with most of these in Free State, Northern Cape and Eastern Cape.
Municipalities responding to the Non-financial census of municipalities survey often state that their initial intention is to provide bucket toilets as a temporary measure while more permanent sanitation options are being considered. Various factors – including a lack of funding – often delay the process, making it difficult to replace bucket toilets once they’ve been introduced.
The evidence indicates a decrease in the number of bucket toilets in recent years, with municipalities making a concerted effort to provide other options such as septic tanks and flush toilets. During 2016, two more municipalities managed to eradicate bucket toilets entirely: Khâi-Ma and Siyathemba, both located in Northern Cape.
In trying to figure out which municipalities struggle the most, it’s useful to consider another way of looking at the data. Nelson Mandela Bay might have the highest prevalence of bucket toilets in the country, but it also has one of the largest populations. If the number of consumer units using bucket toilets is expressed as a percentage of total domestic consumer units receiving some form of toilet facility, Nelson Mandela Bay loses its top spot.
Instead, Siyancuma in Northern Cape comes out as the municipality with the highest percentage of domestic consumer units using bucket toilets, followed by Setsoto (Free Sate) and Mafube (Free State).
So will Nelson Mandela Bay keep its promise of totally eradicating bucket toilets by the end of this year? Stats SA will be your go-to source to answer this question, as the data will appear in its 2017 release of Non-financial census of municipalities report. So watch this space!
Visit the download page for the 2016 Non-financial census of municipalities report here. The PDF report and Excel file with municipal data are available. Data on bucket toilets are available in Table 9.
2 A consumer unit is an entity to which the service is delivered, and which receives one bill if the service is billed, alternatively known as a delivery point. This concept is often referred to as a household by municipalities, but this is not strictly correct, as households and consumer units do not coincide one to one, particularly in blocks of flats, on stands where there are multiple households in the same dwelling, or in additional dwellings, such as garden flats, backyard rooms, etc., and in the case of public taps.
Careers
User Survey
Stats SA is in the process of updating its database of all users. Kindly participate in this short survey and provide your details. All details will be kept confidential and is for the use of Stats SA only.
Name and Surname*
Organisation*
Field of statistical work/interest*
Cell phone number*
Email address*
What information do you want to receive from Stats SA and how often?*
Do you prefer the information emailed to you or downloading it online?*
|
{
"pile_set_name": "Pile-CC"
}
|
Natural history of opportunistic disease in an HIV-infected urban clinical cohort.
To determine the effect of contemporary clinical care on the natural history of opportunistic disease in an urban population infected with human immunodeficiency virus (HIV). Urban university HIV clinic. Retrospective and prospective observational study. 1246 HIV-infected patients with CD4+ counts of 300 cells/mm3 or less. Incidence rates and Kaplan-Meier estimates of the probability of developing opportunistic disease with time, distribution of the CD4+ counts at which opportunistic disease develops, survival after the development of opportunistic disease, and the association between preventive drug therapies and the occurrence of opportunistic infection. The most common opportunistic disease was Candida esophagitis, which had an incidence of 13.3 events per 100 person-years and a 3-year Kaplan-Meier probability of 0.30. Pneumocystis carinii pneumonia, Mycobacterium avium complex bacteremia, cytomegalovirus, and the acquired immunodeficiency syndrome dementia complex occurred at rates of 5 to 9 events per 100 person-years and 3-year Kaplan-Meier probabilities of 0.15 to 0.22. Toxoplasmosis, cryptococcal meningitis, herpes zoster, the wasting syndrome, and Kaposi sarcoma occurred at rates of about 2 to 4 events per 100 person-years and with 3-year Kaplan-Meier probabilities of 0.05 to 0.10. Non-Hodgkin lymphoma, M. tuberculosis infection, progressive multifocal leukoencephalopathy, and cryptosporidiosis were the least common disorders, with an incidence of about 1 to 2 events per 100 person-years and a 3-year Kaplan-Meier probability less than 0.05. Only the incidences of cryptococcal meningitis, secondary P. carinii pneumonia, and herpes zoster decreased (P < 0.05) between 1989-1992 and 1993-1995. Fluconazole use was associated with a decreased relative rate of 0.49 (P = 0.06) for cryptococcal meningitis and a decreased relative rate of 0.61 (P = 0.005) for esophageal candidiasis. Rifabutin use was associated with a decreased relative rate of 0.37 (P = 0.002) for M. avium complex bacteremia, and trimethoprim-sulfamethoxazole use was associated with decreased relative rates of 0.33 (P = 0.02) for secondary P. carinii pneumonia and 0.55 (P = 0.08) for primary P. carinii pneumonia. Candidiasis, herpes zoster, and M. tuberculosis infection first occurred at a median CD4+ count greater than 100 cells/mm3, but all other opportunistic diseases first occurred at a median CD4+ count less than 50 cells/mm3. Median survival after diagnosis varied from 35 days for non-Hodgkin's lymphoma to 680 days for herpes zoster. In the patients studied, the incidences of secondary P. carinii pneumonia, cryptococcal meningitis, and herpes zoster have declined in the past 5 years. The incidences of primary P. carinii pneumonia and Kaposi sarcoma appear to be declining compared with historical estimates. However, although these and other opportunistic diseases continue to be relatively frequent complications of HIV infection, they are first occurring at more advanced immunosuppression than in the past. Continued efforts are needed to develop effective strategies for preventing opportunistic disease in very advanced HIV infection.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
What do a tomato and a bacterium have in common? They both rely on the DNA molecule
to direct their functions and characteristics. While each species has its own
unique combination of DNA, called its genome, the molecule works by
the same rules in all living organisms. What does this mean for a high-tech plant
breeder? That useful genes from one organism can potentially be transferred to
the genome of another.
Scientists have done this with the bacterium Bacillus
thuringiensis, or Bt for short. Bt, which is found naturally
in soil and on the leaves of many plants, is toxic to many plant
pests. The genes that give it that toxicity were first put into
potatoes in 1997. Now many vegetables grown commercially have
the Bt genes in them.
This gene-adding process is called “genetic engineering”
(GE). GE is one form of what’s called “genetic modification”
(GM). Though GE and GM are related, they are not the same. “Genetically
modified” refers to any plant (or other organism) whose DNA
has been manipulated by humans. This could happen through artificial
selection and hybridization or through engineering. “Genetically
engineered” refers specifically to one or more extra genes
that have been inserted into a plant’s genome from an organism
it can’t mingle genes with via sexual reproduction. The products
of such engineering feats— like Bt tomatoes—are called
“transgenic.”
Intuitively, GE doesn’t seem like a very “natural” process.
After all, genes in nature don’t cross so many boundaries.
But the line between natural and unnatural may be grayer than you
think.
In the wild, most plants (actually, most living things) share
genes only with members of their own or very closely related species
living in nearby environments. Hybridization allows plant breeders
to go beyond these limitations that exist in nature. Suppose a
wild tomato exists that seems particularly resistant to cold. A
breeder can cross this plant with one that produces store-bought
tomatoes in order to impart these useful ancestral genes. The resulting
tomatoes may look and taste like the ones we’re familiar
with, and their DNA doesn’t contain any “nontomato” genes.
But this cross isn’t natural: these two plants would never
have found each other in the wild. In fact, the grocer’s
tomato doesn’t even exist in nature. It’s the product
of hundreds of similar hybridizations.
The main difference between a hybridized tomato and a transgenic
one is that the latter has additional bits of DNA that originally
came from another species. It has an altered genome. Critics of
genetic engineering have argued that this process may have unintended
consequences: They are concerned that seeds from GE plants could
carry their genetic alterations into the environment and the food
supply, or that untested GE products may pose health risks for
consumers. Supporters of genetic engineering counter that the benefits
of GE—plants with added nutrition, higher yields, or the
ability to produce pesticides or drugs—outweigh the risks,
and that a world with rapidly increasing population can’t
be fed without them.
Different countries have adopted different policies toward the
growing, selling, and labeling of genetically engineered foods.
For example, GE ingredients in packaged foods have to be identified
as such in Europe, but not in the United States. Until March 2005,
farmers in Brazil were not allowed to grow GE crops, while farmers
in the United States have long been encouraged to do so. In addition,
the companies that develop genetically-engineered crop plants claim
ownership of the seeds and any plants they produce as offspring
for several generations. This has met with resistance from farmers
around the world, some of whom have had pollen from GE plants blown
or carried into their fields, and others in developing countries
who would like the benefits of genetic engineering without being
dependent on seed producers for their seed supply.
As a backyard gardener, you’re not likely to be planting
any transgenics. These engineered plants have been developed for
agriculture. They are patented, and the companies that sell them
would make you aware of that when you buy them. That doesn’t
mean, however, that the plants you grow in your garden are
the products of completely natural processes. There’s manipulations
everywhere out there.
Links presenting various views on transgenics, from the Action
Bioscience Web site, produced by the American Institutes of Biological
Sciences:
Biotechnology and the Green Revolution: An Interview with
Norman Borlaughhttp://www.actionbioscience.org/biotech/borlaug.html
Borlaugh has been involved in crop improvement since the 1940s, and won the
Nobel Peace prize in 1970 for his work. He is an staunch advocator of genetic
engineering to increase crop yields and provide more food in developing nations.
The Ecological Impacts of Agricultural Biotechnology
By Miguel A. Altierihttp://www.actionbioscience.org/biotech/altieri.htmlAltieri teaches agroecology at the University of California at
Berkeley, and also works on sustainable agriculture issues for
the United Nations. He has been outspoken about his concerns regarding
the ecological risks of plant biotechnology.
The Debate Over Genetically Modified Foods
By Kerryn Sakkohttp://www.actionbioscience.org/biotech/sakko.htmlThis article describes some of the differences between breeding
and genetically engineering crops, and presents arguments on both
sides of the issue. Written by an undergraduate student from Australia
who represented her country at the 2004 Youth Science Festival
in Singapore, sponsored by Asia-Pacific Economic Cooperation Forum.
Other Web site resources:
Food Futurehttp://www.foodfuture.org.uk/home.aspxProduced on behalf of the food and drink industry in the
United Kingdom, this site presents many perspectives on questions
such as “Who owns GM technology” and “What about
diversity?”
|
{
"pile_set_name": "Pile-CC"
}
|
Facebook is on the path to platform lock-in with Oculus Quest.
In case you are unfamiliar with the concept, “customer lock-in” is an economics idea that “makes a customer dependent on a vendor for products and services, unable to use another vendor without substantial switching costs.”
With the standalone Oculus Quest, Facebook is funding Oculus Studios projects for both Oculus Quest and Oculus Rift and enabling cross-buy between them. For people who own both an Oculus Rift and Oculus Quest, this means the same Facebook-funded title should be playable on either headset with a single purchase. Third-party studios not funded by Oculus can also opt into the feature should they so choose.
It makes sense strategically — the move is a good one short-term for people buying any of Facebook’s low-cost headsets. If you’re an Oculus Rift owner today, for example, and you’ve purchased Apex Construct you should find the game already waiting in your library when you buy a Quest. Great, right?
Short term? Sure.
Long-term? That’s a bit more complicated.
Steam’s PC Store vs Oculus’ VR Store
Valve Corporation’s Steam is the storefront of choice for millions of gamers with PCs. Valve is a privately owned company based in Bellevue, Washington and, in 2016, it partnered with HTC to invest in the launch of a VR headset that would encourage people to use Steam to purchase VR games.
Critically, owners of Facebook’s Oculus Rift headset (and Microsoft’s Windows VR devices) could also play games from Steam through a Valve-controlled software bridge. If you owned an HTC Vive you could buy games from either Valve or from HTC’s Viveport and play them on the headset. If you owned a Rift, you could buy and play games from Oculus, Steam, or HTC.
Essentially, Valve established Steam as the default place for people to buy VR games because that’s where they already bought their PC games. The only exceptions to this, generally speaking, are Rift owners who buy their content from Oculus because it is the default store. The bottom line? Oculus essentially made their store a walled garden for Rift while Valve made Steam a place for all PC VR headsets.
Facebook, meanwhile, paid for a number of exclusive games like Robo Recall and Lone Echo to be sold only via the Oculus Store. Games on the store only officially work for Facebook headsets. A hack called “Revive”, though, made it possible for some Vive owners to run Oculus-exclusive content they purchased from Facebook.
How Oculus Quest Changes The Game
Things are changing in Facebook’s strategy after the departures of Oculus founders Brendan Iribe and Palmer Luckey.
In Blake Harris’ book The History Of The Future a set of emails are disclosed offering insight into Facebook CEO Mark Zuckerberg’s strategy in VR and AR. According to the book, Zuckerberg was in favor of a “closed” approach to the Rift while Oculus leaders pushed for more openness. An email in the book attributed to Brendan Marten, Facebook’s head of finance for AR/VR, provides a sense of Zuckerberg’s strategy in 2015.
From the book:
Zuckerberg “sees that there are risks to closing the platform, but does not believe they are that great or irreversible, and is willing to take that chance. He believes that you can open it once closed, but going the other way is much harder.”
“Mark believes that some Oculus employees will be unhappy with closing the platform, and even that it goes against the culture and could cause unhappiness across the team, but considers that a reasonable price. . . For example, how many long-time Oculus employees (the ones most likely to strongly object to closing the platform) will walk away from their RSUs? If they don’t, how many will do their jobs worse as a result? Would even highly-principled John [Carmack] walk away?”
“Unless we present Mark with a long-term plan for Rift success that he thinks makes sense, I believe Rift is probably going to be somewhere between resource-starved and orphaned soon.”
“Either: a) we come up with an explanation for openness that Mark can believe makes sense, b) we decide to go closed, or c) there is going to be a head-on collision. And remember, even if we win that head-on collision, Mark controls budget and head count, so the Rift may find itself starved regardless.”
The many billions of dollars Zuckerberg spends on VR and AR boils down to something very simple he laid out in a message to employees: “We are vulnerable on mobile to Google and Apple because they make major mobile platforms. We would like a stronger strategic position in the next wave of computing.”
This is where lock-in comes into play. Right now, the most successful non-mobile VR platform is the wired PlayStation VR system which sold more than 4.2 million headsets since late 2016. PSVR is also home to many of the best VR games centered around intuitive hand interaction. We believe Oculus Rift’s low prices, quality content and significant marketing help Facebook’s headset outsell HTC Vive for PC systems, but have no hard figures to confirm that feeling.
Yet, for many buyers, Valve’s Steam store remains the default place to buy most content.
Zuckerberg does not want that dynamic to continue in the second generation of PC VR. So he’s taking the curated content approach that worked for PSVR and applying it to the Quest VR console. Thus we have consumers ready to experience this:
I bought exclusively on Steam to ensure compatibility with future headsets should I move away from Rift as my primary. Now that many devs are confirming cross-buy with Rift and Quest in the Oculus store, I have shot myself squarely in the foot. — Ben Fineman (@bfineman) April 9, 2019
I believe Fineman’s comment represents the position of a lot of early PC VR buyers.
Platform Services And Lock-In
Buying a VR PC and mounting sensors on your walls to have a top quality VR experience is an enormous undertaking. Given the cost and effort involved, it makes sense that most buyers who picked up PC-based headsets between 2016 and 2019 carefully researched and planned their decisions to protect their investment long-term.
Many VR buyers with a personal computer view it as an “open” platform. In practice, though, “openness” is a more complicated concept than it seems at first.
For example, you generally need Microsoft’s Windows today to enjoy VR apps on PCs. Nonetheless, Valve’s SteamVR underpinnings are supported on Apple computers too and engineers at the company have been working to make it easier for developers to make their apps work fine on Linux as well. In other words, Valve is aware of the “lock-in” both developers and consumers have related to Microsoft’s Windows and they are actively working to weaken those ties.
In comparison, Apple support for Oculus was killed off years ago. Last year, Rift devices worldwide went offline for a full day because somebody (apparently) forgot to renew a security certificate with Microsoft. That sort of thing doesn’t inspire confidence in PC buyers.
Now with Quest, Facebook is poised to enable cross-buy which could begin the process of locking in Quest buyers to the Oculus “platform.” Do you wake up one day and you can’t enjoy a great VR game without logging into your Facebook account first? That reality starts to take shape when people enjoy their Oculus Quest so much they decide they want to upgrade to the increased graphical fidelity of Oculus Rift S.
Differing Paths Detailed
Valve and Facebook are on different paths in 2019.
Facebook’s path offers low hardware prices to consumers and money to developers to build projects first for Oculus and implement features like cross-buy with Rift. Since Facebook is positioning Quest as a standalone VR console, they expect it to appeal to a different set of buyers. Over time, we can expect Facebook’s networking, avatar, analytics, and identity services to become a larger part of the package the company is pushing out to developers and consumers across all VR devices.
Valve’s path is likely more expensive up front for consumers but continues in a direction they’ve stuck to for years — increased immersion leading to better VR games and more satisfying experiences. For example, with a Valve Index Controller you can “wear” a VR controller for the first time rather than “hold” one.
That change comes from years of focused effort developing games, and hardware, in collaboration with trusted developers. We’ve yet to feel the result of that effort, or the effect it has on the value proposition of Valve Index and the Steam ecosystem compared to Facebook’s, or even Microsoft’s. As soon as we have that experience we’ll boil down the differences in clear language so everyone interested in this market can understand the dynamics at play.
For now, though, consumers should keep in mind the decisions made now about which headsets you buy is a vote in favor of something larger than the headset itself. Likewise, the decisions developers make now about which headsets to support, services to use and features to activate in their games could influence the creators who come after.
The VR market is still small enough that there’s a weight to your choices over the next month that could have a lasting impact on policies at these companies for years to come. Choose wisely.
|
{
"pile_set_name": "OpenWebText2"
}
|
Adherence to the Mediterranean diet and risk of stroke and stroke subtypes.
Several meta-analyses including a small number of cohorts showed inverse associations between the Mediterranean Diet (MedDiet) and risk of stroke. However, it remains unclear whether such a relation varies by region of the study population or by major subtypes of stroke. We searched PubMed and EMBASE databases for relevant studies and we further included unpublished results from the Singapore Chinese Health Study (N = 57,078) and the Seguimiento Universidad de Navarra (SUN) study (N = 12,670). We used a random-effects model to calculate summary relative risk (RR) with 95% confidence intervals (CI) of stroke for each 4-point increment of the MedDiet score, roughly corresponding to the difference between extreme quintiles of the MedDiet score among participants of the included studies. The final analyses included 20 prospective cohort studies involving 682,149 participants and 16,739 stroke cases. The summary RRs for each 4-point increment of the MedDiet score were 0.84 (95% CI 0.81-0.88; I2 = 11.5%) for all combined, 0.76 (95% CI 0.65-0.89) for studies in Mediterranean populations and 0.86 (95% CI 0.83-0.89) for those in non-Mediterranean populations. Lower risk of stroke associated with higher MedDiet score also was observed in the analyses stratified by study population and methodological characteristics including study risk of bias, version of the MedDiet index, and definition of moderate alcohol consumption. The MedDiet was similarly associated with lower risk of ischemic stroke (RR 0.86, 95% CI 0.81-0.91; nine studies) and hemorrhagic stroke (RR 0.83, 95% CI 0.74-0.93; eight studies). Our meta-analysis suggests that adhering to the Mediterranean diet was associated with lower risk of stroke in both Mediterranean and non-Mediterranean populations, and for both ischemic stroke and hemorrhagic stroke risk.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Grand design on a small scale; AD showcases a handful of charming structures that evoke an idealized way of life
“We knew we would have lots of work to do, but we didn’t care, because we fell in love with its shape, the land, the view,” Roseline Glazer says of the shingled cottage on Martha’s Vineyard that she shares with her husband, Bill. Originally built by Roger Allen in 1930, architect Joseph W. Dick helped with the renovation. A brick courtyard wraps around the entrance to the kitchen, at the rear of the house. “The chokecherry tree over the cottage is nature’s umbrella,” Glazer remarks. June 2008
|
{
"pile_set_name": "Pile-CC"
}
|
You know how Canadians aren’t really in a mad rush to accept a high number of refugees in a short period of time?
Well, as some of us learned at a technical briefing held in Ottawa by government officials Wednesday morning, neither are a lot of those refugees in a rush to come to Canada.
Turns out, a lot of people want to go home when the war is over.
The 271 Syrian refugees that have arrived since November 4th when the new government took power, were people whose applications were already in process, and then there are the other 9,090 people in process that aren’t here yet. That’s right - ones that the mean old Harper government were working on bringing in. Conservatives focused their efforts on private sponsorships so that would be mainly people with family in Canada who have connections to our country.
Many of the others simply don’t want to come. They don’t hate Canada per se, it’s just that they’d rather go home and that’s why they prefer to stay in refugee camps in Turkey so they’ll be close to home when the war is over.
Funny, that was a point that Stephen Harper made during the election campaign. He quoted people he had spoken to who said that Canada could help by helping those on the ground that didn’t want to leave. For that he was described as uncaring.
This refugee rush has been a white liberal guilt exercise from the beginning - driven by partisan politics and the heartbreaking picture of a dead toddler shared generously by the main-stream media along with the false story that his family had been rejected by Canada.
The Liberals promised 25,000 refugees, a number which has now grown with Immigration Minister John McCallum saying we’ll likely take in close to 50,000 by the end of 2016 but so far, just 1,800 of 41,000 contacted even expressed an interest.
It looks like the hesitancy of Canadians to take part in the refugee rush is matched by the hesitancy of Syrians who don’t seem to want to be part of it either.
JOIN TheRebel.media for more fearless news and commentary you won’t find anywhere else.
Trudeau's Liberals MUST halt plans to bring in 25,000 Syrian refugees
until they can guarantee the safety of the Canadian people.
SIGN THE PETITION at RefugeePause.ca
“Miss Me Yet?” The Harper t-shirt & bumpersticker you asked for!
BUY now ONLY from TheRebel.media:
|
{
"pile_set_name": "OpenWebText2"
}
|
Randomised trials in surgery: the burden of evidence.
A randomised controlled trial (RCT) is considered the hierarchical peak of evidence-based medicine and a general demand for any result to be evaluated by RCTs has evolved. Yet, many advances in operative surgery do not result from RCTs and many controversies remain without an RCT being conducted. A randomised comparison of laparoscopic versus open liver resection has recently been called for. Using such a trial and others as examples, we examine the limitations of randomised design in skill-dependant interventions. Surgical procedures are skill-dependant, constantly developing, irreversible and traumatising. Additionally, placebo control is usually unethical and adequate blinding difficult or impossible to accomplish. Under these circumstances, surgeon and patient participation will be problematic and the resulting data will tend to have low external validity. While some of these obstacles can be modified, others will remain. Nonrandomised, prospective cohort comparison has other weaknesses, but may add complementary data with good external validity. An alternative hierarchy of evidence is warranted in this field.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Wholesale China Jerseys
To reach an acceptable level with respect to revenue collection,Cheap Jerseys Store, which has been a challenge for the National Insurance Scheme (NIS),Authentic NCAA Jerseys Cheap, current measures will have to improve “a hundred times more”, according to Board Chairman, Dr. Roger Luncheon.“That’s a pledge I made to the Opposition and I have no difficulties. I think all of the Directors and Management see the obligation to move our contribution or revenue in more positive grounds to stop having those deficits that were so common in 2011.”He is adamant that the NIS is not a sinking ship and will not need financial assistance from Central Government for its survival in the near future. He also claims the entity started this year with a surplus and is hoping that it is sustained“The difference between those of us who know… I have been the Chairman of NIS now for a lifetime, and I am not really of the opinion that is being bandied around that it is on its knees and couple mornings from now there is going to be a big bailout from Central Government. I certainly don’t think so.”These assertions were made in a recent interview with this publication in relation to the current status of the NIS in light of troubling revelations in its Eighth Actuarial Review released late last year.The review, an independent analysis of the fund as at December 31, 2011,Cheap Jerseys, suggested that the NIS is nearing crisis stage and will be exhausted in less than 10 years,Jerseys NFL Cheap, unless contribution rate adjustments and benefit reforms are done immediately.Dr. Luncheon’s claim of a surplus at the beginning of 2013 came just weeks after the scheme’s General Manager Terry Thomas at a forum at the University of Guyana’s Berbice Campus had stressed that the NIS is operating at a deficit – it is paying out more benefits than contributions received on a monthly basis.Thomas had stated that the Actuarial Review was all about correcting the imbalance.Chairman of NIS Board, Dr. Roger LuncheonIt was pointed out in the review that this could be corrected “by increasing contribution rates; by having much better and fuller compliance by employers and self-employed persons; by stalling or delaying the pension payments—that is—increasing the pensionable qualifying age and it can also be corrected by reducing the pension of those who come into the scheme now”.According to Dr. Luncheon,Wholesale Authentic Jerseys, compliance of citizens to make their contributions remains a major problem. He added that the NIS’s revenue and expenditure are not significantly mismatched.“It is mismatched where expenditure is concerned, particularly benefits expenditure. We spend close to $1B per month in benefits expenditure. We are also about a $1B for overall revenue,Cheap NFL Jerseys,” he said.
|
{
"pile_set_name": "Pile-CC"
}
|
---
abstract: 'We develop a novel preconditioning method for ridge regression, based on recent linear sketching methods. By equipping Stochastic Variance Reduced Gradient (SVRG) with this preconditioning process, we obtain a significant speed-up relative to fast stochastic methods such as SVRG, SDCA and SAG.'
author:
- 'Alon Gonen[^1]'
- 'Francesco Orabona[^2]'
- 'Shai Shalev-Shwartz[^3]'
bibliography:
- 'bib.bib'
title: |
Solving Ridge Regression using\
Sketched Preconditioned SVRG
---
Introduction {#sec:intro}
============
Consider the *ridge regression* problem: $$\label{eq:ridge}
\min_{w \in \reals^d}\ \left\{L(w) = \frac{1}{n}\sum_{i=1}^n \frac{1}{2}(w^\top x_i-y_i)^2 + \frac{\lambda}{2} \|w\|^2\right\},$$ where $\lambda>0$ is a regularization parameter, $x_i\in{\mathbb{R}}^d$ and $y_i\in {\mathbb{R}}$ for $i=1,\cdots,n$ the training data. We focus on the large scale regime, where both $n$ and $d$ are large. In this setting, stochastic iterative methods such as SDCA [@shalev2013stochastic], SVRG [@johnson2013accelerating], and SAG [@roux2012stochastic] have become a standard choice for minimizing the objective $L$. Specifically, the overall complexity of a recent improved variant of SVRG due to [@xiao2014proximal] depends on the average condition number, which is defined as follows. Denote the empirical correlation matrix and its eigenvalue decomposition by $$\label{eq:covSpec}
C := \frac{1}{n} \sum_{i=1} ^n x_i x_i^\top = \sum_{i=1} ^d \lambda_i u_i u_i^\top~.$$ The average condition number of $C+\lambda I$ is defined as the ratio between the trace of the Hessian of $L$ and its minimal eigenvalue: $$\label{eq:initialCondition}
{\hat{\kappa}}:={\hat{\kappa}}(C+\lambda I) = \frac{\tr(C+\lambda I)}{\lambda_d(C+\lambda I)} = \sum_{i=1}^d \frac{\lambda_i+\lambda}{\lambda_d+\lambda} ~.$$ The mentioned variant of SVRG finds an $\epsilon$-approximate minimizer of $L$ in time $\tilde{O}(({\hat{\kappa}}+n)d \log(1/\epsilon))$. Namely, the output of the algorithm, denoted $\hat{w}$, satisfies $\bE[L(\hat{w})]-L(w^\star) \le \epsilon$, where the expectation is over the randomness of the algorithm. For an accelerated version of the algorithm, we can replace ${\hat{\kappa}}$ by $\sqrt{n {\hat{\kappa}}}$ [@shalev2014accelerated; @lin2015universal].
The regularization parameter, $\lambda$, increases the smallest eigenvalue of $C+\lambda I$ to be at least $\lambda$, thus improves the condition number and makes the optimization problem easier. However, to control the under/over fitting tradeoff, $\lambda$ has to decrease as $n$ increases [@shalev2014understanding]. Moreover, in many machine learning applications $\lambda_d$ approaches zero and it is usually smaller than the value of $\lambda$. Overall, this yields a large condition number in most of the interesting cases.
A well-known approach for reducing the average condition number is *preconditioning*. Concretely, for a (symmetric) positive definite (pd) matrix $P \in \reals^{d \times d}$, we define the preconditioned optimization problem as $$\label{eq:condRidge}
\min_{{\tilde{w}}\in \reals^d} {\tilde{L}}({\tilde{w}}):= L(P^{-1/2} {\tilde{w}}) ~.$$ Note that ${\tilde{w}}$ is an $\epsilon$-approximate minimizer of ${\tilde{L}}$ if and only if $w = P^{-1/2} {\tilde{w}}$ forms an $\epsilon$-approximate minimizer of $L$. Hence, we can minimize rather than . As we shall see, the structure of the objective allows us to apply the preconditioning directly to the data (as a preprocessing step) and consequently rewrite the preconditioned objective as a ridge regression problem with respect to the preconditioned data (see ). For a suitable choice of a matrix $P$, the average condition number is significantly reduced. Precisely, as will be apparent from the analysis, the pd matrix that minimizes the average condition number is $P=C+\lambda I$, and the corresponding average condition number is $d$. However, we note that such preconditioning process would require both the computation of $P^{-1/2}$ and the computation of $P^{-1/2} x_i$ for each $i \in [n]$. By first order conditions, computing $(C+\lambda I)^{-1/2}$ is equivalent to solving the original problem in , rendering this “optimal” preconditioner useless. Yet, the optimal preconditioner might not needed in many cases. In fact, a common empirical observation (see ) is that (high-dimensional) machine learning problems tend to have few dominant features, while the other coordinates are strongly correlated with the stronger features. As a result, the spectrum of the correlation matrix decays very fast. Hence, it is natural to expect to gain a lot from devising preconditioning methods that focus on the stronger directions of the data.
Our contributions are as follows. We develop a relatively cheap preconditioning method that, coupled with SVRG, assures to speed-up the convergence in practical applications while having a computational cost comparable to SVRG alone. In order to approximately extract the stronger directions while incurring a low computational cost, we rely on a variant of the Block Lanczos method due to [@musco2015stronger] in order to compute an approximated truncated SVD (Singular Value Decomposition) of the correlation matrix $C$. Finally, by equipping SVRG with this preconditioner, we obtain our main result.
Main Result {#sec:mainResult}
===========
\[thm:main\] Let $k \in [d]$ be a given parameter and assume that the regularization parameter, $\lambda$, is larger than $\lambda_d$. Our preconditioning process runs in time $O(ndk \log(n))$. By equipping the SVRG of [@xiao2014proximal] with this preconditioner, we find an $\epsilon$-approximate minimizer for (with probability at least $9/10$) in additional runtime of $O ((\tilde{\kappa} + n + d )d \log(1/\epsilon) )$, where $\tilde{\kappa}=\frac{k \lambda_k+ \sum_{i>k} \lambda_i}{\lambda}$ or $\tilde{\kappa} = \left(\frac{n(k \lambda_k+ \sum_{i>k} \lambda_i)}{\lambda}\right)^{1/2}$ if we use accelerated SVRG.
When the runtimes of both the (accelerated) SVRG and our preconditioned (accelerated) SVRG are controlled by the average condition number (and both runtimes dominate $ndk$), then ignoring logarithmic dependencies, we obtain a speed-up of order $$\label{eq:ratioRidge}
\textrm{ratio} = \frac{\sum_{i=1}^d \lambda_i}{k\,\lambda_k+\sum_{i>k}\lambda_i}
= \frac{\sum_{i=1}^k \lambda_i +\sum_{i>k}\lambda_i }{k\,\lambda_k ~~~~+\sum_{i>k}\lambda_i}~.$$ (or $\sqrt{\sum_{i=1}^d \lambda_i/(\lambda_k k+\sum_{i>k}\lambda_i)}$ if acceleration is used) over SVRG. If the spectrum decays fast then $k \, \lambda_k \ll \sum_{i=1}^k \lambda_i$ and $\sum_{i>k}\lambda_i \ll k \, \lambda_k$. In this case, the ratio will be large. Indeed, as we show in the experimental section, this ratio is often *huge* for relatively small $k$.
Main challenges and perspective
-------------------------------
While the idea of developing a preconditioner that focuses on the stronger directions of the data matrix sounds plausible, there are several difficulties that have to be solved.
- First, since a preconditioner must correspond to an invertible transformation, it is not clear how to form a preconditioner based on a low rank approximation and, in particular, how should we treat the non-leading components.
- One of the main technical challenges in our work is to translate the approximation guarantees of the Lanczos method into a guarantee on the resulted average condition number. The standard measures of success for low-rank approximation are based on either Frobenius norm or spectral norm errors. As will be apparent from the analysis (see ), such bounds do not suffice for our needs. Our analysis relies on stronger per vector error guarantees due to [@musco2015stronger].
It should be emphasized that while we use a variant of SVRG due to [@xiao2014proximal], we could equally use a variant of SDCA [@shalev2016sdca] or develop such a variant for SAG or SAGA. Furthermore, while we focus on the quadratic case, we believe that our ideas can be lifted to more general setting. For example, when applied to self-concordant functions, each step of Newton’s method requires the minimization of a quadratic objective. Therefore, it is natural to ask if we can benefit from applying our method for approximating the Newton step.
Bias-complexity tradeoff {#sec:reduceReg}
------------------------
As we mentioned above, $\lambda$ controls a tradeoff between underfitting and overfitting. In this view, we can interpret our result as follows. Assuming for simplicity that $n \ge d$ and ignoring logarithmic dependencies, we note that if $$\label{eq:decay}
\lambda = \frac{k \lambda_k+\sum_{i>k}\lambda_i}{nk}~,$$ then the runtime of our preconditioned SVRG is $\tilde{O}(ndk)$. For comparison, the runtime of (unconditioned) SVRG is $\tilde{O}(ndk)$ if $$\label{eq:decaySVRG}
\lambda = \frac{\sum_{i=1}^d \lambda_i}{nk}~.$$ The ratio between the RHS of and is the ratio given in . Hence, for a given “runtime budget” of order $\tilde{O}(ndk)$, we can set the regularization parameter of the preconditioned SVRG to be smaller by this ratio. Similar interpretation holds for the accelerated versions.
Related Work {#sec:related}
============
#### Existing algorithms and their complexities:
Since minimizing is equivalent to solving the system $(C+\lambda I)w=\frac{1}{n} \sum_{i=1}^n y_i x_i$, standard numerical linear algebra solvers such as Gaussian elimination can be used to solve the problem in time $O(nd^2)$.
Iterative deterministic methods, such as Gradient Descent (GD), finds an $\epsilon$-approximate minimizer in time $nd \kappa \log(1/\epsilon)$, where $\kappa=\frac{\lambda_1(C+\lambda I)}{\lambda_d(C+\lambda I)}$ is the condition number of $C+\lambda I$ (see Theorem 2.1.15 in [@nesterov2004introductory]). The Kaczmarz algorithm [@kaczmarz1937angenaherte] has an identical complexity. Both the Conjugate Gradient (CG) method [@hestenes1952methods] and the Accelerated Gradient Descent (AGD) algorithm of [@nesterov1983method] enjoy a better runtime of $nd \sqrt{\kappa}\log(1/\epsilon)$. In fact, CG has a more delicate analysis (see Corollary 16.7 in [@vishnoi2012laplacian]): If all but $c \in [d]$ eigenvalues of $C+\lambda I$ are contained in a range $[a,b]$, then the runtime of CG is at most $nd(c+\sqrt{b/a} \log(1/\epsilon))$. In particular, CG’s runtime is at most $O(nd^2)$. Furthermore, following the interpretation of our main result in , we note that for a “runtime budget” of $\tilde{O}(ndk)$, we can set the regularization parameter of CG to be of order $\lambda_k/k^2$ (which is usually much greater than the RHS of ).
#### Linear Sketching:
Several recently developed methods in numerical linear algebra are based on the so-called *sketch-and-solve* approach, which essentially suggests that given a matrix $A$, we first replace it with a smaller random matrix $AS$, and then perform the computation on $AS$ [@woodruff2014sketching; @clarkson2013low; @sarlos2006improved]. For example, it is known that if the entries of $S$ are i.i.d. standard normal variables and $S$ has $p=\Omega(k/\epsilon)$ columns, then with high probability, the column space of $AS$ contains a $(1+\epsilon)$ rank-$k$ approximation to $A$ with respect to the Frobenius norm. This immediately yields a fast PCA algorithm (see Section 4.1 in [@woodruff2014sketching]).
While the above sketch-and-solve approach sounds promising for this purpose, our analysis reveals that controlling the Frobenius norm error does not suffice for our needs. We need spectral norm bounds, which are known to be more challenging [@witten2013randomized]. Furthermore, as mentioned above, the success of our conditioning method heavily depends on the stronger per vector error guarantees obtained by [@musco2015stronger] which are not obtained by simpler linear sketching methods.
#### Sketched preconditioning:
Recently, subspace embedding methods were used to develop cheap preconditioners for linear regression with respect to the squared loss [@woodruff2014sketching]. Precisely, [@clarkson2013low] considered the case $\lambda=0$ (i.e, standard least-squares) and developed a preconditioning method that reduces the average condition number to a constant. Thereafter, they suggest applying a basic solver such as CG. The overall running time is dominated by the preconditioning process which runs in time $\tilde{O}(d^3+nd)$. Hence, a significant improvement over standard solvers is obtained if $n \gg d$.
The main shortcoming of this method is that it does not scale well to large dimensions. Indeed, when $d$ is very large, the overhead resulted from the preconditioning process can not be afforded.
#### Efficient preconditioning based on random sampling:
While we focus on reducing the dependence on the dimensionality of the data, other work investigated the gain from using only a random subset of the data points to form the conditioner [@yang2014data]. The theoretical gain of this approach has been established under coherence assumptions [@yang2014data].
Preliminaries {#sec:pre}
=============
Additional notation and definitions {#sec:notation}
-----------------------------------
Any matrix $B \in \reals^{d \times n}$ of rank $r$ can be written in (thin) SVD form as $B = U \Sigma V^\top = \sum_{i=1}^r \sigma_i(B) u_i v_i^\top$. The singular values are ordered in descending order. The spectral norm of $B$ is defined by $\|B\|=\sigma_1(B)$. The spectral norm is submultiplicative, i.e., $\|AB\| \le \|A\|\|B\|$ for all $A$ and $B$. Furthermore, the spectral norm is unitary invariant, i.e., for all $A$ and $U$ such that the columns of $U$ are orthonormal, $\|UA\| = \|A\|$. For any $k \in [r]$, it is well known that the truncated SVD of $B$, $B_k := U_k \Sigma_k V_k = \sum_{i=1} ^{k} \sigma_i(B) u_i v_i^\top$, is the best rank-$k$ approximation of $B$ w.r.t. the spectral norm [@trefethen1997numerical]. A twice continuously differentiable function $f: \reals^d \rightarrow \reals$ is said to be $\beta$-smooth if $\|\nabla^2 f(w)\| \leq \beta$ for all $w$, where $\nabla^2 f(w)$ is the Hessian of $f$ at $w$. $f$ is said to be $\alpha$-strongly convex if $\lambda_d(\nabla^2 f(w)) \ge \alpha$ for all $w$. If $g$ is convex and $f$ is $\alpha$-strongly convex, then $f+g$ is $\alpha$-strongly convex.
Stochastic Variance Reduced Gradient (SVRG) {#sec:svrg}
-------------------------------------------
We consider a variant of the Stochastic Variance Reduced Gradient (SVRG) algorithm of [@johnson2013accelerating] due to [@xiao2014proximal]. The algorithm is an epoch-based iterative method for minimizing an average, $F(w) = \frac{1}{N} \sum_{i=1}^N f_i(w)$, of smooth functions. It is assumed that each $f_i:\reals^d \rightarrow \reals$ is convex and $\beta_i$-smooth. The entire function $F$ is assumed to be $\alpha$-strongly convex. The algorithm is detailed in . Its convergence rate depends on the averaged smoothness of the individual functions and the average condition number of $F$, defined as $$\label{eq:conditionSVRG}
\hat{\beta} = \frac{1}{N} \sum_{i=1}^N \beta_i ~~~;~~~ {\hat{\kappa}}_F = \frac{\hat{\beta}}{\alpha} ~.$$
**[@xiao2014proximal]** \[thm:svrg\] Fix $\epsilon>0$. Running SVRG () with any $w_0$, $S \ge \log((F(w_0) - \min_{w \in \reals^d} F(w))/\epsilon)$, $m = \lceil {\hat{\kappa}}_F \rceil$, and $\eta = 0.1/\hat{\beta}$ yields an $\epsilon$-approximate minimizer of $F$. Furthermore, assuming that each single gradient $\nabla f_i(w)$ can be computed in time $O(d)$, the overall runtime is $O(({\hat{\kappa}}_F+N)d \log(\epsilon_0/\epsilon))$.
**Input:** Functions $f_1,\ldots,f_n, \beta_1,\ldots,\beta_n$ **Parameters:** $\bar{w}_0 \in \reals^d$, $m$, $\eta$, $S \in \mathbb{N}$ $\bar{w} = \bar{w}_{s-1}$ $\bar{v} = \nabla F(\bar{w})$ $w_0 = \bar{w}$ Pick $i_t \in [N]$ with probability $q_{i_t}=\beta_{i_t}/\sum \beta_j$ $v_t = (\nabla f_{i_t}(w_{t-1}) -\nabla f_{i_t}(\bar{w}))/q_{i_t}+\bar{v}$ $w_t = w_{t-1} - \eta v_t$ $\bar{w}_s = \frac{1}{m} \sum_{t=1}^m w_t$ **Output:** the vector $\bar{w}_S$
In the original definition of SVRG [@johnson2013accelerating], the indices $i_t$ are chosen uniformly at random from $[n]$, rather than proportional to $\beta_i$. As a result, the convergence rate depends on the maximal smoothness, $\max \{\beta_i\}$, rather than the average, $\hat{\beta}$. It will be apparent from our analysis (see ) that in our case, $\max \{\beta_i\}$ is proportional to the maximum norm of any preconditioned $x_i$. Since we rely on the improved variant of [@xiao2014proximal], our bound depends on the average of the $\beta_i$’s, which scale with the average norm of the preconditioned $x_i$’s. To simplify the presentation, in the sequel we refer to as SVRG.
Randomized Block Lanczos
------------------------
A randomized variant of the Block Lanczos method due to [@musco2015stronger] is detailed[^4] in . Note that the matrix ${\tilde{U}}_k {\tilde{\Sigma}}_k {\tilde{V}}_k^\top$ forms an SVD of the matrix ${\tilde{A}}_k := Q(Q^\top A)_k = {\tilde{U}}_k {\tilde{U}}_k^\top A$.
**Input:** $A \in \reals^{d \times n}, k \le d, \epsilon' \in (0,1)$ $q= \Theta \left(\frac{\log(n)}{\sqrt{\epsilon}} \right)$, $p=qk$, $\Pi \sim \cN(0,1)^{n \times k}$ Compute $K = [A \Pi,(AA^\top) A \Pi,\ldots, (AA^\top)^{q-1} A \Pi]$Orthonormalize $K$’s columns to obtain $Q \in \reals^{d \times qk}$ Compute the truncated SVD $(Q^\top A)_k = {\tilde{W}}_k {\tilde{\Sigma}}_k {\tilde{V}}_k^\top$ Compute ${\tilde{U}}_k = Q {\tilde{W}}_k$ **Output:** the matrices ${\tilde{U}}_k,{\tilde{\Sigma}}_k,{\tilde{V}}_k$
\[thm:musco\] **[@musco2015stronger]** Consider the run of and denote ${\tilde{A}}_k = {\tilde{U}}_k {\tilde{\Sigma}}_k {\tilde{V}}_k = \sum_{i=1}^k {\tilde{\sigma}}_i {\tilde{u}}_i {\tilde{v}}_i^\top$. Denote the SVD of $A$ by $A = \sum_{i=1}^d \sigma_i v_i u_i^\top$. The following bounds hold with probability at least $9/10$: $$\|A - {\tilde{A}}_k\| \le (1+\epsilon') \|A-A_k\| \le (1+\epsilon') \sigma_k$$ $$\begin{aligned}
\forall i \in [k],~~|z_i^\top AA^\top z_i - u_i^\top AA^\top u_i| &= |{\tilde{\sigma}}_i^2-\sigma_i^2 | \notag \\
&\le \epsilon' \sigma_{k+1}^2 ~. \label{eq:perVec}\end{aligned}$$ The runtime of the algorithm is $O\left(\frac{nd k \log(n)}{ \sqrt{\epsilon'}}+ \frac{k^2(n+d)}{\epsilon'} \right)$.
Sketched Conditioned SVRG {#sec:scsvrg}
=========================
In this section we develop our sketched conditioning method. By analyzing the properties of this conditioner and combining it with SVRG, we will conclude .
Recall that we aim at devising cheaper preconditioners that lead to a significant reduction of the condition number. Specifically, given a parameter $k \in [d]$, we will consider only preconditioners $P^{-1/2}$ for which both the computation of $P^{-1/2}$ itself and the computation of the set $\{P^{-1/2}x_i, \ldots, P^{-1/2} x_n\}$ can be carried out in time $\tilde{O}(ndk)$. We will soon elaborate more on the considerations when choosing the preconditioner, but first we would like to address some important implementation issues.
Preconditioned regularization {#sec:regCond}
-----------------------------
In order to implement the preconditioning scheme suggested above, we should be able to find a simple form for the function $\tilde{L}$. In particular, since we would like to use SVRG, we should write $\tilde{L}$ as an average of $n$ components whose gradients can be easily computed. Denote by ${\tilde{x}}_i = P^{-1/2}x_i$ for all $i \in [n]$. Since for every $i \in [n]$, $((P^{-1/2} w)^\top x_i-y_i)^2=(w^\top {\tilde{x}}_i - y_i)^2$, it seems natural to write $\tilde{L}(w)=L(P^{-1/2}w)$ as follows: $$\tilde{L}(w) = \frac{1}{n} \sum_{i=1}^n \underbrace{\frac{1}{2}(w^\top {\tilde{x}}_i-y_i)^2}_{=:\tilde{\ell}_i} +\frac{\lambda}{2} \|P^{-1/2}w\|^2~.$$ Assume momentarily that $\lambda=0$. Note that the gradient of $\tilde{\ell}_i$ at any point $w$ is given by $\nabla \tilde{\ell}_i(w_t)=(w^\top {\tilde{x}}_i-y_i){\tilde{x}}_i$. Hence, by computing all the ${\tilde{x}}_i$’s in advance, we are able to apply SVRG directly to the preconditioned function and computing the stochastic gradients in time $O(d)$.
When $\lambda>0$, the computation of the gradient at some point $w$ involves the computation of $P^{-1} w$. We would like to avoid this overhead. To this end, we decompose the regularization function as follows. Denote the standard basis of $\reals^d$ by $e_1,\ldots,e_d$. Note that the function $L$ can be rewritten as follows: $$\begin{aligned}
L(w) = \frac{1}{n+d} \sum_{i=1}^{n+d} \ell_i(w)~,\end{aligned}$$ where $\ell_i(w) = \frac{n+d}{n} \frac{1}{2} (w^\top x_i-y_i)^2$ for $i=1,\ldots,n$ and $\ell_{n+i}(w)=\lambda(n+d) \frac{1}{2}(w^\top e_i)^2$ for $i=1,\ldots,d$. Finally, denoting $b_i = P^{-1/2} e_i$ for all $i$, we can rewrite the preconditioned function $\tilde{L}$ as follows: $$\begin{aligned}
{\tilde{L}}(w) = \frac{1}{n+d} \sum_{i=1}^{n+d} {\tilde{\ell}}_i(w)~,\end{aligned}$$ where ${\tilde{\ell}}_i(w) = \frac{n+d}{n} \frac{1}{2} (w^\top {\tilde{x}}_i-y_i)^2$ for $i=1,\ldots,n$ and ${\tilde{\ell}}_{n+i}(w)=\lambda(n+d) \frac{1}{2}(w^\top b_i)^2$ for $i=1,\ldots,d$. By computing the ${\tilde{x}}_i$’s and the $b_i$’s in advance, we are able to apply SVRG while computing stochastic gradients in time $O(d)$.
The effect of conditioning
--------------------------
We are now in position to address the following fundamental question: How does the choice of the preconditioner, $P^{-1/2}$, affects the resulted average condition number of the function ${\tilde{L}}$ (\[eq:conditionSVRG\])? The following lemma upper bounds ${\hat{\kappa}}_{{\tilde{L}}}$ by the average condition number of the matrix $P^{-1/2}(C+\lambda I)P^{-1/2}$, which we denote by ${\tilde{\kappa}}$ (when the identity of the matrix $P$ is understood).
\[thm:condEffect\] Let $P^{-1/2}$ be a preconditioner. Then, the average condition number of $\tilde{L}$ is upper bounded by $${\hat{\kappa}}_{\tilde{L}} \le {\tilde{\kappa}}= \frac{\tr (P^{-1/2} (C + \lambda I ) P^{-1/2} ) } {\lambda_d(P^{-1/2} (C + \lambda I ) P^{-1/2})} ~.$$
The proof is in the appendix. Note that an optimal bound of $O(d)$ is attained by the whitening matrix $P^{-1/2} = (C+\lambda I)^{-1/2}$.
Exact sketched conditioning
---------------------------
Our sketched preconditioner is based on a random approximation of the best rank-$k$ approximation of the data matrix. It will be instructive to consider first a preconditioner that is based on an exact rank-$k$ approximation of the data matrix. Let $X \in \reals^{d \times n}$ be the matrix whose $i$-th columns is $x_i$ and let $\bar{X}=n^{-1/2} X$. Denote by $\bar{X}= \sum_{i=1}^{\textrm{rank}(\bar{X})} \sigma_i u_i v_i^\top = U \Sigma V^\top$ the SVD of $\bar{X}$ and recall that $\bar{X}_k= \sum_{i=1}^k \sigma_i u_i v_i^\top$ is the best $k$-rank approximation of $\bar{X}$. Note that $\bar{X}\bar{X}^\top=C$ and therefore $\sigma_i^2 = \lambda_i(C)=\lambda_i$. Furthermore, the left singular vectors of $\bar{X}$, $u_1,\ldots,u_k$, coincide with the $k$ leading eigenvectors of the matrix $C$. Consider the preconditioner, $$P^{-1/2} = \sum_{i=1}^k \frac{u_i u_i^\top}{\sqrt{\lambda_i+\lambda}} + \frac{I-\sum_{i=1}^k u_i u_i^\top}{\sqrt{\lambda_k+\lambda}}~,$$ where $u_{k+1},\ldots, u_d$ are obtained from a completion of $u_1,\ldots,u_k$ to an orthonormal basis.
\[lem:exactCondSketch\] Let $k \in [d]$ be a parameter and assume that the regularization parameter, $\lambda$, is larger than $\lambda_d$. Using the exact sketched preconditioner, we obtain $$\label{eq:exactCondSketchCond}
{\hat{\kappa}}_{\tilde{L}} \le \frac{k \lambda_k+ \sum_{i>k} \lambda_i}{\lambda} + d ~.$$
A simple calculation shows that for $i=1,\ldots,k$, $$\lambda_{i} (P^{-1/2} (C + \lambda I ) P^{-1/2} ) = \frac{\lambda_i+\lambda}{\lambda_i+\lambda} = 1 ~.$$ Similarly, for $i=k+1,\ldots, d$, $$\lambda_{i} (P^{-1/2} (C + \lambda I ) P^{-1/2} ) = \frac{\lambda_i+\lambda}{\lambda_k+\lambda} ~.$$ Finally, $$\lambda_{d} (P^{-1/2} (C + \lambda I ) P^{-1/2}) \ge \frac{\lambda}{\lambda_k+\lambda}~.$$ Combining the above with , we obtain that $$\begin{aligned}
{\hat{\kappa}}_{\tilde{L}} &\le \frac{\tr(P^{-1/2} (C + \lambda I ) P^{-1/2})}{\lambda_d(P^{-1/2} (C + \lambda I ) P^{-1/2})} \\
&\le k\frac{\lambda_k+\lambda}{\lambda} + \sum_{i=k+1}^d \frac{\lambda_i+\lambda}{\lambda} \\
&= \frac{k \lambda_k + \sum_{i>k} \lambda_i}{\lambda} + d ~. \qedhere\end{aligned}$$
Sketched conditioning {#sec:lowCond}
---------------------
An exact computation of the SVD of the matrix $\bar{X}$ takes $O(nd^2)$. Instead, we will use the Block Lanczos method in order to approximate the truncated SVD of $\bar{X}$. Specifically, given a parameter $k \in [d]$, we invoke the Block Lanczos method with the parameters $\bar{X},k$ and $\epsilon'=1/2$. Recall that the output has the form $\tilde{X}_k = {\tilde{U}}_k {\tilde{\Sigma}}_k {\tilde{V}}_k^\top= \sum_{i=1}^k {\tilde{\sigma}}_i{\tilde{u}}_i {\tilde{v}}_i^\top$. Analogously to the exact sketched preconditioner, we define our sketched preconditioner by $$\label{eq:condSketch}
P^{-1/2} = \sum_{i=1}^k \frac{{\tilde{u}}_i {\tilde{u}}_i^\top}{\sqrt{{\tilde{\sigma}}_i^2+\lambda}} + \frac{I-\sum_{i=1}^k {\tilde{u}}_i {\tilde{u}}_i^\top}{\sqrt{{\tilde{\sigma}}_k^2+\lambda}} ~.$$
\[thm:sketchCond\] Let $k \in [d]$ be a parameter and assume that the regularization parameter, $\lambda$, is larger that $\lambda_d$. Using the sketched preconditioner defined in , up to a multiplicative constant, we obtain the bound on the average condition number with probability at least $9/10$.
The rest of this section is devoted to the proof of . We follow along the lines of the proof of . Up to a multiplicative constant, we derive the same upper and lower bounds on the eigenvalues of $P^{-1/2} (C+\lambda I)P^{-1/2}$.
From now on, we assume that the bounds in (where $\epsilon'=1/2$) hold. This assumption will be valid with probability of at least $9/10$. We next introduce some notation. We can rewrite $P^{-1/2} = {\tilde{U}}({\tilde{\Sigma}}^2+\lambda I)^{-1/2} {\tilde{U}}^\top$ where ${\tilde{\Sigma}}$ is a diagonal $d \times d$ with ${\tilde{\Sigma}}_{i,i} = {\tilde{\sigma}}_i$ if $i \le k$ and ${\tilde{\Sigma}}_i = {\tilde{\sigma}}_k$ if $i>k$. and the columns of ${\tilde{U}}$ are a completion of ${\tilde{u}}_1,\ldots,{\tilde{u}}_k$ to an orthonormal basis. Recall that the SVD of $\bar{X}$ is denoted by $\bar{X} = \sum_{i=1}^d \sigma_i u_i v_i^\top = U \Sigma V^\top$.
**(Upper bound on the leading eigenvalue)** We have $$\lambda_1(P^{-1/2} (C+\lambda I)P^{-1/2}) \le 17~.$$
Since $\lambda_1(P^{-1/2} (C+\lambda I)P^{-1/2}) = \|P^{-1/2} (C+\lambda I)P^{-1/2}\| = \|P^{-1/2} C P^{-1/2} + \lambda P^{-1}\|$, using the triangle inequality we have that $$\lambda_1(P^{-1/2} (C+\lambda I)P^{-1/2}) ~\le~ \|P^{-1/2} C P^{-1/2}\| + \lambda \|P^{-1}\| ~.$$ By the definition of $P$ we have that $\|P^{-1}\| = \frac{1}{{\tilde{\sigma}}_k^2 + \lambda}$ and therefore the second summand on the right hand side of the above is at most $\frac{\lambda}{{\tilde{\sigma}}_k^2+\lambda} \le 1$. As to the first summand, recall that $C = \bar{X} \bar{X}^\top$ and therefore $\|P^{-1/2} C P^{-1/2}\| = \|\bar{X}^\top P^{-1/2} \|^2$. We will show that $\|\bar{X}^\top P^{-1/2} \| \le 4$ which will imply that $\|P^{-1/2} C P^{-1/2}\| \le 16$. To do so, we first apply the triangle inequality, $$\begin{aligned}
&\|\bar{X}^\top P^{-1/2}\| = \| ({\tilde{X}}_k + (\bar{X} - {\tilde{X}}_k))^\top P^{-1/2} \| \\
&\quad \le \| {\tilde{X}}_k^\top P^{-1/2} \| + \|(\bar{X} - {\tilde{X}}_k)^\top P^{-1/2} \| ~.\end{aligned}$$ Let us consider one term at the time. Recall that ${\tilde{X}}_k = {\tilde{U}}_k {\tilde{\Sigma}}_k {\tilde{V}}_k^\top$. Since ${\tilde{U}}_k^\top {\tilde{U}}\in \reals^{k,d}$ is a diagonal matrix with ones on the diagonal, and since the spectral norm is invariant to multiplication by unitary matrices, we obtain that $$\begin{aligned}
&\|{\tilde{X}}_k^\top P^{-1/2}\| = \|{\tilde{V}}_k {\tilde{\Sigma}}_k {\tilde{U}}_k^\top {\tilde{U}}({\tilde{\Sigma}}^2+\lambda I)^{-1/2} {\tilde{U}}^\top\| \\
&\quad = \|{\tilde{\Sigma}}_k {\tilde{U}}_k^\top {\tilde{U}}({\tilde{\Sigma}}^2+\lambda I)^{-1/2} \| \\
&\quad = \max_{i \in [k]} \frac{{\tilde{\sigma}}_i}{ \sqrt{{\tilde{\sigma}}_i^2+\lambda}}
\le \max_{i \in [k]} \frac{{\tilde{\sigma}}_i}{{\tilde{\sigma}}_i+\sqrt{\lambda}}
\le 1 ~.\end{aligned}$$ Next, by the submutiplicativity of the spectral norm, $$\|(\bar{X} - {\tilde{X}}_k)^\top P^{-1/2} \| \le \|\bar{X}-{\tilde{X}}_k\| \cdot \|P^{-1/2}\| ~.$$ implies that $\|\bar{X}-{\tilde{X}}_k\| \le \tfrac{3}{2} \sigma_k$ and $$\begin{aligned}
&\|P^{-1/2}\| = \frac{1}{ \sqrt{{\tilde{\sigma}}_k^2+\lambda}}
\le \frac{1}{\sqrt{{\tilde{\sigma}}_k^2}} \le \frac{1}{ \sqrt{\sigma_k^2 - (1/2) \sigma_{k+1}^2}} \\
&\quad \le \frac{1}{ \sigma_{k} \sqrt{\tfrac{1}{2}} } = \frac{ \sqrt{2}}{ \sigma_k}
<\frac{ 2}{\sigma_k}~.\end{aligned}$$ Hence, $\|\bar{X}-{\tilde{X}}_k\| \cdot \|P^{-1/2}\| \le 3$. Combining all of the above bounds concludes our proof.
**(Refined upper bound on the last $d-k$ eigenvalues)** For any $i \in \{k+1,\ldots,d\}$, $$\lambda_i \left(P^{-1/2} (C+ \lambda I)P^{-1/2} \right) \le \frac{2 (\lambda_i+\lambda)}{\lambda_k+\lambda}~.$$
Using the Courant minimax principle [@bhatia2013matrix], we obtain the following bound for all $i \in \{k+1,\ldots,d\}$: $$\begin{aligned}
&\lambda_i \left(P^{-1/2} (C+ \lambda I ) P^{-1/2} \right) \\
&= \max_{\substack{\mathcal{M} \subseteq \reals^d:\\ \dim(\mathcal{M}) =i}}\min_{\substack{x \in \cM:\\x \neq 0}} \frac{ x^\top P^{-1/2} (C+\lambda I) P^{-1/2} x}{\|x\|^2} \\
& = \max_{\substack{\mathcal{M} \subseteq \reals^d:\\ \dim(\mathcal{M}) =i}}\min_{\substack{x \in \cM:\\ x \neq 0}} \frac{ x^\top P^{-1/2} (C+\lambda I) P^{-1/2} x}{\|P^{-1/2}x\|^2} \cdot \frac{\|P^{-1/2}x\|^2} {\|x\|^2} \\
& \le \left(\max_{\substack{\mathcal{M} \subseteq \reals^d:\\ \dim(\mathcal{M}) =i}}\min_{\substack{x \in \cM: \\x \neq 0}} \frac{ x^\top P^{-1/2} (C+\lambda I) P^{-1/2} x}{\|P^{-1/2}x\|^2}\right) \times\\
& \qquad \left( \max_{\substack{x \in \reals^d:\\ x \neq 0}} \frac{x^\top P^{-1} x}{\|x\|^2} \right)\\
&= \lambda_i \left(C+\lambda I \right) \cdot \lambda_1(P^{-1}) = (\lambda_i+\lambda) \cdot (\tilde{\sigma}_k^2+\lambda)^{-1} ~.
$$ Finally, using we have that ${\tilde{\sigma}}_k^2 \ge \sigma_k^2 - \frac{1}{2} \sigma_{k+1}^2 \ge \tfrac{1}{2} \sigma_k^2 = \tfrac{1}{2} \lambda_k$ and therefore, $$(\tilde{\sigma}_k^2+\lambda)^{-1} ~\le~ (\tfrac{1}{2} \lambda_k +\lambda)^{-1} ~\le~ 2\, (\lambda_k+\lambda)^{-1} ~. \qedhere$$
\[lem:lowerSmall\] **(Lower bound on the smallest eigenvalue)** $$\lambda_d( P^{-1/2} C P^{-1/2}) \ge \frac{\lambda}{19 (\lambda_k+\lambda)}~.$$
Note that $$\label{eq:digging0}
\lambda_d (P^{-1/2} (C+\lambda I) P^{-1/2}) = \frac{1}{\|P^{1/2} (C+\lambda I)^{-1} P^{1/2}\|}~,$$ so we can derive an upper bound on $\|P^{1/2} (C+\lambda I)^{-1} P^{1/2}\|$. Consider an arbitrary completion of $\tilde{v}_1,\ldots,\tilde{v}_k$ to an orthonormal set, $\tilde{v}_1,\ldots,\tilde{v}_d \in \reals^n$. Let $\tilde{V} \in \reals^{n \times d}$ be the matrix whose $i$-th column is ${\tilde{v}}_i$. Since the spectral norm is unitary invariant and both ${\tilde{U}}$ and ${\tilde{V}}$ have orthonormal columns, $$\begin{aligned}
&\|P^{1/2} (C+\lambda I)^{-1} P^{1/2}\| \notag\\
&= \|{\tilde{U}}({\tilde{\Sigma}}^2+\lambda I)^{1/2} {\tilde{U}}^\top (C+\lambda I)^{-1} {\tilde{U}}({\tilde{\Sigma}}^2+\lambda I)^{1/2} {\tilde{U}}^\top\| \notag \\
&= \|{\tilde{V}}({\tilde{\Sigma}}^2+\lambda I)^{1/2} {\tilde{U}}^\top (C+\lambda I)^{-1} {\tilde{U}}({\tilde{\Sigma}}^2+\lambda I)^{1/2} {\tilde{V}}^\top\| ~.\end{aligned}$$ Denote by ${\tilde{Z}}= {\tilde{U}}({\tilde{\Sigma}}^2+\lambda I)^{1/2} {\tilde{V}}^\top$. By the triangle inequality and the submutiplicativity of the spectral norm, $$\begin{aligned}
&\|{\tilde{Z}}^\top(C+\lambda I)^{-1} {\tilde{Z}}\| \le \|\bar{X}^\top (C+\lambda I)^{-1} \bar{X}\| \notag\\
&\qquad + \|({\tilde{Z}}-\bar{X})^\top (C+\lambda I)^{-1} ({\tilde{Z}}-\bar{X})\| \notag \\
& \le \|\bar{X}^\top (C+\lambda I)^{-1} \bar{X}\|+ \|{\tilde{Z}}-\bar{X}\|^2 \|(C+\lambda I)^{-1}\|~. \label{eq:digging15}\end{aligned}$$ To bound the first summand of , we use the unitary invariance to obtain $$\begin{aligned}
&\|\bar{X}^\top (C+\lambda I)^{-1} \bar{X} \| = \|V \Sigma U^\top U (\Sigma^2 +\lambda I)^{-1} U^\top U \Sigma V^\top\| \notag\\
&\quad = \|\Sigma (\Sigma^2 +\lambda I)^{-1} \Sigma \| = \max_i \frac{\lambda_i^2}{\lambda_i^2+\lambda} \le 1~.\end{aligned}$$ For the second summand of , note that $\|(C+ \lambda I)^{-1}\|= \frac{1}{\lambda_d+\lambda}$ and that, using the triangle inequality, $$\begin{aligned}
&\|{\tilde{Z}}- \bar{X}\| = \|({\tilde{U}}{\tilde{\Sigma}}{\tilde{V}}^\top - \bar{X}) + ({\tilde{Z}}- {\tilde{U}}{\tilde{\Sigma}}{\tilde{V}}^\top) \|\\
&\quad \le \|{\tilde{U}}{\tilde{\Sigma}}{\tilde{V}}^\top - \bar{X}\| + \|{\tilde{U}}(({\tilde{\Sigma}}^2+\lambda I)^{1/2}- {\tilde{\Sigma}}) {\tilde{V}}^\top\|~.\end{aligned}$$ By using unitary invariance together with the inequality $\sqrt{{\tilde{\sigma}}_i^2+\lambda} - {\tilde{\sigma}}_i \le \sqrt{\lambda}$ (which holds for every $i$), we get $$\|{\tilde{U}}(({\tilde{\Sigma}}^2+\lambda I)^{1/2}- {\tilde{\Sigma}}) {\tilde{V}}^\top\| = \|({\tilde{\Sigma}}^2+\lambda I)^{1/2}- {\tilde{\Sigma}}\| \le \sqrt{\lambda}~.$$ Hence, using the inequality $(x+y)^2 \le 2x^2 + 2y^2$, we obtain $$\begin{aligned}
\|{\tilde{Z}}- \bar{X}\|^2 &\le 2 \|{\tilde{U}}{\tilde{\Sigma}}{\tilde{V}}^\top - \bar{X}\|^2 + 2\lambda ~.\end{aligned}$$ We next derive an upper bound on $\|{\tilde{U}}{\tilde{\Sigma}}{\tilde{V}}^\top - \bar{X}\|$. Since ${\tilde{U}}{\tilde{\Sigma}}{\tilde{V}}^\top = \tilde{X}_k + {\tilde{\sigma}}_k \sum_{i=k+1}^d {\tilde{u}}_i {\tilde{v}}_i^\top$, $$\|{\tilde{U}}{\tilde{\Sigma}}{\tilde{V}}^\top-\bar{X}\| \le \|\tilde{X}_k - \bar{X}\|+{\tilde{\sigma}}_k \left\|\sum_{i=k+1}^d {\tilde{u}}_i {\tilde{v}}_i^\top\right\| ~.$$ Using we know that $\|\tilde{X}_k - \bar{X}\| \le 1.5 \, \sigma_k$ and that ${\tilde{\sigma}}_k \le \sqrt{\sigma_k^2+0.5 \, \sigma_{k+1}^2} \le 1.5 \, \sigma_k$. Combining this with the fact that $\|\sum_{i=k+1}^d {\tilde{u}}_i {\tilde{v}}_i^\top\| = 1$, we obtain $$\begin{aligned}
\|{\tilde{U}}{\tilde{\Sigma}}{\tilde{V}}^\top-\bar{X}\| \le 3\,\sigma_k ~.\end{aligned}$$ Combining the above inequalities, we obtain $$\begin{aligned}
\|P^{1/2} (C+\lambda I)^{-1} P^{1/2}\|
&\le 1 + \frac{2 \cdot(3 \sigma_k)^2+2\lambda}{\lambda_d+\lambda} \\
&\le \frac{19(\lambda_k+\lambda)}{\lambda} ~,\end{aligned}$$ and using we conclude our proof.
**(of )** The three last lemmas imply that the inequalities derived during the proof of remain intact up to a multiplicative constant. Therefore, the bound on the condition number also holds up to a multiplicative constant. This completes the proof.
Sketched Preconditioned SVRG
----------------------------
By equipping SVRG with the sketched preconditioner , we obtain the Sketched Preconditioned SVRG (see ).
**(of )** The theorem follows from and .
**Input:** $x_1,\ldots, x_n \in \reals^d, y_1,\ldots,y_n \in \reals, \epsilon>0$ **Parameters:** $\lambda>0, k \in [d]$ Let $\bar{X} \in \reals^{d,n}$ be the matrix whose $i$’th column is $(1/n) x_i$ Run the Block Lanczos method () with the input $\bar{X}, k, \epsilon'=1/2$ to obtain ${\tilde{X}}_k = {\tilde{U}}_k {\tilde{\Sigma}}_k {\tilde{V}}_k$ Let ${\tilde{u}}_i$ be the columns of ${\tilde{U}}_k$ and $\tilde{\sigma}_i$ be the diagonal elements of ${\tilde{\Sigma}}_k$ Form the preconditioner $P^{-1/2}$ according to Compute ${\tilde{x}}_i = P^{-1/2} x_i, b_i = P^{-1/2} e_i$ Let $\ell_i(w) = \frac{n+d}{n} \frac{1}{2} (w^\top {\tilde{x}}_i-y_i)^2$ for $i=1,\ldots,n$ and $\ell_i(w) = \lambda(n+d) (w^\top b_i)^2$ for $i=n+1,\ldots,n+d$ Let $\beta_i = \frac{n+d}{n} \|{\tilde{x}}_i\|^2$ for $i=1,\ldots,n$ and $\beta_i = \lambda(n+d) \|b_i\|$ for $i=n+1,\ldots,n+d$. Let $\hat{\beta} = \frac{1}{n} \sum_{i=1}^{n+d} \beta_i$ Run SVRG () Return $\hat{w} = P^{1/2} \tilde{w}$
The Empirical Gain of Sketched Preconditioning {#sec:empirical}
==============================================
\[b\] [0.23 ]{}
\[b\] [0.23]{}
\[b\] [0.23]{}
\[b\] [0.23]{}
[0.23 ]{}
[0.23 ]{}
[0.23 ]{}
[0.23 ]{}
[0.23 ]{}
[0.23 ]{}
In this section we empirically demonstrate the gain of our method. We consider both regression problems and binary classifications tasks, where the square loss serves as a surrogate for the zero-one loss. We use the following datasets:
- *Synthetic*: We draw two random $5000 \times 20000$ matrices, $X^{(1)}$ and $X^{(2)}$, whose singular vectors are drawn uniformly at random and the $q$-th singular value is $1/q$ and $1/q^2$, respectively. We then normalize the columns. For each $X=X^{(j)}$, we consider a regression problem, where the labels are generated as follows: we first draw a vector $w^\star \in \cN(0,1)^{5000}$ and then set $y_i = {w^\star}^\top X_{\cdot,i}+z_i$, where $z_i \sim \cN(0,0.1)$.
- *MNIST*:[^5] A subset of MNIST, corresponding to the digits $4$ and $7$, where the task is to distinguish between the two digits. Here, $n=12107, d=784$.
- *RCV1*:[^6] The Reuters RCV1 collection. Here, $n=20242, d=47236$ and we consider a standard binary document classification task.
- *CIFAR-10*:[^7] Here, $n=50000, d=3072$. Following [@frostig2015regularizing], the classification task is to distinguish between the animal categories to the automotive ones.
- *real-sim*:[^8] Here, $n=72309, d=20958$, and we consider a standard binary document classification task.
Inspecting our theoretical speed-up {#sec:ratioEmp}
-----------------------------------
Recall that the ratio quantifies our theoretical speedup. Hence, we first empirically inspect the prefixes of the corresponding quantities (as a function of $k$) for each of the datasets (see ). We can see that while in MNIST and CIFAR-10 the ratio is large for small values of $k$, in RCV1 and real-sim the ratio increases very slowly (note that for the former two datasets we use logarithmic scale).
Empirical advantage of Sketched Preconditioned SVRG {#sec:emp}
---------------------------------------------------
We now evaluate and compare it to the SVRG algorithm of [@xiao2014proximal]. To minimally affect the inherent condition number, we added only a slight amount of regularization, namely, $\lambda=10^{-8}$. The loss used is the square loss. The step size, $\eta$, is optimally tuned for each method. Similarly to previous work on SVRG [@xiao2014proximal; @johnson2013accelerating], the size of each epoch, $m$, is proportional to the number of points, $n$. We minimally preprocessed the data by average normalization: each instance vector is divided by the average $\ell_2$-norm of the instances. The number of epochs is up to $60$. Note that in all cases we choose a small preconditioning parameter, namely $k=30$, so that the preprocessing time of is negligible. There is a clear correspondence between the ratios depicted in and the actual speedup. In other words, the empirical results strongly affirm our theoretical results.
Acknowledgments {#acknowledgments .unnumbered}
===============
We thank Edo Liberty for helpful discussions. The work is supported by ICRI-CI and by the European Research Council (TheoryDL project).
Omitted Proofs
==============
**(of )** We first show that the average smoothness of $\tilde{L}$ is bounded by $$\label{eq:condEffectSmooth}
\frac{1}{n+d} \sum_{i=1}^{n+d} \tilde{\beta}_i \le \tr \left(P^{-1/2} \left(C + \lambda I \right) P^{-1/2} \right)~.$$ Note that for any $w$, $$\nabla^2 \tilde{\ell}_i(w) = \begin{cases} \frac{n+d}{n}{\tilde{x}}_i {\tilde{x}}_i^\top & 1 \le i \le n, \\ \lambda (n+d) b_{i-n} b_{i-n}^\top & n < i \le n+d~. \end{cases}$$ Therefore, using the fact that the spectral norm of a rank-$1$ psd matrix is equal to its trace, we obtain $$\begin{aligned}
\frac{1}{n+d} \sum_{i=1}^n \tilde{\beta}_i &= \frac{1}{n+d} \frac{n+d}{n} \sum_{i=1}^n \|{\tilde{x}}_i {\tilde{x}}_i^\top\| + \frac{1}{n+d}\lambda (n+d) \sum_{j=1}^d \|b_j b_j^\top\| \\
& = \frac{1}{n} \sum_{i=1}^n \tr({\tilde{x}}_i {\tilde{x}}_i^\top) + \lambda \sum_{j=1}^d \tr(b_i b_i^\top) \\
&= \frac{1}{n}\tr( \sum_{i=1}^n P^{-1/2} x_i x_i^\top P^{-1/2}) + \lambda \, \tr \sum_{j=1}^d (P^{-1/2} e_i e_i^\top P^{-1/2}) \\
&=\tr(P^{-1/2} ( C + \lambda I)P^{-1/2})~.\end{aligned}$$ Hence, we deduce (\[eq:condEffectSmooth\]).
We will conclude the theorem by showing that $\tilde{L}$ is $\lambda_d (P^{-1/2} (C + \lambda I) P^{-1/2})$-strongly convex. Indeed, a similar calculation shows that the Hessian of $L$ at any point $w$ is given by $$\begin{aligned}
\nabla^2 \tilde{L}(w) &= P^{-1/2} (C+ \lambda I) P^{-1/2} ~.\end{aligned}$$ Hence, we conclude the claimed bound.
[^1]: School of Computer Science, The Hebrew University, Jerusalem, Israel
[^2]: Yahoo Labs, New York, NY, USA
[^3]: School of Computer Science, The Hebrew University, Jerusalem, Israel
[^4]: More precisely, Algorithm 2 in [@musco2015stronger] returns the projection matrix ${\tilde{U}}_k {\tilde{U}}_k^\top$, while we also compute the SVD of ${\tilde{U}}_k {\tilde{U}}_k^\top A$. The additional runtime is negligible.
[^5]: http://yann.lecun.com/exdb/mnist/
[^6]: https://www.csie.ntu.edu.tw/ cjlin/libsvmtools/datasets/
[^7]: http://www.cs.toronto.edu/ kriz/cifar.html
[^8]: https://www.csie.ntu.edu.tw/ cjlin/libsvmtools/datasets/
|
{
"pile_set_name": "ArXiv"
}
|
Q:
Jinja2: How to use named blocks inside included templates, inside extendable template
I am having the Issue with Jinja2 Extend and Import.
base_admin.html
<html>
<body>
<div class="outerbody">
<somehtml code>
{% include "base_admin_nav.html" %}
{% include "base_admin_sidebar.html" %}
{% include "base_admin_content.html" %}
</div>
</body>
</html>
base_admin_content.html
<div class="innerbody">
{% block body_content %}
{% endblock %}
</div>
admin.html
{% extends 'base_admin.html' %}
{% block body_content %}
<div>BodyContent</div>
{% endblock %}
The code inside body_content is not passed to base_admin_content.html. Any workarounds?
Note
This is not duplicate of this one
jinja2: blocks in included files..
The include is done in different files here
Defining {% macro admin_content() %} insdide base_admin_content.html and importing it inside base_admin.html using
{% from "base_admin_content.html" import admin_content with context %}
{{ admin_content() }}.
also has no effect.
A:
Edited - to reflect changes in original question
Ok, now that I know you definitely need the includes, here's how I would do it: instead of including the base_admin_content.html file, you should include the admin.html file directly into base_admin.html. The admin.html file will extend base_admin_content.html and everything should work just fine:
base_admin.html
<html>
<body>
<div class="outerbody">
<somehtml code>
{% include 'admin.html' %}
</div>
</body>
</html>
admin.html
{% extends 'base_admin_content.html' %}
{% block body_content %}
<div>BodyContent</div>
{% endblock %}
base_admin_content.html
{% block innerbody %}
<div class="innerbody">
{% block body_content %}
{% endblock %}
</div>
{% endblock %}
Why does this work but your original code does not?
In your base_admin.html file you have
{% include 'base_admin_content.html' %}
Where we have no reference to admin.html, which is why nothing from the admin.html file shows up!. Therefore, we should do this:
{% include 'admin.html' %}
Because that does contain a reference to base_admin_content in the extends line:
{% extends 'base_admin_content.html' %}
Hopefully that makes sense...
|
{
"pile_set_name": "StackExchange"
}
|
Histological and ultrastructural localization of antigen B in the metacestode of Taenia solium.
The morphological localization of antigen B (AgB) in the tissues of the Taenia solium metacestode was studied by immunological and biochemical methods. Indirect immunofluorescence carried out on vibratome sections showed that AgB is widely distributed throughout the tissue. A more intense fluorescence was observed in the tegumentary cytons of the bladder wall and in the lumen of the spiral canal of the invaginated scolex. Ultrastructural analysis of larvae washed in PBS after dissection from meat and then incubated with rabbit antibodies against AgB, followed by peroxidase-labeled goat anti-rabbit IgG, did not exhibit electron-dense material on the external surface. Larvae fixed in glutaraldehyde immediately after dissection and exposed to the immunoperoxidase reagents did exhibit electron-dense material on microtriches, indicating that AgB is only loosely bound to the external surface. Crude extracts of surface-radioiodinated cysticerci analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) contained no labeled proteins with the molecular weight of AgB. Autoradiography of the immunoelectrophoretograms in which the crude extract was confronted with antibodies to AgB demonstrated that this antigen was not labeled, and therefore is not exposed on the tegumentary surface. The results suggest that AgB is synthesized by the tegumentary cytons of the parasite and secreted through the tegumental membrane into the host tissues and the lumen of the spiral canal.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Introduction {#s1}
============
> If you come to any more conclusions about polymorphism, I should be very glad to hear the result: it is delightful to have many points fermenting in one\'s brain, and your letters and conclusions always give one plenty of this same fermentation.
>
> \- Charles Darwin, letter to Joseph Hooker, 1846 [@pone.0093622-Darwin1]
Polymorphism has fascinated evolutionary biologists since the time of Darwin [@pone.0093622-Darwin2], [@pone.0093622-Darwin3]. Polymorphism is defined as that there exist more than one distinct phenotype of a species occupying the same habitat at the same time [@pone.0093622-Ford1], [@pone.0093622-Smith1]. Polymorphism does not include continuous variations, but only discrete variations or in the case of continuous traits, such as body size and color, strongly bimodal or multimodal phenotype variation distributions. The existence of more than one distinct phenotype of a species demands an explanation, because the theory of natural selection predicts that the fittest phenotype should drive the other, lesser fit phenotypes to extinction. In general, polymorphism is maintained if the "fitness curves" of the polymorphic phenotypes intersect, where the crossover-point is an evolutionarily stable state, realizing a polymorphic evolutionarily stable strategy (ESS) [@pone.0093622-Smith2], [@pone.0093622-Smith3]. Common features of the evolution and the maintenance of behavioral polymorphism include: 1) that time or resources can be invested in more than one activity that contributes to the fitness; 2) that the individuals have rules about how to allocate time and resources among the alternative activities; and 3) that there is a trade-off between the activities that contribute to the fitness, i.e., the allocation of time and resources invested in one activity could be invested in another [@pone.0093622-Brockmann1]. Frequency-dependent selection [@pone.0093622-Smith3], [@pone.0093622-Anderson1], [@pone.0093622-Gadgil1] is considered the most important explanation for the maintenance of polymorphism in a population. Frequency-dependent selection occurs when the fitness of the phenotypes depends on their frequencies in the population, and the fitness curves intersect at a crossover frequency where the phenotypes are equally successful. Alternative mating strategies (or alternative reproductive behaviors) [@pone.0093622-Rubenstein1]--[@pone.0093622-Brockmann2] is the area of biological research most closely related to this study. Different mating tactics has been observed in a wide variety of species, both in males (e.g., [@pone.0093622-Bisazza1]--[@pone.0093622-Utami1]) and in females (e.g., [@pone.0093622-Roulin1]--[@pone.0093622-Svensson1]). However, there are relatively few cases where the differences in mating behavior have been confirmed to have a genetic basis [@pone.0093622-Zimmerer1]--[@pone.0093622-Sinervo1], and even fewer studies that have suggested equal average reproductive success, i.e., shown crossing of the fitness curves, of alternative phenotypes [@pone.0093622-Shuster1], [@pone.0093622-Ryan1], [@pone.0093622-Sinervo2].
The use of robot evolution experiments to study biological phenomena has gained traction in recent years [@pone.0093622-Floreano1], as a complementary approach to biological studies and theoretical models. In comparison to biological studies, robot evolution has the advantage that the evolution of hundreds of generations of robot controllers can be completed within hours or days. The experiments can easily be repeated for different parameter settings and environmental conditions, which allows for quantitative testing and analysis of robustness and stability. In comparison with theoretical and numerical models, robot models can capture the often complex physical interactions between the agent and the environment, including other agents. Floreano and Keller with different co-authors have used robot evolution experiments to investigate the emergence and reliability of communication [@pone.0093622-Floreano2]--[@pone.0093622-Wischmann1], to quantitatively test Hamilton\'s rule for the evolution of altruism [@pone.0093622-Waibel1], and to test the influence of genetic architecture and mating frequency on the division of labor in social insect societies [@pone.0093622-Tarapore1].
A distinctive feature of our earlier proposed embodied evolution framework [@pone.0093622-Elfwing1] is that there is no explicit fitness function or algorithm for selecting individuals for recombination and mutations. Instead, offspring can only be created by the physical exchange of genotypes between two mating robots. In general, the choice of selection method requires careful consideration when using artificial evolution to study ESSs. A strong theoretical assumption underlying ESS analysis is that the population is infinitely large. Fogel *et al.* [@pone.0093622-Fogel1]--[@pone.0093622-Fogel3] demonstrated in simulation experiments, using the Hawk-Dove game, that for finite populations the results differed, at best, significantly from the theoretical ESS values and, at worst, bore no resemblance to the ESS. They, therefore, questioned the usefulness of ESSs to explain real biological phenomena in populations with limited population sizes. In response, Ficici and Pollock [@pone.0093622-Ficici1] showed that the difference between the theoretical ESS predictions and the observed simulation results could be accounted by the two selection methods used by Fogel *et al.*.
The purpose of this study is to demonstrate that evolutionary stable alternative mating strategies can emerge in a small robot colony without any predefined mating preferences as a result of the trade-off between the resources spent on energy conservation and the resources spent on courtship of mating partners. As alternative mating strategies is a natural precursor for the evolution of sexual dimorphism, this line of research has the potential of increasing our understanding of the emergence of different sexes.
To investigate the ecological conditions for evolution of alternative mating strategies, we performed artificial evolution experiments, in simulation, with a small colony of Cyber Rodent robots [@pone.0093622-Doya1] using our proposed embodied evolution framework [@pone.0093622-Elfwing1]. We performed the experiments in simulation because of the infeasibility of running hundreds of generations of evolution in hardware. In previous work [@pone.0093622-Elfwing1], we have shown that learned and evolved behaviors in simulation have similar performance and behavior when transferred to the hardware setting. Each individual interacted in small groups of four robots and during its lifetime of 288 seconds an individual experienced three periods of group interactions, where the participants in each group were randomly selected. We placed the four robots in an arena (2.5×2.5 m) with 4 to 16 energy sources. The robots were equipped with two wheels, an infrared port for the exchange of genotypes, and a camera that could detect energy sources, and tail-lamps and faces of other robots ([Figure 1](#pone-0093622-g001){ref-type="fig"}). The robots could execute three basic behaviors, foraging, waiting (for a potential mating partner), and mating, which were learned by reinforcement learning [@pone.0093622-Sutton1]. The mating strategy, i.e., the selection of basic behaviors, was controlled by a linear neural network and the (five) neural network weights were adapted by the evolutionary process ([Figure 2](#pone-0093622-g002){ref-type="fig"}). From a biological point of view, the population consisted of simultaneous hermaphrodites, who could reproduce offspring by mating (i.e., an exchange of genotypes with a mating partner). For each of the individuals involved in a mating event, the probability of reproducing offspring was linearly dependent on the individual\'s internal energy level (see [Methods](#s4){ref-type="sec"} and [@pone.0093622-Elfwing1]). This created a trade-off, where, in relative terms, an individual could maximize its fitness by maximizing either the frequency of mating events or the energy level at the mating events.
{#pone-0093622-g001}
{#pone-0093622-g002}
Results {#s2}
=======
After the evolutionary process converged after 1,000 generations of experiments, we frequently observed a variety in the mating behavior when a robot could see both an energy source and the tail-lamp of a potential mating partner. We classified their mating strategies into two types: 1) *Forager strategy* in which an individual never waited for a potential mating partner and only tried to mate if it saw the face of another robot, and 2) *Tracker strategy* in which an individual waited for potential mating partner to turn around and where the threshold for waiting depended on its current energy level, the distance to closest energy source, and the distance to tail-lamp of the closest potential mating partner (in our preliminary report [@pone.0093622-Elfwing2], we called them *roamer* and *stayer*, respectively, borrowing the terminology by Sandell and Liberg [@pone.0093622-Sandell1]). Among the high fitness populations (5 highest fitness populations in each of the 7 environments with different levels of energy source density), 31% (11/35) consisted predominantly of foragers and 49% (17/35) consisted predominantly of trackers. In 20% (7/35) of the populations, and remarkable four of the seven highest fitness populations, there emerged a polymorphic population of foragers and trackers with distinct differences in genotype, phenotype, and behavior. Our analyses show that the polymorphic population could constitute an ESS with an evolutionarily stable state of approximately 25% foragers in the population. In one instance, the trackers were split into two subpopulations: trackers who almost always waited for potential mating partners (strong trackers), and trackers who only waited if the energy level was high and an energy source was close (weak trackers). The analyses show that a population consisting of three phenotypes (foragers, strong trackers, and weak trackers) also could constitute a globally stable polymorphic ESS with several attractors.
The basic behaviors (foraging, waiting, and mating) of each individual were learned from scratch in each generation by reinforcement learning and accelerated by the evolution of the shaping rewards. For a detailed description and analysis of the evolution of shaping rewards and meta-parameters in our embodied evolution framework, see [@pone.0093622-Elfwing1]. The fitness measured by the number of reproduced offspring correlated significantly with the learning performance of the mating behavior ( and ; [Figure 3](#pone-0093622-g003){ref-type="fig"}). The efficiency of the mating behavior depended strongly on the evolved shaping reward functions. On average, one time step (240 ms) reduction in the time to execute a successful mating lead to an increase in fitness by approximately one offspring.
{#pone-0093622-g003}
The purpose of performing experiments with different number of energy sources in the environment was to investigate the effect of energy source density on the emergence of different mating strategies. In general, higher energy source density resulted in higher energy levels at the mating events, which increased the probability of reproducing offspring and, therefore, the fitness (see [Figure 3](#pone-0093622-g003){ref-type="fig"}). However, we could not find an effect of energy source density on the evolution of different mating strategies.
Analysis of Mating Strategies {#s2a}
-----------------------------
Analyses of the behaviors in the final 20 generations of the evolutionary process show that almost all individuals in all experiments executed an opportunistic behavior when either only an energy source was visible or only a face of another robot was visible. In the former situation all individuals always executed the foraging behavior and in the latter situation they always executed the mating behavior. An interesting case happened when a robot could see both an energy source and the tail-lamp of another robot from its behind. While some robots (i.e., foragers) chose to take the foraging behavior of approaching the energy source, others (i.e., trackers) took the mating behavior of keeping track of the tail-lamp and approaching when the face became visible, as illustrated in [Figure 4](#pone-0093622-g004){ref-type="fig"}. The choice in general depended on the robot\'s stored energy level and the distances to the energy source and the tail-lamp.
{#pone-0093622-g004}
In order to characterize these behavioral phenotypes, we took the average energy level threshold for the mating behavior, by computing the mean value over 676 equidistant visual inputs of the energy source and the tail-lamp (see [Methods](#s4){ref-type="sec"}). With this criterion, the individual mating strategies emerged in the experiments could be classified: *Forager strategy*, in which an individual always took foraging behavior with equal or very close to the maximum threshold (1), and 2) *Tracker strategy*, in which an individual waited for the potential mating partners to turn around when its energy level is above a certain threshold, with smaller than 1.
We identified the three different types of populations by computing the median and the standard deviation of the average threshold for all 1600 individuals in the last 20 generations in each of the 70 experiments. The populations with median values larger than 0.98 were considered forager populations. The remaining populations were identified by their standard deviations. Those with standard deviations smaller than 0.19 were considered tracker populations and those with standard deviations larger than 0.19 were considered polymorphic populations, all of which had multimodal distributions.
We found 33, 28, and 9 populations of forager, tracker, and polymorphic strategies, respectively in the 70 populations. If we focus on the higher fitness populations (5 highest fitness populations in each of the 7 environments with different levels of energy source density), 31% (11/35) were foragers, 49% (17/35) were trackers, and 20% (7/35) were polymorphic. Remarkably, four of the seven highest fitness populations were polymorphic.
Analysis of Genotypes, Phenotypes, and Behaviors {#s2b}
------------------------------------------------
The best example of an emerged polymorphic mating strategy is shown in [Figures 5](#pone-0093622-g005){ref-type="fig"}--[7](#pone-0093622-g007){ref-type="fig"}, which emerged in the experiment with the highest average fitness in the environment with four energy sources. [Figure 5](#pone-0093622-g005){ref-type="fig"} shows the distribution of the values of , the bias for the mating behavior, and , the weight for the distance to another robot\'s face, of all 1600 individuals in the final 20 generation. We classified the population into the two subpopulations by *k*-means clustering in the and weight space and show the forager subpopulation in black and the tracker subpopulation in red in this and the following figures. [Figure 6](#pone-0093622-g006){ref-type="fig"} shows the histogram of average energy thresholds of selection of the mating behavior. The genetically classified foragers and trackers formed clearly distinct distributions. [Figure 7](#pone-0093622-g007){ref-type="fig"} shows the average percentage of the lifetimes executing the three basic behaviors of foraging, waiting (choice of mating despite the face is not visible), and mating.
{#pone-0093622-g005}
{#pone-0093622-g006}
{#pone-0093622-g007}
The result clearly shows a polymorphic population with distinct bimodal distributions in genotype, phenotype, and behavior for a forager subpopulation and a tracker subpopulation. For the foragers, the median of average energy threshold was equal the maximum threshold of 1. They spent on average 77.5% of their lifetimes executing the foraging behavior and almost no time (0.6%) executing the waiting behavior. In comparison, for the trackers the median of threshold was 0.40. They spent on average 57.6% of their lifetimes executing the foraging behavior and 21.5% executing the waiting behavior.
Evolutionary Stability of Polymorphism {#s2c}
--------------------------------------
To investigate the evolutionarily stability of the emerged polymorphic population of foragers and trackers, we performed additional experiments in which we fixed the proportion of the two phenotypes in the polymorphic population shown in [Figures 5](#pone-0093622-g005){ref-type="fig"}--[7](#pone-0093622-g007){ref-type="fig"}. Individuals from the two genotypes were selected randomly from the final (1000th) generation of the evolutionary experiment, which consisted of 10 foragers and 70 trackers. The experiment was repeated 100 times for each proportion of the two phenotypes. [Figure 8](#pone-0093622-g008){ref-type="fig"} shows the average number of mating events, [Figure 9](#pone-0093622-g009){ref-type="fig"} shows the average normalized energy level at the mating events, and [Figure 10](#pone-0093622-g010){ref-type="fig"} shows the average number of offspring (i.e., the fitness) for the two phenotypes, as functions of the tracker proportion in the population. The number of mating events increased, both for the population as a whole and for the two phenotypes, as the number of trackers in the population increased. The increase in the number of mating events was much larger for the foragers (black line in [Figure 8](#pone-0093622-g008){ref-type="fig"}), from 10 to almost 13 mating events, as the tracker proportion increased from 0% to 87.5%. In comparison, the number of mating events for the trackers (red line in [Figure 8](#pone-0093622-g008){ref-type="fig"}) increased with only approximately 1 mating event, from 12 to 13, as the tracker proportion increased from 12.5% to 100%. The average energy levels were approximately constant over all phenotype proportions for both foragers and trackers (dotted lines in [Figure 9](#pone-0093622-g009){ref-type="fig"}) with a mean value of 0.81 for the foragers and 0.77 for the trackers. The experimental result clearly show ([Figure 10](#pone-0093622-g010){ref-type="fig"}) that the emerged population of foragers and trackers constitute a polymorphic ESS with an evolutionarily stable state of around 25% foragers in the population. The stable state corresponded to group-interactions with only one forager in the environment where the three trackers would all wait and adjust their trajectories according to the forager\'s behavior.
{#pone-0093622-g008}
{#pone-0093622-g009}
{ref-type="disp-formula"} and [9](#pone.0093622.e056){ref-type="disp-formula"}.](pone.0093622.g010){#pone-0093622-g010}
The foragers demonstrated negative frequency-dependent selection, i.e., the fitness increased as they became rarer. The foragers could be considered as a parasitic phenotype, because they relied on a high proportion of trackers to achieve a large number of mating events and, thereby, high fitness. In relative terms, the foragers tried to maximize the second term of the fitness function ([Equation 1](#pone.0093622.e027){ref-type="disp-formula"}) by maximizing their own energy at the mating events, . The trackers demonstrated positive frequency-dependent selection, i.e., the fitness increased as they became more common. However, the proportion of trackers in population had much less effect on the trackers\' fitness. The average fitness increased only about 8% as the tracker proportion increased from 25% to 100%, compared to the foragers that increased their average fitness 21% as the tracker proportion increased from 0% to 75%. In relative terms, the trackers maximized the first term of the fitness function, the mating frequency , by using the waiting behavior to attract potential mating partners. The result also shows that the evolved mating strategy could have a significant impact on the average population fitness. There was an almost 4 offspring increase in fitness between a population with only foragers (fitness of about 16 offspring) and a population with only trackers or an evolutionarily stable polymorphic population (fitness of about 20).
Mating Dynamics of Foragers and Trackers {#s2d}
----------------------------------------
Now let us investigate the dynamic interactions of the two subpopulations behind the observed fitness curves in [Figure 10](#pone-0093622-g010){ref-type="fig"}.
The average fitness , i.e., the number of reproduced offspring, is a function of the average number of mating events, , the average normalized energy level at the mating occasions (for offspring created by the individual), , and the mating partners\' average normalized energy level (for offspring created by the mating partners), :
In order to model the average fitness, we take the following hypotheses: H1) The trackers achieved almost perfect mixing of mating partners. H2) The foragers mated more frequently with trackers than the proportion of trackers in the population. H3) The mating frequencies of the foragers, , and the trackers, , were functions of the tracker proportion . Under these conditions, we predict the fitness curves of the two subpopulations.
Let us denote the average numbers of mating events in a lifetime of a forager with foragers by , with trackers by , and with either by . We also denote the average numbers of mating events of a tracker , , and in the same convention.
H1 can be represented asThis was true for all tested phenotype proportions, as shown by the red line in [Figure 11](#pone-0093622-g011){ref-type="fig"}.
{ref-type="disp-formula"} with .](pone.0093622.g011){#pone-0093622-g011}
H2 means thatwith a positive . This was the case in our experiment with , as shown by the black line in [Figure 11](#pone-0093622-g011){ref-type="fig"}.
From H1 ([Equation 2](#pone.0093622.e038){ref-type="disp-formula"}) and the symmetry condition , we have . From this and H2 ([Equation 3](#pone.0093622.e040){ref-type="disp-formula"}), the ratio of the average numbers of mating events by foragers and trackers () can then be derived as In our experiment, the ratio of forager mating events to tracker mating events increased from 0.88 to 1.02 when the tracker proportion increased from 0.125 to 0.875, as shown by the blue line in [Figure 11](#pone-0093622-g011){ref-type="fig"}. This fits well with the model prediction in [Equation 5](#pone.0093622.e047){ref-type="disp-formula"} with , shown by the dotted blue line in [Figure 11](#pone-0093622-g011){ref-type="fig"}.
From H3, the average fitness ([Equation 1](#pone.0093622.e027){ref-type="disp-formula"}) of the foragers, and the trackers, , can be rewritten as The dotted lines in [Figure 10](#pone-0093622-g010){ref-type="fig"} shows the estimation of the average fitness of the foragers and the trackers, using [Equations 7](#pone.0093622.e054){ref-type="disp-formula"} and [9](#pone.0093622.e056){ref-type="disp-formula"}, respectively, with constant -values, and with the best linear fit of from the data (see dotted lines in [Figures 9](#pone-0093622-g009){ref-type="fig"} and [8](#pone-0093622-g008){ref-type="fig"}, respectively).
Population with Three Mating Strategies {#s2e}
---------------------------------------
As shown in [Figure 6](#pone-0093622-g006){ref-type="fig"}, the distribution of the average energy threshold for waiting, , of all trackers in the final 20 generations was very broad. Interestingly, in the final generation, the trackers\' phenotype distribution was bimodal with two distinct peaks ([Figure 12](#pone-0093622-g012){ref-type="fig"}): a larger subpopulation (46 individuals) of *strong trackers* with a smaller median -value of approximately 0.22 and a smaller subpopulation (24 individuals) of *weak trackers* with a larger median -value of approximately 0.48. The separation in genotype between the strong and weak trackers was not as distinct, but it is clearly visible in and weight space ([Figure 13](#pone-0093622-g013){ref-type="fig"}).
{#pone-0093622-g012}
{#pone-0093622-g013}
To investigate the evolutionarily stability of a population consisting of foragers, weak trackers, and strong trackers, we conducted additional experiments in which we fixed the proportions of the three phenotypes and measured their fitnesses. As in the earlier experiments, the genotypes of the three subpopulations were randomly selected from the genotypes in the final generation of the evolutionary experiments and repeated 100 times for each proportion of the three phenotypes. The result of the experiments is summarized in the DiFinetti diagram in [Figure 14](#pone-0093622-g014){ref-type="fig"}. The proportion of each phenotype increases from the side of the triangle to the opposite vertex. The black circles represent the tested phenotype proportions. In general, the average population fitness (shown by the background coloring of the diagram) increased with the number of strong trackers in the population. The average population fitness was 16.2 offspring in population with only foragers, 18.2 offspring in population with only weak trackers, and 19.7 offspring in a population with only strong trackers, which was close to the maximum average fitness of 19.9 for a population consisting of 87.5% strong trackers and 12.5% foragers.
{#pone-0093622-g014}
Assuming a fixed population size, the proportion of phenotype *i* in the next generation, , can be calculated by the discrete replicator dynamic equation:where, is the average fitness of phenotype *i* with proportion of the population in the current generation. The black arrows in [Figure 14](#pone-0093622-g014){ref-type="fig"} show the average directions and magnitudes of the changes in phenotype proportions. Three populations consisting of only two of the phenotypes (as seen along the three sides of the triangle) were evolutionarily stable with populations proportions close to: 1) 31.25% foragers and 68.75% strong trackers (bottom side); 2) 87.5% weak trackers and 12.5% strong trackers (right side); and 3) 12.5% foragers and 87.5% weak trackers (left side). Populations consisting of all three phenotypes were globally stable, because all arrows in the inner triangle, representing populations consisting of at least 12.5% of each of the three phenotypes, point inside the triangle. The diagram suggests that there were two evolutionarily stable attractors where the magnitude of the change in the proportion of the phenotypes was zero, , with centers close to populations consisting of: 4) 25% foragers, 50% weak trackers, and 25% strong trackers () and 5) 25% foragers, 25% weak trackers, and 50% strong trackers (). For populations with 12.5% to 25% foragers, the average magnitudes of the change in phenotype proportions were small, i.e., low selection pressure, which meant that phenotype ratios could move relatively easily between the attractors by random genetic drift. For example, in the final 20 generations of the evolutionary experiment (illustrated by the gray line ending in the gray triangle), the phenotype ratios were first close to the fourth attractor (fitness of 19.1), and ended up close to the fifth attractor (fitness of 19.3).
Discussion {#s3}
==========
In this study, we demonstrated that polymorphic mating strategies can emerge in a small robot colony under homogeneous evolutionary conditions, without a selection scheme or an explicit fitness function that promoted a certain outcome. Our study is, to our knowledge, the first to demonstrate the emergence of polymorphic ESSs within a robot evolution framework. This gives further evidence that artificial robot evolution (for an overview see [@pone.0093622-Floreano1]) can be a feasible and a valuable approach for investigating hypotheses of biological phenomena.
The importance of specific details of the genetic algorithm and the structure of the genotype were illustrated in this study. A condition for the evolution of the polymorphic ESSs consisting of foragers and trackers was the small proportion of genotype controlling the mating strategy, in combination with the relatively low crossover rate. The mating strategy was controlled by only 5 out of the 51 genes, located at the beginning of the genotype, and with a crossover rate of 0.1 there was only a 0.8% probability that an offspring would have a mating strategy controlled be a mixture of genes from both parents. A much more frequent mixing of the mating strategy genes would have made it more difficult or even impossible to evolve and maintain separate genetic traits in the same population. An assumption underlying evolutionary game theory [@pone.0093622-Smith3], is that the payoffs that agents are assumed to be without noise. It is therefore very encouraging that evolutionarily stable polymorphic ESSs could emerge in our experiment with a small population size and with large variances in the performance of similar, and even identical, individuals. The lifetime learning of the basic behaviors by reinforcement learning introduced additional stochasticity. Even in the last part of the evolutionary process, a few individuals failed to capture any batteries or engage in any mating activity. This was usually caused by that the individual got trapped in a corner of the environment and failed to learn how to navigate out of it.
The forager strategy in the evolved polymorphic populations can be seen as a cheater strategy. To achieve high fitness, a forager relies on that all the other individuals in the environment (i.e., trackers) will adjust their behaviors according to the trajectory of the forager. The forager, therefore, avoids the cost of searching for mating opportunities. There exists a rich literature on the potential of cheating in hermaphrodite mating systems (for an overview see [@pone.0093622-Leonard1]). Usually, cheating refers to the attempt of individuals to take on the male role over the female role in mating encounters to avoid the cost offspring reproduction.
The most similar study to ours was conducted by Rold *et al.* [@pone.0093622-DaRold1]. They co-evolved a population of predefined male and female robots. The robots, as in our experiments, remained alive by capturing energy sources and reproduced by physical mating, consisting of touching a robot of the opposite sex. The only difference between males and females was that the males remained reproductive throughout their lives, while the females became non-reproductive for a fixed period of time after an reproductive mating event. In their experiment, the reproduced offspring were not the result of a genetic exchange between mating robots. Instead, the males and females were evolved separately with the number of reproductive mating events used as fitness objective. The evolved behaviors of the males and females had distinct differences and their behaviors corresponded to observed behaviors of males and females in biological studies. Interestingly, the evolved behaviors of the males and females also matched the behavior of the foragers and trackers, respectively, in our study. The males opportunistically ate all the food they could find while looking for reproductive females. The reproductive females were less active and adopted a mating strategy of waiting for males to mate with them. This give some support to a hypothesis that polymorphic mating strategies, emerged due to basic trade-off between the resources spent on energy conservation and the resources spent on courtship of mating partners, is a precursor of sexual dimorphism. In our experiments, polymorphism could arise because the foragers and trackers optimized, in relative terms, different parts of the fitness function ([Equation 1](#pone.0093622.e027){ref-type="disp-formula"}). The foragers maximized their own energy level, , by spending all their lives foraging for energy sources except for when a a potential mating partner was directly visible, while the trackers maximized the mating frequency, , by spending considerable amount of their lives waiting for potential mating partners. The evolution of "proto-sexes" is a research venue we plan to explore in future work. In the current experimental setup, both the sender and receiver can reproduce offspring at the mating events and the cost of mating is equal, whether offspring are reproduced or not (see [*Methods*](#s4){ref-type="sec"}). A more biological plausible setup would be that only one of the agents took on the female role, e.g., the receiver, and also bore the main cost of reproducing offspring. The goal would then be to investigate if, and in such case under which conditions, a breeding system with distinct male and female roles could evolve from an initial population of hermaphrodites without any predefined mating preferences, and maybe even more exotic breeding systems such as *androdioecy* (males and hermaphrodites) and *gynodioecy* (females and hermaphrodites).
Methods {#s4}
=======
Four Cyber Rodent mobile robots [@pone.0093622-Doya1] were placed in a 2.5×2.5 m arena, together with 4, 6, 8, 10, 12, 14, or 16 energy sources ([Figure 1](#pone-0093622-g001){ref-type="fig"}). The task of the robots were to survive by maintaining their internal energy level through foraging of energy sources and by reproduction of offspring through physical exchange of genotypes by infrared communication. We performed the experiments in a simulation environment, developed to mimic the features of the real Cyber Rodent hardware platform. The robots were equipped with a camera system with color blob detection, used to extract the distances and relative angles to nearest energy source (blue), the nearest tail-lamp of another robot (green), and the nearest face of another robot (red). Mimicking the real robotic hardware, the field of view of the simulated vision system set to . Within an angle range of , the robots could detect energy sources up to 2 m, tail-lamps up to 1.5 m, and faces up to 1 m. Outside this range, the detection capability decreased linearly down to 0.2 m for the maximum angles.
We performed 1000 generations of evolution and for each energy source density, we ran 10 evolution experiments. To be able to conduct robot evolution experiments with only only four robots, we utilized time-sharing in subpopulations of 20 individuals within each robot. Each individual in a subpopulation took control, in random order, of the robot for three time-sharings of 400 time steps, i.e., the total lifetime was 1200 time steps. An individual had a maximum internal energy level () of energy units. Each time step, the energy level decreased by unit and a capture of an energy source increased the energy level by units. At birth, an individual had an internal energy level of units. If an individual\'s energy was depleted, then the individual died and was removed from its subpopulation. When a robot captured an energy source it disappeared from its current position and reappeared in new, randomly selected, position.
We did not apply an explicit fitness function or a centralized selection process, instead offspring were created by a mating. The individuals controlling the robots could create offspring by a physical exchange of genotype through infrared communication. The infrared communication ports were located slightly to the right of center in the front of the Cyber Rodent robots, directed straight forward. In the simulation environment, the maximum range of the communication was set to 1 m and the angle range was set to . An individual could initiate the infrared communication by executing a predefined action selected by the mating behavior. For a mating event to be successful, both robots had to be within each others mating range before and after the individuals controlling the robots executed the actions of their currently selected reinforcement learning modules. The probability, for each of the two individuals involved in a mating event, of reproducing offspring was linearly depended on the individual\'s energy level (). A reproductive mating event created two offspring in the individual\'s subpopulation by applying one-point crossover with a probability of . The genes of the two newly created genotypes were then mutated with probability of , by adding value from a Gaussian distribution with zero mean and a standard deviation of 0.1. After all individuals in a subpopulation had survived for a full lifetime or died a premature death, a new subpopulation was created by randomly selecting a fixed number, i.e., 20 in our experiments, of the offspring reproduced during the last generation.
The genotype consisted of 51 real-valued genes: 1) 5 genes controlling the mating strategy by encoding the weights of the top-level neural network that selected basic behaviors ([Figure 2](#pone-0093622-g002){ref-type="fig"}); 2) 42 genes determining the parameters of the additional reward signals for the basic behaviors in the form of potential-based shaping rewards [@pone.0093622-Ng1]; and 3) 4 genes determining the meta-parameters of the reinforcement learning algorithm. The five-dimensional input to the neural network consisted of a constant bias of 1 (), the individual\'s internal energy (), and the inverse distances to the nearest energy source (), tail-lamp (), and face (). The sensory inputs were linearly scaled to a range of . If a visual target was not visible, the corresponding input value was set to −1. In each sensory-motor cycle (time step), the output of the neural network () determined which of two reinforcement learning modules that was selected. If the output was greater than zero the mating module was selected, otherwise the foraging module was selected. After a successful mating event, whether it reproduced offspring or not, an individual could not select the mating module again until it had captured an energy source or until time steps had passed. During this time, the tail-lamp was turned off.
In the case when only an energy source and a tail-lamp were visible, the energy thresholds for the selection of the mating module, , was given bywhich depended on the distance to the closest energy source () and the distance to the closest tail-lamp (). In order to derive the average energy threshold , we computed the mean of over 676 values of and (26 equidistant values between 0 and 1 for each of the two sensory inputs).
The reinforcement learning modules learned their behaviors from scratch in each generation with the aid of evolutionarily tuned potential-based shaping rewards and meta-parameters. The foraging module executed a foraging behavior using the relative angle and the distance to the nearest energy source as state variables. The mating module executed either a mating behavior or a waiting behavior, depending on the current sensory inputs. If a face of another robot was visible, the mating behavior was executed using the relative angle and the distance to the nearest face as state variables. Otherwise, the waiting behavior was executed using the relative angle and the distance to the nearest tail-lamp as state variables. The behaviors were learned by the Sarsa reinforcement learning algorithm [@pone.0093622-Rummery1], [@pone.0093622-Sutton2] with tile coding [@pone.0093622-Sutton2] and potential-based shaping rewards [@pone.0093622-Ng1]. The global reward for the reinforcement learning modules was set to +1 for a successful mating event and +1 for a capture of an energy source, otherwise the reward was set to 0.
The additional experiments, conducted to investigate the evolutionarily stability of the emerged polymorphic ESSs, were performed in a similar manner as the evolution experiments. The only difference was that, in each generation, the subpopulations were created by randomly selecting the genotypes of the different phenotypes from the final generation of the evolution experiment according to the predefined phenotypes ratios.
For a detailed description of our embodied evolution framework and algorithm specifics, see [@pone.0093622-Elfwing1].
[^1]: **Competing Interests:**The authors have declared that no competing interests exist.
[^2]: Conceived and designed the experiments: SE KD. Performed the experiments: SE. Analyzed the data: SE. Wrote the paper: SE KD.
|
{
"pile_set_name": "PubMed Central"
}
|
Q:
Multiplier algebra of $A \otimes \mathcal{K}$
If $A$ is unital C$^*$-algebra, is it true that the multiplier algebra of $A \otimes \mathcal{K} $ is $ A \otimes \mathcal{B}(\mathcal{H})$? Where $\mathcal{K}$ is C$^*$-algebra of compact operators on the Hilbert space $\mathcal{H}$.
A:
If $A=C_0(X)$ and $B$ is a $C^\ast$-algebra then $M(A\otimes B)$ is the set of strictly continuous functions $\beta X\to M(B)$, where $\beta$ stands for Stone-Čech compactification.
If you take $X$ to be compact and $B=\mathcal{K}$ then we are in your setting.
But $C(X)\otimes\mathcal{B(H)}$ is the set of norm-continuos functions $\beta X=X\to\mathcal{B(H)}$. The strict topology is the $\sigma$-strong-$^\ast$ topology on $\mathcal{B(H)}$, which is different form the norm topology. This should answer your question in the negative.
A:
The fact stated in the answer by vap is proven in the paper "Multipliers of C*-algebras" by Akemann, Pedersen and Tomiyama (see Theorem 3.3, I guess). Moreover, they prove in Theorem 3.8 that multiplier algebras are not very well behaved with respect to minimal tensor products:
Let $A$ and $B$ be $C^*$-algebras and assume that $B$ has a countable
approximate unit, but no unit (think of $\mathcal{K}$ here in your
case) and that $A$ is infinite dimensional. Then
$$
M(A) \otimes M(B) \subsetneq M(A \otimes B)
$$
where the tensor product is the minimal one.
So, in particular, for any infinite dimensional unital $C^*$-algebra $A$, the tensor product $A \otimes \mathcal{B}(\mathcal{H})$ is always a proper subalgebra of $M(A \otimes \mathcal{K})$.
|
{
"pile_set_name": "StackExchange"
}
|
17.2.12
This was based on a field study I did last autumn. I was totally captivated by the beauty of the open interior of this forest. I named it Autumn Cathedral because the majesty of the tall trees being crowned with such beauty reminds me of a cathedral, and of the saying from Psalm 19, "Day to day pours forth speech, and night to night declares knowledge. There is no speech, nor are there words; their voice is not heard; yet their voice goes out through all the earth, and their words to the end of the world."
|
{
"pile_set_name": "Pile-CC"
}
|
13, l: 5}.
26/51
Three letters picked without replacement from {h: 5, l: 5, y: 1, s: 7}. What is prob of sequence yls?
35/4896
What is prob of sequence lzml when four letters picked without replacement from {a: 6, m: 4, n: 2, l: 1, t: 1, z: 4}?
0
Three letters picked without replacement from {h: 2, a: 2, v: 2, g: 2, l: 3, u: 1}. What is prob of sequence ggl?
1/220
Three letters picked without replacement from avavlvllv. What is prob of sequence val?
1/21
Four letters picked without replacement from gioqgisa. Give prob of sequence ggio.
1/420
Two letters picked without replacement from {d: 4, h: 6}. What is prob of sequence hd?
4/15
What is prob of sequence icov when four letters picked without replacement from {f: 9, c: 3, i: 1, v: 3, o: 2}?
1/4080
Two letters picked without replacement from caagymum. What is prob of sequence yu?
1/56
Two letters picked without replacement from kkkkpkkkkkkkkjkkkkkk. Give prob of sequence pj.
1/380
Four letters picked without replacement from joekerkvjrre. What is prob of sequence kvkr?
1/1980
Two letters picked without replacement from mpmpmpmppkkp. What is prob of sequence km?
2/33
Four letters picked without replacement from wwzcozozzczwzuww. What is prob of sequence ozwo?
1/728
Calculate prob of sequence iihi when four letters picked without replacement from {x: 2, b: 2, h: 2, d: 4, i: 7}.
1/136
Calculate prob of sequence jiz when three letters picked without replacement from mjikzkjjmjk.
2/495
Two letters picked without replacement from {q: 1, g: 1, z: 3, o: 1, k: 2, c: 4}. What is prob of sequence kg?
1/66
Two letters picked without replacement from {w: 3, l: 2, s: 3, c: 10, a: 1}. Give prob of sequence la.
1/171
What is prob of sequence pfs when three letters picked without replacement from {a: 3, f: 1, p: 1, v: 5, s: 4}?
1/546
Two letters picked without replacement from {y: 1, t: 1, x: 5, h: 7, d: 1}. What is prob of sequence tt?
0
Two letters picked without replacement from {j: 1, l: 1, x: 1, u: 1, y: 1, o: 2}. What is prob of sequence ux?
1/42
Four letters picked without replacement from vvvvvv. Give prob of sequence vvvv.
1
What is prob of sequence ev when two letters picked without replacement from {e: 5, v: 11}?
11/48
Two letters picked without replacement from rrroreorooerrr. Give prob of sequence or.
16/91
Four letters picked without replacement from {v: 1, r: 1, w: 6, c: 2, h: 3}. What is prob of sequence hhrh?
1/2860
Two letters picked without replacement from {s: 8, u: 1, j: 5, n: 4}. What is prob of sequence js?
20/153
Four letters picked without replacement from qqqqqgqqqqeegqeqq. What is prob of sequence qqgg?
11/2380
Two letters picked without replacement from {j: 1, l: 1, x: 1, o: 1, u: 2, q: 1}. What is prob of sequence qu?
1/21
What is prob of sequence qmms when four letters picked without replacement from qxhzhsqqmzmh?
1/1980
What is prob of sequence qwwq when four letters picked without replacement from {q: 4, p: 1, w: 2}?
1/35
Four letters picked without replacement from eeeeeeeeeeeeeiieee. What is prob of sequence eeii?
1/153
Two letters picked without replacement from {f: 1, x: 1, o: 2, m: 2, b: 2}. Give prob of sequence om.
1/14
Two letters picked without replacement from kyymsmyykymmmmmhhy. What is prob of sequence hy?
2/51
Four letters picked without replacement from {t: 5, b: 3, o: 2, i: 3}. Give prob of sequence iibo.
3/1430
Two letters picked without replacement from hhhthhhhhhhhhthhhhhh. Give prob of sequence th.
9/95
Calculate prob of sequence vo when two letters picked without replacement from ovoviiigvi.
1/15
Three letters picked without replacement from kmkkkkkk. Give prob of sequence kmk.
1/8
Calculate prob of sequence ar when two letters picked without replacement from aaaaraaraaaaaaaaaaa.
17/171
Two letters picked without replacement from bggggbggbgg. Give prob of sequence gb.
12/55
Three letters picked without replacement from mmmmm. Give prob of sequence mmm.
1
Two letters picked without replacement from {n: 2, l: 2, m: 6, d: 1}. What is prob of sequence dm?
3/55
Calculate prob of sequence ss when two letters picked without replacement from sssszszs.
15/28
Calculate prob of sequence lly when three letters picked without replacement from {l: 3, y: 9}.
9/220
Two letters picked without replacement from lml. Give prob of sequence lm.
1/3
What is prob of sequence sq when two letters picked without replacement from {s: 1, g: 1, q: 2, r: 1, h: 2}?
1/21
Three letters picked without replacement from poopo. Give prob of sequence ooo.
1/10
Calculate prob of sequence ww when two letters picked without replacement from wwgwcykwgywg.
5/33
What is prob of sequence cwge when four letters picked without replacement from {g: 2, w: 8, e: 5, c: 1, q: 3}?
5/5814
Three letters picked without replacement from pntnptpnnptpn. Give prob of sequence nnp.
25/429
Calculate prob of sequence emaa when four letters picked without replacement from emetatattatemtama.
3/952
What is prob of sequence tt when two letters picked without replacement from tggggtttgt?
2/9
Four letters picked without replacement from {w: 8, f: 5}. Give prob of sequence ffwf.
4/143
Calculate prob of sequence jb when two letters picked without replacement from {j: 11, c: 3, b: 4}.
22/153
Three letters picked without replacement from {e: 2, f: 4, a: 2, z: 2, r: 2, c: 2}. Give prob of sequence car.
1/273
Two letters picked without replacement from jbxj. What is prob of sequence xb?
1/12
What is prob of sequence dd when two letters picked without replacement from ddddggggddgdgggddg?
4/17
Three letters picked without replacement from oeoooeeooooooo. Give prob of sequence ooo.
165/364
Four letters picked without replacement from nzznnzn. Give prob of sequence znzn.
3/35
Calculate prob of sequence dx when two letters picked without replacement from qqqqxqqxqkkqhqdqqz.
1/153
What is prob of sequence qy when two letters picked without replacement from yfqyyqfqqooqyyqyqyyo?
14/95
Calculate prob of sequence iiv when three letters picked without replacement from {v: 5, k: 8, r: 1, i: 3, y: 2, l: 1}.
1/228
Calculate prob of sequence xx when two letters picked without replacement from qquwxwwwxxqxuxquaq.
10/153
Calculate prob of sequence wx when two letters picked without replacement from wxwvgyyyvy.
1/45
Three letters picked without replacement from {s: 1, y: 1, x: 4, r: 1, w: 4, t: 1}. What is prob of sequence rtx?
1/330
What is prob of sequence ll when two letters picked without replacement from {y: 1, x: 1, l: 2}?
1/6
What is prob of sequence tt when two letters picked without replacement from affftfafatat?
1/22
Calculate prob of sequence xi when two letters picked without replacement from hksixmx.
1/21
What is prob of sequence agw when three letters picked without replacement from obaabbbgw?
1/252
What is prob of sequence gg when two letters picked without replacement from {q: 1, g: 6, t: 3, x: 4}?
15/91
What is prob of sequence kn when two letters picked without replacement from {x: 4, k: 1, a: 5, n: 1}?
1/110
What is prob of sequence gg when two letters picked without replacement from {c: 1, b: 7, g: 2}?
1/45
Three letters picked without replacement from {x: 1, p: 1, q: 1, i: 1, s: 1}. What is prob of sequence sqq?
0
Three letters picked without replacement from {n: 3, v: 2, w: 2, t: 2}. What is prob of sequence tnt?
1/84
Calculate prob of sequence xnx when three letters picked without replacement from {x: 5, u: 7, a: 1, o: 1, n: 1, k: 2}.
1/204
Four letters picked without replacement from veddddeddddv. What is prob of sequence eved?
4/1485
Three letters picked without replacement from {g: 7, y: 6}. Give prob of sequence ygy.
35/286
What is prob of sequence fr when two letters picked without replacement from {f: 1, r: 5, x: 2, u: 8, h: 2}?
5/306
What is prob of sequence btqj when four letters picked without replacement from jtootttbjjqqqj?
2/1001
What is prob of sequence nv when two letters picked without replacement from {z: 2, n: 1, g: 1, v: 1}?
1/20
What is prob of sequence km when two letters picked without replacement from mmmmmmmmmmmmmwmkmm?
8/153
What is prob of sequence wi when two letters picked without replacement from ffsmsmisfw?
1/90
Two letters picked without replacement from ohxxoooooo. What is prob of sequence o
|
{
"pile_set_name": "DM Mathematics"
}
|
{
"dev": false,
"isRemoteAppiumServer": false,
"remoteAppiumServerUri": "http://<remoteurl>:<port>/wd/hub",
}
|
{
"pile_set_name": "Github"
}
|
Background {#Sec1}
==========
Uveitis is one of the major causes of blindness worldwide \[[@CR1]\]. Certain uveitis patients may complicate with cataract \[[@CR1], [@CR2]\]. Cataract in patients with uveitis results from both the process of primary disease and the persistent treatment with corticosteroids \[[@CR2]--[@CR4]\]. Among all the surgical approaches, phacoemulsification is the most effective method, because of a smaller incision of the cornea and a less stimulation of the vitreous \[[@CR5]\]. The ultrasound energy applied in phacoemulsification generate the heat, which may cause postoperative anterior chamber inflammation, corneal edema, incision burns, corneal endothelial cells loss, and even decompensation of corneal endothelium \[[@CR3], [@CR4]\]. Besides, the postoperative inflammatory reaction of cataracts caused by uveitis is usually more severe than that in senile cataracts because of the adhesive iris or intraocular environment prone to inflammation.
Hypothermia has been proved to be a protective factor to improve the body's tolerance to ischemia and hypoxia in numbers of studies \[[@CR6]--[@CR9]\]. Since hypothermia could decrease the tissue metabolism to reduce hypoxic-ischemic injury, therapeutic hypothermia is applied in many diseases. In the ocular field, it has been reported in some experiments that local hypothermia can protect the retina from acute ischemic injury and reduce postoperative inflammation in vitrectomy \[[@CR10], [@CR11]\]. In cataract surgeries, applying an ice-cold eye mask after operation could increase the comfort level and reduce inflammation with no adverse effects \[[@CR12]\]. In our previous study, we found that hypothermic perfusion in the phacoemulsification of senile hard nuclear cataract was safe and could effectively protect the corneal endothelium, decrease corneal edema and reduce anterior chamber inflammation in the early postoperative stage \[[@CR13]\]. Therefore, we hypothesized that hypothermia could also protect the cornea during phacoemulsification and reduce the early postoperative anterior chamber inflammation in patients with uveitis.
In this prospective randomised controlled trial, we aim to decrease the temperature of anterior chamber during phacoemulsification to reduce the postoperative reaction in uveitic cataract. Phacoemulsification was performed under an intraocular perfusion temperature at 4 °C (hypothermia) or 24 °C (room temperature) to evaluate the effectiveness and safety of hypothermic perfusion in the phacoemulsification of uveitic cataract surgeries.
Methods {#Sec2}
=======
Study design {#Sec3}
------------
A prospective, randomised, controlled clinical trial was conducted at the department of ophthalmology in the First Affiliated Hospital of Chongqing Medical University, Chongqing, China, a tertiary care center. The trial adhered to the principles of the Declaration of Helsinki. The study was approved by the ethical review committee of the First Affiliated Hospital of Chongqing Medical University, Chongqing, China, and the trial is registered in the Chinese Clinical Trial Registry (registration number ChiCTR1800016145). All the participants provided written informed consent before enrolment. The protocol and the possible risks and benefits of the trial were explained to all the participants.
Patient selection {#Sec4}
-----------------
The trial started in December 2017 and concluded in May 2019. One hundred and six patients with uveitic cataract were enrolled in this trial. The inclusion criteria were: age of more than 6 years, diagnosis of uveitis-associated cataract, and eyes had no active inflammation for 6 months prior to surgery. The exclusion criteria were: pregnant women, patients enrolled in other clinical studies at the same time, other ocular surgeries at the same time or before, previous usage of contact lenses, patients with other severe ocular diseases, systemic diseases including diabetes or hypertension, and patients who declined to consent.
Study protocol {#Sec5}
--------------
The patients were randomly divided into 2 groups (the 4 °C group and the 24 °C group). We chose 4 °C as the treatment group and 24 °C as the control group because 4 °C was available in the refrigerator, and the room temperature in the operating room was 24 °C. The perfusion fluid was stored in the 4 °C refrigerator before the surgery for at least 4 h in the 4 °C group. The perfusion fluid was not pre-treated in the 24 °C group. Both groups underwent phacoemulsification and intraocular lens implantation by the same surgeon. All the eyes were topically anesthetized with benoxinate. Phacoemulsification through a standard clear corneal incision with 2.8 mm in width and 1.5--1.75 mm in length and implantation of an intraocular lens were performed by the same surgeon with the same machine (Stellaris, Bausch & Lomb, USA). The central corneal thickness and the corneal thickness at the incision was measured by anterior segment optical coherence tomography (OCT) (Cirrus HD-OCT 1000, Carl Zeiss Meditec, Inc. Dublin, CA), the macular fovea thickness was observed by fundus OCT (Spectralis OCT, Heidelberg Engineering Corporation, Heidelberg, Germany). Corneal endothelial cell count (Topcon SP-3000P, Topcon Corporation, Tokyo, Japan), slit-lamp examination for the anterior and posterior segment observation, best corrected visual acuity (BCVA) and intraocular pressure (IOP) were examined before operation and on the 1st day and 7th day after operation. BCVA was recorded in Log MAR scores with the standard logarithmic visual acuity chart. Significant clinical visual improvement was defined as an improvement of more than 2 lines on the visual acuity chart \[[@CR14]\]. Anterior chamber flare and cells were graded with the Standardization of Uveitis Nomenclature (SUN) Grading System \[[@CR15]\] on the 1st day and 7th day postoperatively by a same examiner. Patients and examiners were masked to the grouping. Phacoemulsification time, ultrasound energy and perfusion time were recorded during the surgery. A topical combination of levofloxacin, diclofenac sodium and tobramycin-dexamethasone four times a day and tobramycin-dexamethasone ointment every night was applied after surgery. Systemic medications included prednisone or cyclosporine, and the dose varies according to the patients' condition.
Randomization and masking {#Sec6}
-------------------------
As a pilot study, we planned to enroll 100 patients instead of calculate sample size. Random numbers were generated by computer. All the patients were randomised to 2 groups before surgery. In order to restrict bias, the study was single-masked. Patients examiners, and evaluating investigators were masked to the grouping. The temperature of perfusate could be sensed by the surgeon during the phacoemulsification, the surgeon was therefore not masked. The surgeon played no other role in the study. To ensure that neither participants nor examiners would know the grouping, the random numbers were kept in envelopes to avoid selection bias. Unmasked patients were withdrawn from this study.
Statistical analysis {#Sec7}
--------------------
Data analysis was performed using SPSS version 21.0 (SPSS Inc., Chicago, IL, USA). Normality distribution of continuous variables was tested using the Kolmogorov-Smirnov test. The t-test and Mann-Whitney U test were used to compare variables. Pearson Chi-Square test with Fisher's exact test was applied for analyzing categorical variables. *P*-values lower than 0.05 were considered statistically significant.
Results {#Sec8}
=======
A total of 106 eyes from 106 patients were included in the study. Two patients unmasked were withdrawn from this study. The remaining 106 eyes were randomly divided into 4 °C group and 24 °C group, and each group included 53 eyes. There were no statistical differences in gender, age, systemic anti-inflammatory medicine use, average phacoemulsification time, average ultrasound energy or perfusion time between the two groups (*P* \> 0.05) (Table [1](#Tab1){ref-type="table"}). Table 1Demographic data of study subjects4 °C group24 °C groupT value*P* value**Patients**5353N/AN/A**GenderMale**21 (39.6%)28 (52.8%)N/A0.173**Female**32 (60.4%)25 (47.2%)**Age** (year, mean ± SD)40.2 ± 17.642.3 ± 16.9−0.6320.851**Patients using systemic corticosteroids**45 (84.9%)43 (81.1%)N/A0.605**Dose of prednisone** (mg/day)18.5 ± 3.318.4 ± 2.80.0460.963**Patients using cyclosporine**36 (67.9%)30 (56.6%)N/A0.229**Dose of cyclosporine** (mg/day)106.3 ± 16.2114.2 ± 22.4−1.6130.113**Average phacoemulsification time** (seconds)20.6 ± 18.019.4 ± 21.70.2390.390**Average ultrasound energy** (%)5.4 ± 2.34.5 ± 2.31.5540.506**Perfusion time** (seconds)152.6 ± 62.2143.3 ± 92.70.3520.548
Anterior chamber inflammation {#Sec9}
-----------------------------
The anterior chamber inflammation was described by the aqueous flare and the aqueous cells according to the Standardization of Uveitis Nomenclature (SUN) Grading System. The eyes had no aqueous flare or aqueous cells in the anterior chamber before the surgery. The aqueous flare score was 0.83 ± 0.76 in the 4 °C group, which was significantly lower than that in the 24 °C group 1.51 ± 1.02 (P = 0.006) on the first day after operation (Fig. [1](#Fig1){ref-type="fig"}). The aqueous cells score was 0.17 ± 0.38 in the 4 °C group, which was also significantly lower than that in the 24 °C group 0.62 ± 0.94 (*p* = 0.025) (Fig. [2](#Fig2){ref-type="fig"}). There was no significant difference in the aqueous flare score or aqueous cells score between the two groups on the 7th day (*p* \> 0.05). Fig. 1Aqueous flare score measured on the first day after operation in 4 °C group and 24 °C groupFig. 2Aqueous cells score measured on the first day after operation in 4 °C group and 24 °C group
Corneal analysis {#Sec10}
----------------
The corneal thickness was measured by anterior segment OCT. The mean corneal thickness of incision was 907.66 ± 85.37 μm in the 4 °C group, which was thinner than that in the 24 °C group (963.75 ± 103.81 μm) (*P* = 0.005) on the 1st day post operation. The mean central corneal thickness was 596.53 ± 54.30 μm in the 4 °C group and 603.88 ± 71.12 μm in the 24 °C group, without significant difference between two groups (*P* = 0.654). The corneal endothelial cells were also observed on the 1st day after operation. The mean endothelial cell density was 2845 ± 538.9 cells/mm2 in the 4 °C group and 2473 ± 475.8 cells/mm2 in the 24 °C group (*P* = 0.907). The mean percentage of hexagonal cells was 49.9 ± 12.7% in the 4 °C group and 48.6 ± 12.6% in the 24 °C group (*P* = 0.609). There was no significant difference on the parameters concerning on cornea between the two groups on the 7th day post operation (*p* \> 0.05).
Macular fovea thickness {#Sec11}
-----------------------
Because of the cataract and the miotic pupils, some of the macular fovea thickness cannot be observed by OCT before the surgery. The mean macular fovea thickness before surgery was 232.0 ± 26.9 μm in the 4 °C group and 231.6 ± 38.5 μm in the 24 °C group, without significant differences between two groups (*p* = 0.971). On the first day after operation, the mean macular fovea thickness was 232.1 ± 24.3 μm in the 4 °C group and 223.8 ± 34.3 μm in the 24 °C group. There was no significant difference between the two groups (*p* = 0.303). There was no significant difference before and after surgery in the 4 °C group (*p* = 0.989) and the 24 °C group (*p* = 0.480).
Clinical parameters {#Sec12}
-------------------
Significant clinical visual improvement was defined as an improvement of more than 2 lines on the visual acuity chart. There was no difference between the two groups in the visual improvement on the first day post operation (*p* = 0.183).
Transiently increased IOP was found in 17 eyes in the 4 °C group and 16 eyes in the 24 °C group postoperatively without significant difference (*p* = 0.834). No macular edema, endophthalmitis or other intraoperative or post-operative complications was observed in the two groups during the study.
Uveitis subtypes {#Sec13}
----------------
The subtypes of uveitis included Vogt-Koyanagi-Harada (VKH) syndrome (27 patients), anterior uveitis (19 patients), Behcet's disease (14 patients), pediatric uveitis (11 patients), Fuchs' syndrome (11 patients), and other subtypes of uveitis (23 patients) which consisted of retinal vasculitis (3 patients), scleritis (1 patient), lymphoma (1 patient), syphilis (1 patient), glaucomatocyclitic crisis (1 patient), panuveitis (1 patient), Blau syndrome (1 patient), and unclassified uveitis (14 patients) (Table [2](#Tab2){ref-type="table"}). Table 2Subtypes of uveitis4 °C group24 °C groupTotal**Vogt-Koyanagi-Harada disease (VKH)**11 (20.8%)16 (30.2%)27 (25.5%)**Anterior Uveitis**13 (24.5%)6 (11.3)19 (17.9%)**Behcet's disease (BD)**7 (13.2%)7 (13.2%)14 (13.2%)**Childhood uveitis**8 (15.1%)4 (7.5%)12 (11.3%)**Fuchs' syndrome**5 (9.4%)6 (11.3%)11 (10.4%)**Other subtypes**9 (17.0%)14 (26.4%)23 (21.6%)**Total**5353106
We found that the mean corneal thickness of incision was 0.874 ± 0.074 mm in the 4 °C group, which was significantly thinner than that in the 24 °C group (0.973 ± 0.116 mm) (*p* = 0.027) on the first day after operation in VKH syndrome patients. Other index including corneal endothelial cells analysis, macular fovea thickness and anterior chamber inflammation were not significantly different (p \> 0.05).
Discussion {#Sec14}
==========
Uveitis is a refractory and recurrent disease resulting in blindness. Although phacoemulsification can effectively remove the cataract caused by uveitis, the complexity of uveitis poses for the surgeon special challenges. The cataract caused by uveitis usually complicated with iris atrophy, posterior synechiae, a miotic pupil, pupillary membrane, invisible foveal reflection and bleeding from abnormal fragile iris vessels \[[@CR16]\], which leads to additional surgical procedures, develops more intraoperative complications, and has poorer postoperative visual acuity compared with the senile cataract surgery \[[@CR17]\]. The postoperative anterior chamber inflammatory reaction is usually more serious than senile cataract \[[@CR18]\]. We therefore aim to find a method to alleviate the inflammatory response after phacoemulsification in cataract eyes complicated with uveitis. In our study, hypothermic perfusion was used to decrease the anterior chamber temperature during phacoemulsification to reduce postoperative inflammatory response. The results showed that compared to 24 °C (room temperature), perfusion fluid at 4 °C during phacoemulsification was safe, and it resulted in milder anterior chamber inflammatory reaction and incisional corneal edema in the early postoperative stage.
Hypothermic perfusion could alleviate postoperative inflammatory responses. In our study, the hypothermic perfusion group had milder anterior chamber inflammatory reaction in the early stage after cataract surgery than the room temperature group. This might be due to the hypothermic perfusion in the anterior chamber that increased the tolerance of the anterior segment to ischemia and hypoxia, or decreased cellular metabolism. Hypothermia has been applied in many fields. It has been shown that transient hypothermia is an effective factor to induce tolerance to ischemic or hypoxic injury in nervous and cardiac systems \[[@CR19], [@CR20]\]. Local hypothermia has been also proved to protect the retina from acute ischemic injury caused by increased IOP and reduce postoperative inflammation in vitrectomy \[[@CR10], [@CR11]\]. Hypothermia could protect retinal ganglion cells against an ischemic insult and prolongs their ischemic tolerance time \[[@CR21]\]. It has been demonstrated that the neuronal energy metabolism has been reduced by hypothermic preconditioning \[[@CR22]\]. Meanwhile, hypothermia could decrease cellular metabolism, reduce vascular endothelial growth factor-A (VEGF-A) and sustain pigment epithelium--derived factor (PEDF) expression to create an anti-angiogenic environment in culture of human retinal pigment epithelium (RPE) cells \[[@CR23]\]. The global or ocular hypothermic preconditioning could also significantly protect retina from ischemia/reperfusion injury by reducing retinal glutamate uptake and glutamine synthetase activity \[[@CR24]\]. Cold exposure induces the expression of cold-shock proteins to inhibit the apoptotic process and make the visual neurons survive the retrograde insult \[[@CR25]\].
In addition, it has been observed that hypothermia could alleviate postoperative corneal edema at the incision, especially in eyes with VKH syndrome. This might be because hypothermic perfusion can reduce thermal damage caused by phacoemulsification. Ultrasonic energy in phacoemulsification is used to emulsify the lens nucleus. The vibration of the metal tip creates friction and generates heat. Excessive heat burns the cornea, causing corneal edema, melting and scarring of cornea \[[@CR26]\]. In our experiments, the temperature of the perfusate which took away some of the heat generated by phacoemulsification has been decreased to reduce incisional corneal edema at the early postoperative stage.
In our previous work, similar study of hypothermic perfusion in hard nuclear senile cataract by animal experiments and clinical trials has also been performed \[[@CR13]\]. There was a significant difference in the central and incisional corneal thickness, the corneal endothelial cells loss, and the anterior segmental inflammation between the 4 °C group and the 24 °C group in hard nuclear cataract surgeries. However, compared with reducing corneal edema, hypothermia tended to play a more vital role on alleviating anterior chamber inflammation in uveitic cataract. There were some reasons to explain the results. Firstly, the cataract caused by uveitis generally occurred in younger patients in comparison with senile cataracts \[[@CR27]\]. The most common subtype would be posterior sub-capsular cataracts instead of nuclear sclerosis cataracts \[[@CR28]\]. Thus, it required shorter phacoemulsification duration and less ultrasound energy, leading to milder corneal damage. Secondly, postoperative inflammatory reaction was usually more serous in uveitis complicated cataract as compared with senile cataract. Thirdly, the timing of the surgeries was strictly controlled to avoid severe intraocular inflammation occurred post cataract operation. It was generally recommended that a successful cataract surgery complicated with uveitis requires a quiet eye devoid of active inflammation for at least 3 months \[[@CR29]\]. In our study, all the patients were followed-up for at least 6 months before the surgery to ensure there was no active inflammation preoperatively. That was the key point to reduce postoperative inflammatory reaction. Fourthly, perioperative medication was also important. The patients were given adequate perioperative prophylactic anti-inflammatory therapy by a topical combination of diclofenac sodium and tobramycin-dexamethasone, as well as systemic prednisone and/or cyclosporine. It was recommended to use supplemental corticosteroids and immunosuppressive agents to control postoperative inflammation and reduce the recurrence of uveitis \[[@CR27]--[@CR29]\].
The results of our trial showed that hypothermic perfusion was effective in reducing the early inflammatory response after cataract surgery of eyes with VKH syndrome. Uveitis associated with VKH syndrome is an autoimmune disease which progresses to chronic recurrent granulomatous intraocular inflammation. Patients with VKH syndrome are more likely to have ocular complications, such as cataract, high intraocular pressure, and fundus lesions, which lead to poor visual prognosis \[[@CR30], [@CR31]\]. It was reported that the incidence of cataract in patients with VKH syndrome ranges from 5 to 45% \[[@CR32]\]. Chronic recurrent VKH syndrome had a worse cataract surgery prognosis. We therefore need to reduce the inflammatory response after cataract surgery to minimize recurrences of uveitis. Therefore, we recommend to use the hypothermic perfusion fluid to reduce anterior chamber inflammation and incisional corneal damage at the early postoperative stage for VKH syndrome cataract patients.
There were several limitations in our study. The trial was single-masked (participants and examiners were masked while the surgeon was not), which could not completely avoid the subjective bias from the researchers. Besides, the anterior chamber inflammation was measured using the SUN Grading System, which might increase the bias of subjective interpretation. Laser flare and cell meter would be more objective and will be conducted in the next step. In this study, the clinical parameters have been observed in the early postoperative stage. The follow-up period should be extended to observe long term postoperative reactions including macular changes in future studies. Further study with accurate record of perfusion amount and its effect on endothelial cell loss would be conducted. And further multi-center large sample study by different tertiary centers would be necessary to analyze the beneficial effects of intraoperative hypothermic perfusion on different subtypes of uveitis.
Conclusions {#Sec15}
===========
In conclusion, hypothermic intraocular perfusion during the phacoemulsification is safe, and it can effectively inhibit the anterior chamber inflammation and reduce corneal edema at the incision at the early postoperative stage in cataract patients with uveitis, which could be recommended in the uveitic cataract surgeries.
IOP
: Intraocular pressure
OCT
: Optical coherence tomography
BCVA
: Best corrected visual acuity
SUN
: Standardization of Uveitis Nomenclature
VKH syndrome
: Vogt-Koyanagi-Harada syndrome
VEGF-A
: Vascular endothelial growth factor-A
PEDF
: Pigment epithelium--derived factor
RPE
: Retinal pigment epithelium.
**Publisher's Note**
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The authors would like to thank Junyan Zhang from Bothwin Clinical Study Consultant, WA, USA for English editing and suggestions on this study.
KH, LJ and WW participated in the design of the study. LJ, YX, LX, LX, BW, RL, YX, ZL, LZ, YC and YJ helped with the acquisition of the data. LJ and WW carried out the statistical analysis and drafted the manuscript. KH and PY revised the final manuscript. All authors read and approved the manuscript.
This work was supported by the National Natural Science Foundation of China (grant number 81870650, 81100657, 81570832 and 81970832), the Project Foundation of Chongqing Science and Technology Commission of China (grant number cstc2018jcyjA0429), Chongqing Education Commission Project Fund of China (grant number CQGJ17062B), and the Project of Chongqing Health Commission combined with Science and Technology Commission of China (grant number 2018MSXM003 and 2018GDRC008\]. The funders had no role in the study design, data collection, analysis or writing of the manuscript.
The datasets analysed are available from the corresponding author on reasonable request.
The study adhered to the principles of the Declaration of Helsinki and was approved by the ethical review committee of the First Affiliated Hospital of Chongqing Medical University, Chongqing, China, with the ethics approval reference number 20181401. All the participants provided written informed consent before enrolment. The protocol and the possible risks and benefits of the trial were explained to all the participants.
Not Applicable.
The authors declare that they have no competing interests.
|
{
"pile_set_name": "PubMed Central"
}
|
List of Amstrad CPC games
This list contains game titles released for the Amstrad CPC home computer series. This number is always up to date by this script.
0–9
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
Z
See also
Lists of video games
List of Amstrad PCW games
References
CPC Game Reviews by Nicholas Campbell
Amstrad CPC at Adventureland
CPC-power Database of CPC software (in French)
The 464 Project - 464 Games on the Amstrad CPC
The 464 Project Part 2 - 128 MORE Games on the Amstrad CPC
Amstrad CPC
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Nenad Filipović (racewalker)
Nenad Filipović (Serbian Cyrillic: Ненад Филиповић; born 5 October 1978) is a Serbian racewalker. He competed in the 50 km event at the 2008, 2012 and 2016 Olympics with the best result of 30th place in 2008. His twin brother Predrag is also an Olympic racewalker.
References
Category:Sportspeople from Leskovac
Category:Serbian male racewalkers
Category:1978 births
Category:Living people
Category:Athletes (track and field) at the 2008 Summer Olympics
Category:Athletes (track and field) at the 2012 Summer Olympics
Category:Athletes (track and field) at the 2016 Summer Olympics
Category:Olympic athletes of Serbia
|
{
"pile_set_name": "Wikipedia (en)"
}
|
The informed consent aftermath of the genetic revolution. An Italian example of implementation.
A great part of human genetics research is carried out collecting data and building large databases of biological samples that are in a non-anonymous format. These constitute a valuable resource for future research. The construction of such databases and tissue banks facilitates important scientific progress. However, biobanks have been recognized as ethically problematic because they contain thousands of data that could expose individuals and populations to discrimination, stigmatization and psychological stress if misused. Informed consent is regarded as a cornerstone in the protection of personal autonomy in research involving human subjects. Yet in recent years this fundamental concept has been overwhelmed by the genomic revolution. From a general overview of international literature, it seems evident that informed consent issues have come into sharp focus, in particular in relation to the twin issues of time extension (blanket versus specific/repeated consent) and personal extension (group consent). After an introduction on obtaining informed consent in the context of genetic research, this paper addresses the apparent lack of a single, universal model of obtaining informed consent among populations involved in genetic research and it argues for the need to develop an ethical framework tailored to the specific features of each project. In order to support this theory of contextualizing, the case of a private biotechnology company, SharDNA is presented. The present paper explores the management of its biobank, developed from a genetic research project carried out on isolated populations living on the Italian island of Sardinia. In particular, the paper highlights how the company is tackling the problem of informed consent and other ethical requirements for genetic research, such as the respect of individual privacy, the population approach and the existing Italian legal regulatory framework.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
et a(s) = -s**3 + 4*s**2 + 3*s + 4. Does 11 divide a(3)?
True
Suppose 145 = 2*b - b. Suppose 0 = -5*a + 3*m + b, 9*a - 135 = 4*a + m. Is 13 a factor of a?
True
Suppose -3*p = 5*j - 241, 162 + 183 = 5*p - 3*j. Does 8 divide p?
True
Suppose 5*b = 3*b + 48. Suppose 27*g = b*g + 90. Is 10 a factor of g?
True
Suppose -2*i - 2*i = 8. Let f(d) = -8*d**2 + 2*d - 2. Let w(t) = 16*t**2 - 4*t + 5. Let y(m) = -5*f(m) - 2*w(m). Is y(i) a multiple of 19?
False
Suppose 2*g - 3 = 1. Let z be 1 - (-6*g - -2). Suppose -23 = -2*t + z. Is 8 a factor of t?
False
Is 70 + 1 - (-25 + 24) a multiple of 14?
False
Suppose 504 + 77 = 7*r. Is r a multiple of 26?
False
Let k(v) = 4*v**2 + 7*v. Is 15 a factor of k(-5)?
False
Let h be ((-3)/(-3))/((-3)/18). Let q(v) be the second derivative of -v**3/6 - v**2 + v. Does 3 divide q(h)?
False
Does 5 divide 50/(-3*2/(-3)*1)?
True
Let a be 30/(-4)*12/(-18). Suppose 4*z = a*z - 4. Is z a multiple of 2?
True
Suppose -52 - 16 = -4*a. Suppose a = -2*b + 49. Is b a multiple of 12?
False
Let k = 10 - 7. Let r(j) = -5*j**3 - 4*j**2 + 9 + 6*j**3 + k*j**2. Is r(0) a multiple of 6?
False
Suppose -b + 0 = -45. Does 9 divide b?
True
Is 1026/36*4/3 a multiple of 8?
False
Let k be 7 - ((1 - 1) + 1). Let u = k - 3. Suppose -40 = -u*m + 8. Is 8 a factor of m?
True
Let p(q) = 5*q**2 + 1. Let m = -5 + 3. Let g be p(m). Suppose 4*b + g = 93. Is b a multiple of 12?
False
Let l = 5 - 3. Let g(v) = -v + 0*v - v**l - 29*v**3 + 0 - 1. Is 14 a factor of g(-1)?
True
Suppose -2*z + 2004 = 4*a, 6*a - a - 5*z = 2535. Suppose 2*x - 5*j = 188, x + 4*j - a = -4*x. Let l = x + -45. Is 14 a factor of l?
False
Let x(r) be the third derivative of r**6/120 + r**5/10 - 3*r**4/8 + 3*r**3/2 + 6*r**2. Is x(-7) a multiple of 12?
False
Let c be (51/(-7) + 3)*7. Let j = -18 - c. Is j a multiple of 6?
True
Let t(j) = -5*j - 14. Is t(-7) a multiple of 7?
True
Let n(u) be the second derivative of -u**5/20 + u**4/12 + u**3/6 + u**2 - 2*u. Let v be n(0). Suppose -v*c = -0*c - 48. Is c a multiple of 12?
True
Let b(s) = -s**3 + 7*s**2 - 3*s + 7. Does 21 divide b(5)?
True
Let g = 61 - -91. Let n = g + -105. Does 13 divide n?
False
Let g(c) = 41*c**3 - 2*c**2 + 3*c - 1. Is g(1) a multiple of 41?
True
Let d(g) = 2*g - 11. Let t(n) = -3*n - 9. Let v be t(-8). Let u = v - 5. Is d(u) a multiple of 9?
True
Suppose -2*k + c + 33 = -6*k, 2*k + 2*c + 18 = 0. Is 4 a factor of (-54)/k + (-21)/28?
False
Let j(t) = 7*t - 18. Is j(12) a multiple of 22?
True
Let q = 41 - -24. Does 12 divide q?
False
Let x = -55 - -66. Is 3 a factor of x?
False
Suppose 2*v - 387 = -r + 6*r, 0 = 5*v - 4*r - 959. Suppose -4*d = -0*d - n - v, 0 = 2*d - 3*n - 83. Does 14 divide d?
False
Suppose -t + 6*t + 15 = 5*r, -2*t = 5*r - 29. Suppose 3*w = -4*y + 2*w - 309, -t*y = 3*w + 167. Let s = y + 109. Is 14 a factor of s?
False
Suppose -5*b = -25, -a - 1 - 3 = -b. Suppose 4*d - 3*f = -d - 3, 5*d + f = a. Suppose d = -m - 4*m + 50. Does 6 divide m?
False
Let f(m) = m**3 + 4*m**2 - 11*m + 7. Is f(5) a multiple of 8?
False
Suppose 7*x = 2*x + 270. Suppose n + 0*n = x. Does 14 divide n?
False
Let d(i) = i**3 - 2*i**2 - 3*i. Let m(v) = -v - 4. Let b be m(-9). Suppose -b*c = -14 - 6. Is d(c) a multiple of 15?
False
Let z = -128 + 188. Is 20 a factor of z?
True
Suppose -i + g - 5*g = -112, 3*i - 352 = 4*g. Does 15 divide i?
False
Let k(h) = 26*h**2 - 1. Let g = -1 - -2. Is 25 a factor of k(g)?
True
Let z = -5 + 14. Is 12 a factor of z + 2/(4/6)?
True
Suppose 6*f - 12 = 3*f. Let l = 6 - f. Suppose l*o + 122 = 5*a, -2*a = -4*a + 2*o + 50. Is a a multiple of 8?
True
Let j = -41 - -59. Is 6 a factor of j?
True
Let q = 7 - 2. Suppose 9 = q*a - 16. Is a a multiple of 4?
False
Suppose 0 = -7*r + 6*r + 13. Does 3 divide r?
False
Let r be 2 + 14/(-2 - -1). Let y = 17 + r. Suppose -19 = -b - b + y*c, -2*b + 2*c = -34. Is b a multiple of 11?
True
Suppose -4*v + 124 + 44 = 0. Suppose 4*b - 66 = v. Is 10 a factor of b?
False
Let s(q) = -q - 7. Let p be s(-9). Let h be 88/10 + p/10. Let j = -1 + h. Is 8 a factor of j?
True
Suppose 31 = 5*s - 2*b, s = 2*b + 2 + 1. Suppose -2*p + r + 149 = 0, r = -s*p + 2*p + 355. Does 23 divide p?
False
Let l(p) = -25*p**3 + 2*p**2 + 2*p + 2. Is l(-1) a multiple of 20?
False
Does 9 divide 189/14*64/12?
True
Suppose -5*k + 82 = -u - 34, 3*k + 5*u - 64 = 0. Is 23 a factor of k?
True
Let x(t) = t**2 + 7*t. Let b be x(-4). Let u = 29 + b. Is u a multiple of 12?
False
Let u be (-3)/6*(-1 + 1). Is -1 - -1 - (u - 40) a multiple of 16?
False
Let x(d) be the second derivative of -d**3/3 - 6*d. Let i(g) = -g**3 + 5*g**2 + 5*g + 1. Let q be i(6). Is 4 a factor of x(q)?
False
Let p = 45 + 43. Is 22 a factor of p?
True
Let u(i) be the second derivative of i**3/6 - 2*i**2 - 2*i. Let l be u(7). Suppose 0 = -l*t + t + 8. Does 2 divide t?
True
Suppose -m + 301 = 6*m. Is m a multiple of 16?
False
Suppose -3*a + 3*c + 2*c + 735 = 0, 0 = -5*a - 3*c + 1191. Is 16 a factor of a?
True
Let n be 1/(3/(-12)*-1). Suppose 0 = -n*y - 2*a + 24, -4*y + 0*a + 5*a = -52. Is y a multiple of 5?
False
Let b = -113 - -64. Let u = 34 + b. Is 16 a factor of (-6)/((u/18)/5)?
False
Let z(v) = -v + 2. Let i be z(-2). Suppose x - 236 = 3*r, -5*r + i*x - 567 = -183. Let y = r - -119. Is 15 a factor of y?
False
Let v be (-368)/(-88) + 4/(-22). Suppose 2*f = v*f - 108. Does 18 divide f?
True
Let r(q) = q**3 + 7*q**2 + 4*q - 6. Let b(y) = -y**3 - y**2 - 6. Suppose -2*o + o = 0. Let g be b(o). Is r(g) a multiple of 4?
False
Let i = -6 - -2. Let z = -2 - i. Is 16 a factor of (z + -1)*(-3 + 46)?
False
Suppose -3*l = -15, -2*l = 2*g - 0*l - 94. Let k = 64 - g. Is 12 a factor of k?
False
Let a = -15 + 21. Suppose 5*k + 6 = -b, 5*k + 3*b + b = a. Is 4 a factor of ((-4)/8)/(k/48)?
True
Let l = 10 - 9. Is (-459)/(-18) + l/(-2) a multiple of 10?
False
Suppose -3*v + 26 = 8. Suppose h = v*h - 40. Let o = h - 3. Is 2 a factor of o?
False
Let f(t) = t**2 - t. Let g be f(0). Let o be (2 - (2 + g))/(-1). Is 13 a factor of -2 + 32 + o + 0?
False
Let r = 5 - 4. Let x = 54 + -78. Does 13 divide (-6*x/9)/r?
False
Suppose 14 + 100 = 2*n. Is 17 a factor of n?
False
Let b(h) = 33*h - 12. Does 9 divide b(8)?
True
Let g(k) = 17*k**3 + 11*k**2 + 12*k - 8. Let b(p) = -4*p**3 - 3*p**2 - 3*p + 2. Let u(i) = 9*b(i) + 2*g(i). Is u(-3) a multiple of 16?
False
Let m(c) = 2*c - 2. Let j be m(4). Let v(g) be the third derivative of -g**6/120 + g**5/10 + g**4/12 - g**3/2 + g**2. Does 9 divide v(j)?
True
Let f(z) = 3*z**2 - 9*z - 18. Does 13 divide f(8)?
False
Let m(l) be the first derivative of l**3/3 - l**2/2 - 7*l - 1. Does 12 divide m(-5)?
False
Let l = -61 - -178. Is l a multiple of 31?
False
Suppose -5*f + 2*g + 7 = 0, -16 + 5 = -5*f + g. Suppose -124 = -f*z - z. Suppose -4*t + 81 + z = 0. Is 10 a factor of t?
False
Let r be 4/(-18) - 790/90. Let x = 39 - r. Is x a multiple of 8?
True
Suppose 0 = 4*x + 8, -3*x - 2 + 6 = 2*h. Suppose h = 2*g + 3*z, 3*z = -4*g - z + 16. Does 7 divide g?
True
Let q(p) = -2*p**3 - 6*p**2 - 7. Is 32 a factor of q(-5)?
False
Let z be (-4 + 0)/(2 + -4). Suppose z*a - 5*c - 172 = -7*c, 0 = 2*a + 4*c - 174. Suppose -w + a = 4*w. Does 8 divide w?
False
Suppose -2*f + 224 - 52 = 0. Does 21 divide f?
False
Is 27 a factor of (-5 + (-1 - -24))/(6/27)?
True
Let g = 231 - 148. Suppose 5*p + r = 376, -p + g - 7 = r. Is 15 a factor of p?
True
Suppose -3*u - 51 = -6*x + 3*x, 2*x + 5*u - 6 = 0. Does 6 divide x?
False
Let b be (18/15)/(-3)*-5. Suppose -2*k = b*k - 224. Does 15 divide k?
False
Suppose -3*z - 4*a + 86 = 0, -5*z + 3*a - a + 126 = 0. Is z a multiple of 13?
True
Let v = -79 - -104. Is v a multiple of 2?
False
Suppose -v + 4*v + 3*b + 588 = 0, 2*b = -4*v - 776. Suppose -s = j - 0*s + 14, -5*j = 4*s + 74. Does 11 divide v/j - (-1)/3?
True
Let a(x) = 2*x**2 + 13*x + 27. Is 36 a factor of a(-9)?
True
Suppose 2*h + 11 = -1. Is 13 a factor of ((-148)/h)/((-4)/(-6))?
False
Let k(m) = -40*m - 8. Is k(-2) a multiple of 36?
True
Suppose 115 = k + 4*k. Does 23 divide k?
True
Let u(a) = 4*a**2 - 6*a. Let g(y) = -y**2 - 8*y - 4. Let r be g(-5). Let h = -7 + r. Does 20 divide u(h)?
True
Let m(w) = 3*w - w - w - 4. Let n be m(6). Suppose n*b - 7*b + 90 = 0. Is b a multiple of 9?
True
Suppose -2*o = q - 784, 1558 = -2*o + 6*o - 3*q. Does 39 divide o?
False
Suppose 0*z +
|
{
"pile_set_name": "DM Mathematics"
}
|
Why You Should Buy: Whether youíre watching a movie on your tablet, speeding through a book on your e-reader, replying to emails on your work phone or sending hilarious group texts on your cell, you use a lot of juice. Keep all of your gadgets charged and in-charge with Three 3-in-1 Charging Cables from ePacific Mall.http://www.epacificmall.com/gadgets/3-in-1-lightning-30-pin-micr...
What You Will Get: Low batteries be gone with this package that includes three charging cables, each able to charge up to 3 devices at once (if youíre counting thatís 9 cords for charging!). The devices are: lightning, 30-Pin, Micro USB to USB charger.
What You Should Know: These super-slim cables are easily transportable for charging on-the-go, and with a three-pack, you can plant one at home, one at work and keep one on you at all times. Plus, since they are compatible with iPhone 5, iPhone 4/4S, 3G/3GS, iPads, iPods, Android, HTC, Samsung and many other devices, you can be charged and in-charge, always.
Redemption InstructionsPurchase this CBS Local Offers Voucher and receive a unique voucher code in your email. Please click here to enter your voucher code and place your order. Tax and shipping are not included and must be paid separately at checkout. Please allow up to 3 weeks for shipping. Promotional value expires 10/31/2013.
Disclaimer: By providing links to other sites, FatWallet.com does not guarantee, approve or endorse the information or products available at these sites, nor does a link indicate any association with or endorsement by the linked site to FatWallet.com.
Members of our community may attach files to a post in accordance with the User Agreement. FatWallet is not responsible for the content, accuracy, completeness or validity of any information contained in any attached file. Files have *not* been scanned for viruses. Be especially wary of Excel files which may contain malicious content.
FatWallet coupons help you save more when shopping online. Use our Coupons Search to browse coupons and offers from thousands of stores, gathered into one convenient location.
Forums
As part of our FatWallet Community, you can share deals with almost a million shoppers in our forums. Forum content is generated by consumers for consumers. Share deals, money-saving tips, and more. It's FREE, fun, and addicting.
Support
Our customer experience team is here around the clock - real people ready to assist.
|
{
"pile_set_name": "Pile-CC"
}
|
Bombyx adipokinetic hormone receptor activates extracellular signal-regulated kinase 1 and 2 via G protein-dependent PKA and PKC but β-arrestin-independent pathways.
Neuropeptides of the adipokinetic hormone (AKH) family are among the best studied hormone peptides. They play important roles in insect hemolymph sugar homeostasis, larval lipolysis, and storage-fat mobilization. Mechanistic investigations have shown that, upon AKH stimulation, adipokinetic hormone receptor (AKHR) couples to a Gs protein and enhances adenylate cyclase activity, leading to intracellular cAMP accumulation. However, the underlying molecular mechanism by which this signaling pathway connects to extracellular signal-regulated kinase 1/2 (ERK1/2) remains to be elucidated. Using HEK293 cells stably or transiently expressing AKHR, we demonstrated that activation of AKHR elicited transient phosphorylation of ERK1/2. Our investigation indicated that AKHR-mediated activation of ERK1/2 was significantly inhibited by H-89 (protein kinase A inhibitor), Go6983, and GF109203X (protein kinase C inhibitors) but not by U73122 (PLC inhibitor) or FIPI (PLD inhibitor). Moreover, AKHR-induced ERK1/2 phosphorylation was blocked by the calcium chelators EGTA and BAPTA-AM. Furthermore, ERK1/2 activation in both transiently and stably AKHR-expressing HEK293 cells was found to be sensitive to pretreatment of pertussis toxin, whereas AKHR-mediated ERK1/2 activation was insensitive to siRNA-induced knockdown of β-arrestins and to pretreatment of inhibitors of EGFR, Src, and PI3K. On the basis of our data, we propose that activated AKHR signals to ERK1/2 primarily via PKA- and calcium-involved PKC-dependent pathways. Our current study provides the first in-depth study defining the mechanisms of AKH-mediated ERK activation through the Bombyx AKHR.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
The New Zealand Herald | A taste of what’s to come from smartphones
December 28, 2013
Few of us can conceive of the pace with which technological power is developing.
Ray Kurzweil (author, futurist, and a director of engineering at Google) predicts that by 2025 we’ll have a computer which has the processing power of the human brain, and by 2045 it’ll have the processing power of six billion brains — i.e., everyone on the planet. [...]
|
{
"pile_set_name": "Pile-CC"
}
|
You're currently on:
8in. Fence Pot Holder
Product Description
Wood Fence Top Flower Pot Holder- 8" Ring with wood fenc bracket easily adjust to fit over 2x4 or 2x6 fence top without the need to make any holes. A great way to add color and charm to any plain fence.
|
{
"pile_set_name": "Pile-CC"
}
|
A boat can be more easily maneuvered by the use of a bow thruster which consists of a water pump and a valve arrangement for diverting the water into either of two pipelines that open at the opposite sides of the boat near the bow. It would be desirable if the thrust obtainable from a water pump placed on board a vessel could be used to help maneuver the boat, or if required, to propel the boat forwardly, as when the propeller has been broken or where the propeller could cause injury to persons or articles in the water.
|
{
"pile_set_name": "USPTO Backgrounds"
}
|
“[T]he tornado was so ferocious it pulsed with 2,700 degrees of heat,” the San Francisco Chronicle reported. “It ripped roofs off houses and flung cars, power line towers and a steel shipping container into the air.”
Not only was this “raging cyclone of flame” the worst tornado of any kind in California history, but Stoke is the latest of six firefighters who lost their lives this year attempting to contain the most intense wildfires in California history.
An entire neighborhood in the city of Santa Rosa was wiped out by the fires in 2017
One of the fires spanned a record-breaking 300,000 acres. That’s the size of Los Angeles. Another first: Since last year, the fires have increasingly swept through more populated areas — like the Napa and Sonoma County blazes last yearthat killed at least 41 people and destroyed parts of the city of Santa Rosa.
But there is another not-at-all-natural factor that is a constant from Greece to California and beyond: global warming and the extreme weather patterns that result from it. Each day’s news seems to bring new examples and evidence:
JUST AS there can be no doubt that a climate change disaster is already taking place, there can be no doubt about its cause: capitalism.
The source of environmental destruction, spreading pollution, immense waste and the failure to address the long-term effects of all of these is an economic and social system that prioritizes the profit and power of a few over the needs of the many.
But when you read mainstream news reports about the capitalism-made disaster in California, you’re much more likely to see blame heaped on individuals — those who might have set a fire, others who accidentally dropped a lit cigarette, or even a flat tire on a car that caused sparks in a dry area.
Even when news reports mention “climate change,” they don’t focus on any kind of thoroughgoing critique of the free-market system and its in-built dynamics that lead to environmental destruction.
If we are ever going to reverse course for the sake of the world’s people and the planet, we have to name the system of capitalism in order to understand what we’re struggling against. Only from here can we start to put forward real solutions to the ecological crisis we face today — and begin to envision what a different world we could live in if human and ecological needs were put ahead of profits.
THE ECOLOGICAL crisis we face is rooted in the way the capitalist system works.
Capitalism is driven by the endless pursuit of profit. It is a system that has no consideration for the long-term health and well-being of the vast majority of people on the planet, let alone the natural world — the land we live on, the water we drink, the air we breathe.
Expropriation and exploitation of both land and labor are vital to capitalism. They are turned into commodities to be bought and sold on the market. As Chris Williams writes in his book Ecology and Socialism, Karl Marx described how capitalism’s robbery of both the worker and the soil together was the “original source of all wealth.”
The history of the U.S. demonstrates this violent truth. Capitalism in the U.S. was born out of a colonial settler state that stole, expropriated and accumulated Indigenous communal lands, and that kidnapped Africans and forced them to provide labor. The U.S. also relied on one of the world’s most exploited working classes to create the wealth needed to become the number-one economic and military power in the world.
Capitalism is based on a minority owning and controlling both the “means of production” (the factories and offices and land) and what the working class produces with its labor. The profits are privatized while costs are socialized. Capitalists accumulate as much wealth as possible, while taking no responsibility for the negative consequences for society, including environmental devastation.
The whole world suffers the destruction and chaos of living under extreme conditions of climate change caused by a capitalist system that chooses short-term profit over long-term sustainability.
There is also an inbuilt destructiveness to the system because of endless competition among capitalists as they attempt to capture markets and profits. One consequence of this is that the militaries of different states compete by means of arms — to win geopolitical influence and gain access to natural resources, cheap labor and markets.
Under capitalism, new technology is only used if it is profitable to do so. Its purpose is not first of all to meet more people’s needs more efficiently, but to help businesses get ahead of their competition and seize the market. As a result, overproduction and endless waste are an inherent part of capitalism.
For instance, the goal of producing homes under capitalism is not to house more people, but to make the most money for the people who own — and if that means leaving some completely built homes empty to raise the price, so be it.
The disregard of the 1 Percent for the long-term consequences of their actions is stunning. One case in point is the U.S. energy industry over the past decade.
It has been indisputable for 10 years and far longer that the extraction and burning of fossil fuels is causing rising global temperatures, which leads to all kinds of ecological destruction. But instead of reversing course — and using the technology that already exists to produce sustainable energy — the U.S. expanded its production of oil over the past 10 years.
The main reason for this is the boom in hydraulic fracturing, or fracking — a process in which a toxic mix of water and chemicals is pumped into shale rock at high pressure to release natural gas and oil deposits.
The extraction boom — which took place even as scientists were detecting ever-more alarming evidence of ecological crisis — was about getting cheaper energy for U.S. industry and also exporting oil and gas on the world market. U.S. rulers have enjoyed record profits during this decade, in part because of the extraction boom, while they continued to push austerity measures onto ordinary people.
TODAY, A lot of attention is focused on what the Trump administration is doing to the environment. And for good reason: the administration is gutting the Environmental Protection Agency, threatening to withdraw from climate change treaties, attacking the emissions control programs in California, stripping the Endangered Species Act and more.
But we shouldn’t forget that the fracking boom that helped drive the U.S. economy forward for the past 10 years began under the presidency of Barack Obama — with his encouragement.
If the Democratic Party is the champion of the environment, as it claims, then why did the Democratic National Committee (DNC) overturn a ban on candidates taking contributions from fossil fuel companies? In reality, both the Republicans and Democrats represent a ruling class agenda that prioritizes the accumulation of more and more wealth from our labor, alongside the extraction of the earth’s precious and limited resources.
Leaders of the Democratic Party like Obama and California Gov. Jerry Brown are credited as being groundbreakers against climate change. But Obama’s embrace of an “all of the above” energy policy that encouraged the extraction boom proves otherwise.
Where he took some pro-environment measures, the pressure of a movement demanding climate justice has never been far behind. For example, Obama canceled the permit for the construction of the Keystone XL pipeline only when activists committed to protest and direct action refused to be silent.
At the end of Obama’s presidency, he blocked the Dakota Access Pipeline. But this followed a historic struggle led by the Lakota (the Standing Rock Sioux), with the support of other Indigenous nations here and around the world, along with tens of thousands of non-Native activists, who all came together at Standing Rock.
In California, Gov. Brown has at least continued to press for increased use of renewable energy to meet targets for lower CO2 emissions — but one motivation is to allow for continued oil and gas extraction and production. California is the fourth-largest producer of fossil fuels among the 50 states, after Texas, North Dakota and Alaska.
Brown’s true record isn’t lost on climate justice activists and organizations. This September, during the Global Climate Action Summit that Brown will host in San Francisco, climate justice organizations are calling global days of action to “demand our elected leaders commit to no new fossil fuels and a just and fair transition to 100 percent renewable energy.”
EVERY YEAR, an environmental NGO calculates “Earth Overshoot Day” — the day when humans have consumed more resources than the planet can regenerate in a year. This year, that was August 1 — just one month past the midpoint of the year.
1. Take the carbon out of your commute.
2. Strive for zero waste.
3. Eat less meat and more veggies.
4. Tread lightly when you travel.
Unfortunately, these individual solutions so popular with the mainstream media won’t solve the climate crisis.
Why not? Take another California example: The state has suffered an ongoing cycle of droughts, which has put freshwater aquifers in jeopardy. The solution put forward by political and business leaders is for individuals to consume less water — and many working class people have happily done so, with the state meeting community conservation targets.
But big agriculture and industry are responsible for consuming the vast majority of the world’s water: over 90 percent. Ordinary people have no control over that at all. Instead of blaming them for supposedly consuming too much, we need to question how food is produced as commodities and how these commodities are transported.
What we need is a rational agriculture system that takes into account conservation and sustainability. You can only truly do this under a social and economic system that is planned, where everyone has a democratic say in how resources are used.
It’s also impossible to have a sustainable society without ending the poverty and inequality rooted in this system. The focus on individual consumers as the source of environmental problems doesn’t account for the fact that millions of people are already living with less and struggling to survive.
It isn’t possible to build a utopian oasis in our neighborhoods when almost half of the world’s population — over 3 billion people — live on less than $2.50 a day.
Even in the U.S., one in six people live in poverty, according to the government’s vastly understated official statistics — and nearly half fall in the struggling-to-get by category of up to twice the official poverty line. If you are a woman, Native American, Black, Latino, an immigrant or LGBTQ, you are even more likely to live in poverty.
INSTEAD OF individual solutions, we need to work toward systemic changes to the whole way our economy is designed and how production works. Once we understand individual solutions aren’t enough, we can discuss the social forces and struggles necessary to bring about system change.
The problem is not emissions, the problem is capitalism. The climate issue is like every other issue. It’s very important to working people, poor people, people around the world. It threatens their food, their water, their lives.
Extending the political and economic influence of workers is crucial to solving the climate crisis. If workers extend their control and power over politics and economic decision-making, I very much doubt, if it goes to the full process and conclusion, that what’s left standing would be called capitalism.
Any solution to climate change requires a reasonable, democratic and planned economy that stops making products to sell for a profit, that ends wars, that radically curbs waste and pollution and that transforms all the other things keeping energy and transportation demands so high.
This would have to begin with people organizing throughout society, especially in their workplaces, to have a say in what we produce and how we produce it. We need a revolution of the vast majority in society that takes political and economic power away from the rich — and that transforms ourselves and our way of life.
Chris Williams describes our vision for a socialist future:
Every single facet of industrial life — energy production most urgently, but also transportation, housing, trade, agriculture, manufacture of commodities, and waste production and treatment — all require gigantic systemic change and complete structural reorganization. It will be nothing short of totally remodeling the world on a social, political, technological, cultural and infrastructure level...
We cannot make these changes as individuals. The reconstruction of agriculture along sustainable lines, along with the expansion of alternative energy-harnessing technologies is a social project. These are the kinds of changes that need to occur to actually make a difference on the required socio-ecological level.
With a socialist society and a planned economy, some changes would be a longer process — like reimagining the very way our cities are built, our connections to where we work, how we build shelters and produce our food.
But other changes could be made in a shorter time span — for instance, moving to full renewable energy sources like wind, solar, tidal and geothermal — if the majority of people control society to meet their needs, not for profit.
THIS IS a long-term vision of a different world, but that doesn’t mean our struggles should wait. We need to fight for reforms today that could slow the climate disaster and elevate the worst circumstances for people whose communities are deemed sacrifice zones by capitalists and corporations.
We also need a movement based on solidarity that aims to address the historic injustices suffered by Indigenous peoples — by returning lands that were stolen and cleaning up those that have been polluted. In the process, we can open up the possibility of learning from Indigenous cultures, including their conservation practices.
And we have to organize for justice for immigrants and refugees who are going to continue to be on the move around the world, due to the unfolding climate crisis among other reasons.
Socialists want to build a culture around a reconnection to the land — based on the knowledge that humans are not separate from nature.
We need global solutions and coordination to combat climate change and build a sustainable world. That means working now in all the struggles taking place in society, but with the vision of a socialist future ahead.
|
{
"pile_set_name": "Pile-CC"
}
|
Outcome of laparoscopic inguinal hernia repair in a South African private practice setting.
The aim of this study was to determine the recurrence and complication rates of laparoscopic inguinal hernia repair performed in a private practice in Cape Town. An unselected cohort of 507 patients who underwent laparoscopic totally extraperitoneal (TEP) inguinal hernia repair before September 2005 were included in this study, thus ensuring a minimum 5-year follow-up. Patient demographic data, clinical notes, operating notes and outpatient follow-up notes were studied. Patients were interviewed telephonically regarding hernia recurrence, chronic pain and technique preference if they had previously undergone an open repair. All data collected were recorded on an electronic spreadsheet. The primary outcome parameter was recurrence. The secondary outcome parameters were postoperative and long-term complications. Of the 507 patients, 267 were contactable telephonically. There were 384 hernia repairs with a mean follow-up of 8.8 years. There were 9 recurrences (2.3%). The overall complication rate was 7.9%. Two per cent of patients suffered from chronic groin pain with gradual improvement since surgery. Sixteen per cent of patients had had previous open repair of an inguinal hernia, either on the ipsilateral or the contralateral side, and all judged the open repair to have been more painful. The recurrence and complication rates for laparoscopic TEP inguinal hernia repair in this practice are low and comparable to the best reported series. There is a low incidence of persistent postoperative pain with the laparoscopic technique, and it is the technique preferred by patients who previously underwent an open repair.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
package com.telerik.examples.examples.calendar;
import android.content.Context;
import android.view.View;
import android.view.ViewGroup;
import android.widget.ArrayAdapter;
import android.widget.ImageView;
import android.widget.TextView;
import com.telerik.examples.R;
import java.util.ArrayList;
public class HotelListAdapter extends ArrayAdapter<Hotel> {
private ArrayList<Hotel> hotels;
public HotelListAdapter(Context context, int resource) {
super(context, resource);
this.hotels = new ArrayList<Hotel>();
Hotel hotel = new Hotel();
hotel.setName("Buckingham gate");
hotel.setPrice(265);
this.hotels.add(hotel);
hotel = new Hotel();
hotel.setName("Three Lords Hotel");
hotel.setPrice(85);
this.hotels.add(hotel);
hotel = new Hotel();
hotel.setName("King's Court");
hotel.setPrice(155);
this.hotels.add(hotel);
hotel = new Hotel();
hotel.setName("Oxford Palace");
hotel.setPrice(105);
this.hotels.add(hotel);
hotel = new Hotel();
hotel.setName("Queen's Hotel");
hotel.setPrice(233);
this.hotels.add(hotel);
hotel = new Hotel();
hotel.setName("London Bridge Hotel");
hotel.setPrice(302);
this.hotels.add(hotel);
hotel = new Hotel();
hotel.setName("Hyde Park Towers Hotel");
hotel.setPrice(201);
this.hotels.add(hotel);
}
@Override
public View getView(int position, View convertView, ViewGroup parent) {
View view = View.inflate(getContext(), R.layout.calendar_range_list_item, null);
Hotel currentHotel = this.getItem(position);
TextView hotelName = (TextView) view.findViewById(R.id.txtHotelName);
TextView priceTag = (TextView) view.findViewById(R.id.txtPrice);
hotelName.setText(currentHotel.getName());
priceTag.setText("$ " + currentHotel.getPrice());
ImageView hotelImage = (ImageView) view.findViewById(R.id.hotelImage);
hotelImage.setImageResource(getContext().getResources().getIdentifier("calendar_pic_" + (position + 1), "drawable", getContext().getPackageName()));
return view;
}
@Override
public int getCount() {
return this.hotels.size();
}
@Override
public int getViewTypeCount() {
return 1;
}
@Override
public Hotel getItem(int position) {
return this.hotels.get(position);
}
}
|
{
"pile_set_name": "Github"
}
|
A meta-analysis of the rates of Staphylococcus aureus and methicillin-resistant S aureus contamination on the surfaces of environmental objects that health care workers frequently touch.
Health care workers may potentially spread Staphylococcus aureus and methicillin-resistant S aureus (MRSA) to patients by contaminated high-touch items. We aimed to determine the pooled rates of S aureus and MRSA contamination and influencing factors. A literature search of the PubMed, ScienceDirect, Embase, Ovid, and Scopus databases was performed. Pooled contamination rates were determined using random effect models. Subgroup and meta-regression analyses were conducted to identify factors potentially influencing the rates of S aureus and MRSA contamination. Sensitivity and publication bias analyses were performed. Thirty-eight studies were included in the meta-analysis. The pooled contamination rates were 15.0% (95% confidence interval [CI], 9.8%-21.1%) for S aureus and 5.0% (95% CI, 2.7%-7.7%) for MRSA. The subgroup analyses indicated that the pooled rate of S aureus contamination was significantly higher for studies conducted in South America, in developing countries, and during 2010-2015. The pooled rate of MRSA contamination was significantly higher for studies conducted in Africa. The meta-regression analysis suggested that the pooled rate of S aureus contamination was lower for studies conducted in developed countries (odds ratio, 0.664; 95% CI, 0.509-0.867; P = .004). No bias was found in the publication of the rates of S aureus and MRSA contamination. S aureus and MRSA contamination statuses of high-touch items are worrisome and should be paid greater attention. Developing country status was a risk factor for S aureus contamination.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
FARMVILLE, Va. — A new ad featuring President Obama urging voters to go to the polls to help protect his legacy has begun airing black radio in Ohio and Florida.
The ad features audio from the speech Obama delivered to the Congressional Black Caucus Foundation's Phoenix Awards dinner, his last as president. In the speech, a fired up Obama ridiculed Trump's black outreach and asked black voters to turn out if they wanted to give him "a good sendoff."
"And after we have achieved historic turnout in 2008 and 2012 especially in the African American community I will consider it a personal insult, an insult to my legacy, if this community lets down its guard and fails to activate itself in this election. You want to give me a good sendoff? Go vote," Obama said.
"You may have heard Hillary’s opponent in this election say that there’s never been a worse time to be a black person. I mean, he missed that whole civics lesson about slavery and Jim Crow — but we've got a museum for him to visit," Obama said during the speech, referring to the opening of the National Museum of African-American History and Culture.
In Ohio yesterday, Hillary Clinton urged turn out during a visit yesterday to Akron, and has received the endorsement of NBA star LeBron James, one of the state's favorite sons.
Clinton has between a 4% to 6% lead nationally, but is consistently behind in polls in Ohio. Florida, meanwhile, has been a virtual tie for much of the race. Although a Quinnipiac University Poll showed Clinton up 5% in Florida, low turn out there could have dire consequences for her.
Jeff Johnson, a Priorities USA strategists said Trump has "insulted and demeaned" the black community in this campaign and through his life. Obama's call to action, he said, is crucial to making sure Trump never becomes president. "This November we have to build on President Obama’s legacy by turning out in force and electing Hillary Clinton."
|
{
"pile_set_name": "OpenWebText2"
}
|
KCL
KCL or KCl may refer to:
Science and technology
Potassium chloride (KCl), a metal halide salt
Keycode lookup, keycode log, or keycode list
Kirchhoff's current law, in physics
Kyoto Common Lisp, an implementation of Common Lisp
Other uses
King's College London, a public research university in London, UK and a constituent college of the University of London
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Test Handicap
The Test Handicap was an American Thoroughbred horse race first run on July 25, 1896 at Brighton Beach Race Course in Brighton Beach, Brooklyn, New York where it continued annually through 1909. A race for horses age three and older, it was run on dirt over a distance of one mile.
On June 11, 1908, the Republican controlled New York Legislature under Governor Charles Evans Hughes passed the Hart-Agnew anti-wagering bill. The owners of Brighton Beach Race Course and other racing facilities in New York State struggled to stay in business without wagering revenue. Racetrack owners had no choice but to drastically reduce the purse money being paid out which saw important stakes worth as little as twenty-five percent of what they were just two years earlier. Although the Test Handicap had been scheduled for July 23, 1908, all stakes races were canceled at Brighton Beach Race Course and put on hiatus.
There was no Test Handicap in 1908 and 1909 but the following year Empire City Race Track took over the race dates belonging to the Bright Beach Race Course and in 1910 hosted the Test Handicap. The race was won by Everett, a three-year-old colt owned by James Francis Johnson's Quincy Stable. However, further restrictive legislation was passed by the New York Legislature in 1910 which resulted in the deepening of the financial crisis for track operators and led to a complete shut down of racing across the state during 1911 and 1912. When a Court ruling saw racing return in 1913 it was too late for the Brighton Beach facility and it never reopened.
Records
Speed record: (at 1 mile)
1:38.00 - Voter (1900) (New World Record).
1:38.00 - Hermis (1904)
Most wins by a jockey:
2 - Danny Maher (1898, 1899)
Most wins by a trainer:
3 - Thomas Welsh (1897, 1901, 1903)
Most wins by an owner:
2 - Julius Fleischmann (1897, 1903)
2 - James R. Keene (1900, 1905)
Winners
References
Category:Discontinued horse races
Category:Open mile category horse races
Category:Horse races in New York (state)
Category:Brighton Beach Race Course
Category:Yonkers Raceway
Category:Recurring sporting events established in 1896
Category:Recurring sporting events disestablished in 1910
Category:1896 establishments in New York (state)
Category:1910 disestablishments in New York (state)
|
{
"pile_set_name": "Wikipedia (en)"
}
|
Wind turbine blades are a means of converting power from wind into mechanical power to propel a generator of a wind turbine. Wind turbine blade design is mainly conducted under consideration of aerodynamic and mechanical considerations, i.e. the airfoil shape or profile is designed as a compromise between aerodynamic optimization and mechanical characteristics, for example strength, for various wind conditions. An optimal blade of a wind turbine exhibits a low cut in wind speed, good and efficient aerodynamic performance for various wind speeds, which means for low to high wind speed, minimal loads on blade and wind turbine structure, i.e. under turbulent and high wind speed conditions, and low acoustic noise.
It is for instance important that the blade can resist high forces and stresses close to the hub. Blades are therefore thick and wide in the region of a root section close to the hub. At the root, the blade is usually narrow and tubular to fit onto the hub and to provide sufficient strength. The blade profile becomes thinner and thinner as to obtain acceptable aerodynamic properties. The lift force will increase as the speed increases towards a tip of the blade, caused by a larger diameter at the same rotational frequency toward the tip. Decreasing the chord width towards the tip will contribute to counteract this effect. The blade tapers from a point somewhere near the root towards the tip. Furthermore the blade is twisted along its axis to account for a change in direction of the airflow to the wind resulting from rotation. The speed of a blade section is increasing the further it is located towards a tip of the blade.
Nevertheless, conventional blades of wind turbines have a static airfoil and thus exhibit limited possibilities to adjust for wind conditions. The only possible means of adjustment and optimization of the aerodynamic properties for a wind speed and turbulence is the adjustment of a pitch angle for the whole blade. From an aerodynamic point of view an optimal blade of a wind turbine would not only comprise a variable pitch angle for each section of the blade but also comprise of an adjustable airfoil to account for different wind and turbulence conditions.
It is commonly known to enhance and optimize the performance of wind turbine blades with devices added onto the wind turbine blade. Such devices are among others active or passive components such as flaps, vortex generators or stall strips. The actuation of flaps can for instance be conducted with electricity, hydraulic or piezoelectric means. The bending of a blade is considered to stay in a relation to blade loading and wind conditions. A passive solution is known, where a shape and/or angle of a flap relative to a blades chord line changes depending on the bending of a wind turbine blade.
Another difficulty with wind turbine blades is the occurrence of acoustic noise when in operation. The tip speed of a blade in operation is for instance 80 m/s. One means to reduce the noise is the attachment of a serrated plate in the region of the trailing edge projecting over the trailing edge. The EP 1314885 discloses valuable information on wind turbine blades technology and performance.
Generally, it is desirable to improve the aerodynamic performance and efficiency of wind turbine blades, i.e. in low wind speed conditions. Moreover, the mechanical loading of the blade may be to be minimized. A secondary factor is the reduction of acoustic noise of the wind turbines blades when in operation.
Further relevant state of the art is disclosed in EP 1623111 B1, US 2011/0116927 A1, WO 2004/088130 A1 and EP 2034178 A2. The EP 1314885 discloses an apparatus improving the efficiency of a wind turbine with a panel connected to the trailing edge of the wind turbine blade. EP 2034178 A2 discloses a fairing plate to avoid an air gap when the flap is deflected, but which has no mechanical function.
|
{
"pile_set_name": "USPTO Backgrounds"
}
|
Q:
loop not going to next page
I just made a python script that goes to a lawyers profile to scrape their details. It is working fine for first page but the loop is not going to second page. The script is scraping data from first page only. I want to scrape all pages. Please help me I am new to python.
Here is the code:
import requests
from lxml import html
root_url = 'http://lawyerlist.com.au/'
def get_page_urls():
for no in ('1','2'):
page = requests.get('http://lawyerlist.com.au/lawyers.aspx?city=Sydney&Page=' + no)
tree = html.fromstring(page.text)
return (tree.xpath('//td/a/@href'))
for li in (get_page_urls()):
pag=requests.get(root_url + li)
doc = html.fromstring(pag.text)
for name in doc.xpath('//tr/td/h1/text()'):
print(name)
A:
The get_page_urls function only returns the urls of the first page because of the return statement in the for loop. Use a yield statement to turn the function into a generator and then iterate over each page of urls like so:
import requests
from lxml import html
root_url = 'http://lawyerlist.com.au/'
def get_page_urls():
for no in ('1','2'):
page = requests.get('http://lawyerlist.com.au/lawyers.aspx?city=Sydney&Page=' + no)
tree = html.fromstring(page.text)
yield tree.xpath('//td/a/@href')
for page_of_urls in get_page_urls():
for li in page_of_urls:
pag=requests.get(root_url + li)
doc = html.fromstring(pag.text)
for name in doc.xpath('//tr/td/h1/text()'):
print(name)
|
{
"pile_set_name": "StackExchange"
}
|
<?xml version='1.0' encoding='utf-8'?>
<!DOCTYPE hibernate-configuration
PUBLIC "-//Hibernate/Hibernate Configuration DTD//EN"
"http://www.hibernate.org/dtd/hibernate-configuration-3.0.dtd">
<hibernate-configuration>
<session-factory name="java:Pentaho/SessionFactory" >
<!-- CacheProviderClass -->
<property name="connection.datasource">Hibernate</property>
<property name="dialect">org.hibernate.dialect.HSQLDialect</property>
<property name="transaction.factory_class">org.hibernate.transaction.JTATransactionFactory</property>
<property name="transaction.manager_lookup_class">org.hibernate.transaction.JBossTransactionManagerLookup</property>
<property name="show_sql">false</property>
<property name="hibernate.cache.use_query_cache">true</property>
<property name="hibernate.generate_statistics">true</property>
<property name="session_factory_name">hibernate/HibernateFactory</property>
<property name="hibernate.jdbc.use_streams_for_binary">true</property>
</session-factory>
</hibernate-configuration>
|
{
"pile_set_name": "Github"
}
|
Hill Casino Erfahrungen Test Der..
quasar gaming
Rotating Reels Slot Feature and Free to Play Games
Online slots are fun for many reasons: the thrill of waiting for the reels to stop spinning, the interaction between player and machine, and (on modern video bonus titles) the video game-style play. The machine itself is exciting, with lights flashing, animations and video clips, and sound effects alerting you to a win. Yes, the. Awesome Spinning 7's slot demo created by Amaya Gaming is nice video slot, which looks like classic slot. It has 5 reels and 3 rows with numerous paylines and interactive bonuses, but you can easily confuse it with a classic slot, because of its symbols. Indeed, icons, which are your source of income in this free slot game. It is easy to play free slots online, you just need to find a game you like and start spinning the reels. Your goal is simple: Placing the same symbols on a pay line side by side in a certain number according to the machine you are playing. The amount of prize you win is determined according to the payout table, and you have.
To score a win, you just need luck. Betzest Casino goes live this November Not many casinos receive the kind of attention and reviews that Betzest casino is getting even before it is officially opened to the world. This is actually an advantage... Even though the expectation might be that Rotating Reels will give players the chance to earn more payouts, the reality is different. Benefits of Rotating Reels Rotating Reels mix things up a bit. Wild cards are popular with players, too, but not as popular as scatter symbols and bonus symbols.
Rotating Reels Slot Feature and Free to Play Games - Online Roulette
I always make small bets, to spin the slots more and more, 0. Free slots are good for casual gaming, but if you want to take risks and make a profit, you should choose to play slots with real money. Given that this is a pretty easy to understand feature, and one that obviously makes for a more exciting bonus round, we hope that other developers will find ways of incorporating it into their own games. Your geographical location, your computer's features, your... It seems a certainty more games will use the feature in the near future but its impact on the industry is likely to be short and sweet. We encourage you to start playing for free and when it is time to move on to the real money games, visit our online casino reviews page. Whether you play a game for free or with real money, the rules are the same. Popular Slots with Rotating Reels Due to the limited number of slots that have used the Rotating Reels feature, there is not a wide choice of games to choose from when it comes to picking them out. Power Cash Lucky enough for a pile of cash? The random wild functions mostly in the same way as any other, by transforming symbols, the difference is in the way that it appears, which as the name might give away — is entirely random, depending on the game theme and structure. On the other hand, twelve scatter symbols on the reels have the chance of winning you big. Steam Tower Steam your way to the jackpot. Secret of the Stones. Nudges Largely forgotten in the modern ways win format of both online and land casino slots is the classic nudge feature. Reel Bandits Bandits riding into a jackpot sunset. Genesis Gaming are the biggest name to do so but Booming Games have also entered the market with their own version. These free spins are triggered in various ways though normally it involves hitting 3 or more free spin or scatter symbols on the reels. Paylines Pay lines are one of the fundamental features of all slot machines. A certain portion of every Ruby Slippers - Rizk Casino goes toward an accumulating jackpot. All reel slots games that can be played on mobile devices mobile phone, tablet PC, etc. However, it's also important to remember that some slots require you to play the 'max bet' to be eligible for the jackpot. Where Play The Money Drop Slots Online at Casino.com India game, either randomly or through a special symbol, would allow you the option to move down a reel either singly or a multiple of times to possibly form a new winning combination, this was mostly manual, but some games featured auto-nudging too. They are a rare breed on the slots market but it is a feature that is fast capturing the imagination of players. More Languages - English Français. Established in 1980, Novomatic has a collection of more than 400 titles. The withdrawal was simple, fast and secure, as promised!
|
{
"pile_set_name": "Pile-CC"
}
|
Ethnic differences in trichotillomania: phenomenology, interference, impairment, and treatment efficacy.
In this Internet study, we explore differences among minority and Caucasian participants in the phenomenology of, interference and impairment related to, and perceived efficacy of treatments for trichotillomania (TTM) symptoms. A demographic difference was found for number of children only. Results indicate that the minority sample was less likely to report pulling from their eyebrows and eyelashes than the Caucasian sample. Minorities were less likely to report increased tension before a pulling episode. Minorities reported high levels of TTM interference with home management but Caucasians reported higher TTM interference with their academic life. Caucasians with TTM reported higher daily stress than their minority counterparts. Although minorities were less likely to utilize treatment, no significant differences were found for treatment improvement. This Internet study sheds important light on differences in TTM symptoms among minorities and Caucasians. The lack of economic and education differences between groups is a strength of this research.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
# Copyright 2016 Telefonica Investigacion y Desarrollo, S.A.U
#
# This file is part of Orion Context Broker.
#
# Orion Context Broker is free software: you can redistribute it and/or
# modify it under the terms of the GNU Affero General Public License as
# published by the Free Software Foundation, either version 3 of the
# License, or (at your option) any later version.
#
# Orion Context Broker is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero
# General Public License for more details.
#
# You should have received a copy of the GNU Affero General Public License
# along with Orion Context Broker. If not, see http://www.gnu.org/licenses/.
#
# For those usages not covered by this license please contact with
# iot_support at tid dot es
# VALGRIND_READY - to mark the test ready for valgrindTestSuite.sh
--NAME--
String filters for compound values of metadata: string match in subscriptions
--SHELL-INIT--
dbInit CB
brokerStart CB
accumulatorStart --pretty-print
--SHELL--
#
# 01. Create E1 with A1 with metadata M1 with compound value containing String, Integer and Float
# 02. Beginning: Create subscription 1 for mq=A1.M1.s~=test - notification
# 03. Dump and reset accumulator, see one notification
# 04. End: Create subscription 2 for mq=A1.M1.s~=ing - notification
# 05. Dump and reset accumulator, see one notification
# 06. Single-char: Create subscription 3 for mq=A1.M1.s~=_ - notification
# 07. Dump and reset accumulator, see one notification
# 08. No match: Create subscription 4 for mq=A1.M1.s~=fest_string - NO notification
# 09. Dump and reset accumulator, see no notification
# 10. Modify E1/A1 - provoke 3 notifications
# 11. Dump and reset accumulator, see 3 notifications
#
echo "01. Create E1 with A1 with metadata M1 with compound value containing String, Integer and Float"
echo "==============================================================================================="
payload='{
"id": "E1",
"A1": {
"value": "not important",
"metadata": {
"M1": {
"value": {
"i": 10,
"f": 3.14,
"s": "test_string",
"modified": false
}
}
}
}
}'
orionCurl --url /v2/entities --payload "${payload}"
echo
echo
echo "02. Beginning: Create subscription 1 for mq=A1.M1.s~=test - notification"
echo "========================================================================"
payload='{
"subject": {
"entities": [
{
"idPattern": "E.*"
}
],
"condition": {
"attrs": [ ],
"expression": {
"mq": "A1.M1.s~=test"
}
}
},
"notification": {
"http": { "url": "http://localhost:'$LISTENER_PORT'/notify" },
"attrs": [ ]
},
"expires": "2050-04-05T14:00:00.00Z"
}'
orionCurl --url /v2/subscriptions --payload "${payload}"
echo
echo
echo "03. Dump and reset accumulator, see one notification"
echo "===================================================="
accumulatorDump
accumulatorReset
echo
echo
echo "04. End: Create subscription 2 for mq=A1.M1.s~=ing - notification"
echo "================================================================="
payload='{
"subject": {
"entities": [
{
"idPattern": "E.*"
}
],
"condition": {
"attrs": [ ],
"expression": {
"mq": "A1.M1.s~=ing"
}
}
},
"notification": {
"http": { "url": "http://localhost:'$LISTENER_PORT'/notify" },
"attrs": [ ]
},
"expires": "2050-04-05T14:00:00.00Z"
}'
orionCurl --url /v2/subscriptions --payload "${payload}"
echo
echo
echo "05. Dump and reset accumulator, see one notification"
echo "===================================================="
accumulatorDump
accumulatorReset
echo
echo
echo "06. Single-char: Create subscription 3 for mq=A1.M1.s~=_ - notification"
echo "======================================================================="
payload='{
"subject": {
"entities": [
{
"idPattern": "E.*"
}
],
"condition": {
"attrs": [ ],
"expression": {
"mq": "A1.M1.s~=_"
}
}
},
"notification": {
"http": { "url": "http://localhost:'$LISTENER_PORT'/notify" },
"attrs": [ ]
},
"expires": "2050-04-05T14:00:00.00Z"
}'
orionCurl --url /v2/subscriptions --payload "${payload}"
echo
echo
echo "07. Dump and reset accumulator, see one notification"
echo "===================================================="
accumulatorDump
accumulatorReset
echo
echo
echo "08. No match: Create subscription 4 for mq=A1.M1.s~=fest_string - NO notification"
echo "================================================================================="
payload='{
"subject": {
"entities": [
{
"idPattern": "E.*"
}
],
"condition": {
"attrs": [ ],
"expression": {
"mq": "A1.M1.s~=fest_string"
}
}
},
"notification": {
"http": { "url": "http://localhost:'$LISTENER_PORT'/notify" },
"attrs": [ ]
},
"expires": "2050-04-05T14:00:00.00Z"
}'
orionCurl --url /v2/subscriptions --payload "${payload}"
echo
echo
echo "09. Dump and reset accumulator, see no notification"
echo "==================================================="
accumulatorDump
accumulatorReset
echo
echo
echo "10. Modify E1/A1 - provoke 3 notifications"
echo "=========================================="
payload='{
"A1": {
"value": "modified",
"metadata": {
"M1": {
"value": {
"i": 10,
"f": 3.14,
"s": "test_string",
"modified": true
}
}
}
}
}'
orionCurl --url /v2/entities/E1/attrs --payload "${payload}" -X PUT
echo
echo
echo "11. Dump and reset accumulator, see 3 notifications"
echo "==================================================="
accumulatorDump
accumulatorReset
echo
echo
--REGEXPECT--
01. Create E1 with A1 with metadata M1 with compound value containing String, Integer and Float
===============================================================================================
HTTP/1.1 201 Created
Content-Length: 0
Location: /v2/entities/E1?type=Thing
Fiware-Correlator: REGEX([0-9a-f\-]{36})
Date: REGEX(.*)
02. Beginning: Create subscription 1 for mq=A1.M1.s~=test - notification
========================================================================
HTTP/1.1 201 Created
Content-Length: 0
Location: /v2/subscriptions/REGEX([0-9a-f]{24})
Fiware-Correlator: REGEX([0-9a-f\-]{36})
Date: REGEX(.*)
03. Dump and reset accumulator, see one notification
====================================================
POST http://localhost:REGEX(\d+)/notify
Fiware-Servicepath: /
Content-Length: 232
User-Agent: orion/REGEX(\d+\.\d+\.\d+.*)
Ngsiv2-Attrsformat: normalized
Host: localhost:REGEX(\d+)
Accept: application/json
Content-Type: application/json; charset=utf-8
Fiware-Correlator: REGEX([0-9a-f\-]{36})
{
"data": [
{
"A1": {
"metadata": {
"M1": {
"type": "StructuredValue",
"value": {
"f": 3.14,
"i": 10,
"modified": false,
"s": "test_string"
}
}
},
"type": "Text",
"value": "not important"
},
"id": "E1",
"type": "Thing"
}
],
"subscriptionId": "REGEX([0-9a-f]{24})"
}
=======================================
04. End: Create subscription 2 for mq=A1.M1.s~=ing - notification
=================================================================
HTTP/1.1 201 Created
Content-Length: 0
Location: /v2/subscriptions/REGEX([0-9a-f]{24})
Fiware-Correlator: REGEX([0-9a-f\-]{36})
Date: REGEX(.*)
05. Dump and reset accumulator, see one notification
====================================================
POST http://localhost:REGEX(\d+)/notify
Fiware-Servicepath: /
Content-Length: 232
User-Agent: orion/REGEX(\d+\.\d+\.\d+.*)
Ngsiv2-Attrsformat: normalized
Host: localhost:REGEX(\d+)
Accept: application/json
Content-Type: application/json; charset=utf-8
Fiware-Correlator: REGEX([0-9a-f\-]{36})
{
"data": [
{
"A1": {
"metadata": {
"M1": {
"type": "StructuredValue",
"value": {
"f": 3.14,
"i": 10,
"modified": false,
"s": "test_string"
}
}
},
"type": "Text",
"value": "not important"
},
"id": "E1",
"type": "Thing"
}
],
"subscriptionId": "REGEX([0-9a-f]{24})"
}
=======================================
06. Single-char: Create subscription 3 for mq=A1.M1.s~=_ - notification
=======================================================================
HTTP/1.1 201 Created
Content-Length: 0
Location: /v2/subscriptions/REGEX([0-9a-f]{24})
Fiware-Correlator: REGEX([0-9a-f\-]{36})
Date: REGEX(.*)
07. Dump and reset accumulator, see one notification
====================================================
POST http://localhost:REGEX(\d+)/notify
Fiware-Servicepath: /
Content-Length: 232
User-Agent: orion/REGEX(\d+\.\d+\.\d+.*)
Ngsiv2-Attrsformat: normalized
Host: localhost:REGEX(\d+)
Accept: application/json
Content-Type: application/json; charset=utf-8
Fiware-Correlator: REGEX([0-9a-f\-]{36})
{
"data": [
{
"A1": {
"metadata": {
"M1": {
"type": "StructuredValue",
"value": {
"f": 3.14,
"i": 10,
"modified": false,
"s": "test_string"
}
}
},
"type": "Text",
"value": "not important"
},
"id": "E1",
"type": "Thing"
}
],
"subscriptionId": "REGEX([0-9a-f]{24})"
}
=======================================
08. No match: Create subscription 4 for mq=A1.M1.s~=fest_string - NO notification
=================================================================================
HTTP/1.1 201 Created
Content-Length: 0
Location: /v2/subscriptions/REGEX([0-9a-f]{24})
Fiware-Correlator: REGEX([0-9a-f\-]{36})
Date: REGEX(.*)
09. Dump and reset accumulator, see no notification
===================================================
10. Modify E1/A1 - provoke 3 notifications
==========================================
HTTP/1.1 204 No Content
Fiware-Correlator: REGEX([0-9a-f\-]{36})
Date: REGEX(.*)
11. Dump and reset accumulator, see 3 notifications
===================================================
POST http://localhost:REGEX(\d+)/notify
Fiware-Servicepath: /
Content-Length: 226
User-Agent: orion/REGEX(\d+\.\d+\.\d+.*)
Ngsiv2-Attrsformat: normalized
Host: localhost:REGEX(\d+)
Accept: application/json
Content-Type: application/json; charset=utf-8
Fiware-Correlator: REGEX([0-9a-f\-]{36})
{
"data": [
{
"A1": {
"metadata": {
"M1": {
"type": "StructuredValue",
"value": {
"f": 3.14,
"i": 10,
"modified": true,
"s": "test_string"
}
}
},
"type": "Text",
"value": "modified"
},
"id": "E1",
"type": "Thing"
}
],
"subscriptionId": "REGEX([0-9a-f]{24})"
}
=======================================
POST http://localhost:REGEX(\d+)/notify
Fiware-Servicepath: /
Content-Length: 226
User-Agent: orion/REGEX(\d+\.\d+\.\d+.*)
Ngsiv2-Attrsformat: normalized
Host: localhost:REGEX(\d+)
Accept: application/json
Content-Type: application/json; charset=utf-8
Fiware-Correlator: REGEX([0-9a-f\-]{36})
{
"data": [
{
"A1": {
"metadata": {
"M1": {
"type": "StructuredValue",
"value": {
"f": 3.14,
"i": 10,
"modified": true,
"s": "test_string"
}
}
},
"type": "Text",
"value": "modified"
},
"id": "E1",
"type": "Thing"
}
],
"subscriptionId": "REGEX([0-9a-f]{24})"
}
=======================================
POST http://localhost:REGEX(\d+)/notify
Fiware-Servicepath: /
Content-Length: 226
User-Agent: orion/REGEX(\d+\.\d+\.\d+.*)
Ngsiv2-Attrsformat: normalized
Host: localhost:REGEX(\d+)
Accept: application/json
Content-Type: application/json; charset=utf-8
Fiware-Correlator: REGEX([0-9a-f\-]{36})
{
"data": [
{
"A1": {
"metadata": {
"M1": {
"type": "StructuredValue",
"value": {
"f": 3.14,
"i": 10,
"modified": true,
"s": "test_string"
}
}
},
"type": "Text",
"value": "modified"
},
"id": "E1",
"type": "Thing"
}
],
"subscriptionId": "REGEX([0-9a-f]{24})"
}
=======================================
--TEARDOWN--
brokerStop CB
accumulatorStop
dbDrop CB
|
{
"pile_set_name": "Github"
}
|
Speak english muddafugga……
Man listen…….
I get this weird out the blue phone call tonight from Citibank talking about my account past due heading towards delinquent status. Delinquent?!?! Hold the fuck up? I pay my shit on time every month! What you mean delinquent??!?!?
She goes on to explain it and get a payment out of me. Okay, cool, whatever. Get off the phone. I’m checking my account on-line and it has a big phat “$0” in the minimum payment due field. In fact I see all of my recent activity but still nothing under my minimum payment due.
Hmmm…..something’s weird……
So I call Citibank back. I’m explaining to the lady on the phone that I think something is wrong with the website because it’s not showing me a minimum payment due and I know I owe some money. Lady gives me my balance and minimum due. I tell lady I’m not seeing this on the website. She transfers me to a technical specialist.
Here’s where the shit goes wrong…
Immediately when the “specialist” gets on the phone, I hear the accent. She’s not from these parts. My call got transferred to someone over the Atlantic Ocean. DAMN….
As I’m trying to explain the problem for the 3rd time tonight, I tell her, my account info isn’t being updated. My card is due to expire but I haven’t registered the new card. (I figured out what the problem was). So I’m asking her if she can fix the website so the correct information is reflected and to give me a refund of the late fees charged and this dumb chick is telling me that I have the wrong login information.
WHAT THE FUCK ARE YOU TALKING ABOUT?!?!?
We went back and forth, like it was a 3 Stooges, who’s on first routine….Refund? No, login. No refund. No sir, login. Arrrgghhhhhhh…..
Then she goes off on this tangent about how to login to the website. I’m already logged into the website. How the fuck am I telling you that my account information is wrong if I’m not looking at the website?!?! So in the middle of the sentence, I hung up on her.
I called back, hoping, praying I got someone in the States who spoke English. I did. I explained the problem for the 4th time and in 2 minutes everything was resolved.
Yeah umm Mr. and Mrs. Corporations. I know help out there is cheaper but it also irritates me and a million other people in this country. Any economy will struggle if you TAKE AWAY jobs. You don’t need Alan Greenspan or whoever the new guy is, to tell you that.
|
{
"pile_set_name": "Pile-CC"
}
|
An analysis of functional communication training as an empirically supported treatment for problem behavior displayed by individuals with intellectual disabilities.
This paper examines the literature on the use of functional communication training (FCT) as a treatment for problem behavior displayed by individuals with intellectual disabilities (ID). Criteria for empirically supported treatments developed by Divisions 12 and 16 of the American Psychological Association (Kratochwill & Stoiber, 2002; Task Force, 1995) and adapted by Jennett and Hagopian (2008) for evaluation of single-case research studies were used to examine the support for FCT. Results indicated that FCT far exceeds criteria to be designated as a well-established treatment for problem behavior exhibited by children with ID and children with autism spectrum disorder, and can be characterized as probably efficacious with adults.
|
{
"pile_set_name": "PubMed Abstracts"
}
|
Q:
Why do some conferences have a variable amount of reviewers for each submitted paper?
Why do some conferences have a variable number of reviewers for each submitted paper?
For example, I read on https://redd.it/9nomv1 that the AAAI 2019 conference has between 3 and 5 reviewers for each paper submission. Why isn't the number of reviewers the same for all submitted papers? Is that to make up for non-experienced or outside-the-area reviewers, or for reviewers who do not turn in reviews?
A:
Editors usually invite more than the minimum amount of reviewers when they assign a paper for peer review, because not all invited reviewers will agree to do the review. A suggestion that I have seen in a guideline for editors (and which indeed mostly works well in my experience for this specific publication venue) is to invite 5 or 6 reviewers initially when aiming to get 2 reviews.
Occasionally an editor will be lucky and (almost) all invited reviewers will submit the review, leading to a higher than required number of reviews for this manuscript.
In other cases there will be less than the minimum number of reviews even though more reviewers have been invited, and the editor will have to invite even more reviewers.
|
{
"pile_set_name": "StackExchange"
}
|
Yoga is supposed to be about healing the body and calming the mind, but tonight one of the biggest yoga gurus in the world is facing allegations of sexually assaulting five of his followers. But why didn't they go to police until years later? What does it mean to be a guru? Pastor. Teacher. Reporter: A little more than a year ago, bikramoudhury invited "Nightline" to his house for an interview. He's proud of his success. He even gave us a tour of his garage full of luxury gars. This is 1953. His commercialism and brass personality have made him a controversial figure in the yoga world. But he has no time for his critics. People sometimes who are famous and well known, that group of people get jealous. People talk bad about Jesus also. Reporter: Since that interview, five women have come forward filing civil lawsuits alleging sexual assault. Four of them accuse him of rape. This woman asked us not to use her real name. In her lawsuit against bikram, she's identified as Jane doe number three. She has never told her story in public before. She said bikram raped her three times. She said the first was at teacher training. She will be given a scholarship and said he humbly thanked him for it. I was on my knees. It's's respectful to kneel in front of your master, your guru. He said I like you, I like you. And then he said it's cold. Let's go up. Then you're going up in the elevator alone with this man. I was in denial, praying that it would not be. Reporter: She says shortly after they entered his hotel suite, he made his intentions clear. And he said, you know how many people apply for the scholarship every training? And of all the people, you're here. Because I believe in to you. No one believed in me before. I looked at him as a godly figure. Like mother Teresa, you know? I said please don't. You're a guru. The world believes in you. He said nah, I'm just a man. Reporter: She said he raped her and then acted as though nothing happened. She continued, not determined to lose her chance to fulfill her dream but said she made sure they were never alone again at teacher training. After she completed the course, she moved in L.A. To work at bikram's headquarters. He made it very professional, I would be on the payroll, not volunteer. I believed him. I trusted. Reporter: Then she says he raped her a second time at this apartment upstairs on the bare mattress on the floor. He took his pants off. He pulled me and I said please don't. You said it wouldn't happen again. And then he pushed me down with force. And I said no. And he hit me. Reporter: He hit you? Then he forced himself on me. And so I just knew, I'm doomed. Reporter: When somebody says why didn't you go to the cops? Why did you come to L.A.? Why did you allow yourself to be alone with this man again. What do you say? I try not to be alone with him. But why not? Who would believe me? No one would. Everybody was so hypnotized by him. Reporter: She says she was financially dependent on bikram and believed her career as a yoga instructor was on the line. So, she says, she kept it to herself.
This transcript has been automatically generated and may not be 100% accurate.
|
{
"pile_set_name": "Pile-CC"
}
|
Rémi Tremblay by
T
The Académie Française, the exclusive and ancient institution tasked with safeguarding the French language, awarded Haitian novelist Dany Laferrièrem, on May 2015, the honor of being one of the "immortals" of French Canadian culture for publishing his novel How to Make Love to a Negro Without Getting Tired.
here is currently one million French Canadians living outside of Quebec, including around 300,000 Acadians. Over the last few decades, thanks mostly to political correctness, these communities came to redefine their identity on purely linguistic terms. The ethnic term "French Canadian", which described people of French descent, came to be understood as a Canadian speaking French without any reference to culture, ethnicity or religion. In the 1960s, before the Quiet Revolution in Quebec, Quebecers considered themselves to be French Canadians. On the other hand, Acadians, acknowledging their differences and their unique heritage, always rejected the term. There was a bond of solidarity unifying French Canadians living outside Quebec and those in the Belle Province. However, with the Quiet Revolution that brought about the rise of Quebec's neo-nationalism , a nationalism purely based on provincial status and autonomy demands, the cultural and ethnic solidarity that united French Canadians vanished. The communities outside Quebec found themselves isolated and without an ally to defend their linguistic rights like French schools and services.These communities came to see language as the sole difference between them and the majority of Canadians, totally forgetting that their identity could not be reduced to a single aspect of their life. The French Canadians' prime objective became to secure their linguistic rights, and when Ottawa started talking about promoting Francophone immigration in provinces where French minorities existed, the French Canadians welcomed the move and embraced the promotion of multicultural immigration, as long as it was French speaking immigrants. In 2002, the liberal Minister of Citizenship and Immigration, Denis Coderre, launched the Strategic Framework to Foster Immigration to Francophone Minority Communities.The leaders of French Canadian communities started championing higher immigration rates and trying to attract new immigrants to their communities in order to increase their percentage of the population and therefore give credence to their linguistic demands. If their communities increased in demographical weight, their demands could not be rejected. French Canadians, like most European derived people, faced a declining demography thanks to the low fertility rate that plagued them, and, like most European peoples, saw immigration as the miraculous solution to their problems.This seemingly logical strategy proved to be extremely short sighted for many reasons. First, French Canadians' rights were historically justifiable; French Canadians being one of the two founder peoples of Canada. Where they were a minority, they could always invoke their status as founders of Canada to back their claims, but if the French speaking communities become mostly made up of immigrants, why would these immigrants receive services in their own language while more numerous Chinese or Hindus be denied these same rights? Or, as French speaking immigrants start gaining ground over French Canadians, would they not start acting as the true representatives of the French heritage in Canada, pushing the Quebecois aside as senile leftovers from the past?Furthermore, these new immigrants joining the ranks of French Canadian communities seldom reduce their identity to the sole linguistic aspects. They consider themselves to be Muslims, Blacks, Haitians, North Africans, etc. The fact that French Canadians have sabotaged their own organizations to become a purely linguistic lobby ultimately works against them, because, as previously mentioned, without the historical argument, as a founding people who were instrumental in the creation of Canada's institutions, French Canadians have no more legitimacy than any other group.For instance, in Vancouver, Francophones represent one percent of the population while immigrant languages are spoken by 40%. Rights and privileges granted to Francophones for historical reasons become null and void if French communities become a solely linguistic club representing French speaking immigrants. Why would a freshly arrived Haitian receive services in his language and not the more numerous Chinese?The Sainte-Jeanne-d'Arc school in Edmonton also shows the failure of the immigration strategy for French Canadians. Immigrants represent around 20% of Edmonton's French community. When school resumed in September 2014, a petition demanding Black teachers was presented to the school administration. For the 500 people who had signed the letter, it was unthinkable that there was no Black teacher in this school. Therefore, it proves that immigrants, unlike French Canadians, do not reduce their identity to language only, but include racial identity. If French Canadians are ready to sacrifice their identity and consider themselves to be Canadians speaking French without any other specific interests to defend, other groups are not following them in this path.The case of the "Association canadienne-française de l'Ontario" that had to change its name for the less ethnically connoted term "Assemblée de la francophonie de l'Ontario", in order to show more inclusiveness, is also telling.We have also seen in the last decades a disturbing trend; we have seen French Canadians preferring to associate with Muslim Francophones rather than Anglo-Canadians despite the fact that they share the same religion, a similar culture and a common history. Language has become for too many of them the sole issue at stake. If French language rights are indeed primordial and must be defended, it must be done in an intelligent way, not against their ethnic and cultural interests.In fact, with immigration, French minorities became a less significant minority, its percentage of the population ever diminishing, even if we only take into consideration language. The place of French in Canada and even in Quebec is decreasing because of the combined effect of a low fertility rate and immigration.French Canadians are working against their very interest by adopting immigration as a strategy to further their goals; in the end, they will probably become extinct if they do not realize that their identity cannot be summarized by language.Related posts:
|
{
"pile_set_name": "OpenWebText2"
}
|
. I
THE ATTORNEY GENERAL
OP -XAS
April 3, 1968
Hon. Thomas C. Green, P.E. Opinion No. ~-216
Member Secretary, Texas State
Board of Registration for Rer Authority of the Texas
Professional Engineers State Board of Reglstra-
Austin, Texas 78701 tion for Professional
Engineers to pay from
Line Item No. 7 of Its
departmental approprla-
tlon, Senate Bill 15,
Acts 60th Legislature,
1967, Regular Session,
Ch. 784, p. 2175, trav-
eling expenses and the
purchasing of furniture
and necessary office
equipment and supplies
Dear Colonel Green: for its investigator.
You have requested the opinion of this office concerning
the authority of the Texas State Board of Registration for Pro-
fessional Engineers, hereinafter called 'Board" to pay from Line
Item No. 7 of the departmental appropriation to the Board, Senate
Bill 15, Acts 60th Legislature, 1967, Regular Session, Chapter 784,
page 2175, the General Appropriations Bill, the traveling expenses
of Its investigator as well as the purchase price of furniture and
necessary office equipment and supplies for its Investigator.
Line Item No. 7 of the department appropriation to the
Board provides:
“7. Investigation Program: Expenses
of witness fees, court transcripts, docu-
mentary evidence, processing complaints,
court costs, and other necessary expenses
in gathering evidence for the enforcement
- 1035 -
. .
Hon. Thomas C. Green, page 2 (M-216)
In Attorney General's Opinion No. M-179 (1968),
this
office held that:
"The Texas State Board of Registration
for Professional Engineers has authority to
employ an investigator to assist the Board
in the enforcement of Article 3271a, Vernon's
Civil Statutes, and such Investigator may be
compensated from the departmental appropria-
tion to the Texas State Board of Registration
for Professional Engineers in Line Item No. 7
of Senate Bill 15, Acts 60th LegIslatureA 1967,
Regular Session, Chapter 784, page 2175.
The authority to employ an investigator by the Board
and compensate him from the funds appropriated by Line Item 7 of
the departmental appropriation to the Board as being a necessary
expense In gathering evidence for the enforcement of Article 3271a,
Vernon's Civil Statutes, the Engineering Practice Act, carries
with it the further authority to make such expenditures as are
necessary to enable such investigator to carry out his duties of
gathering evidence for the enforcement of Article 3271a. Certainly,
the payment of the investigator’s traveling expenses, while per-
forming his official duties, would be a necessary expense in the
gathering of evidence to enforce Article 3271a; and if the Board
Is of the opinion that such investigator needs certain office
equipment and supplies to enable him to efficiently carry out
his investigative functions, then such expenditure could properly
be paid from the funds appropriated to the Board In Line Item No. 7.
SUMMARY
The Texas State Board of Registration for
Professional Engineers may properly pay from
Line Item No. 7 of Senate Bill 15, Acts 60th
Legislature, 1967, Regular Session, Chapter
784,page 2175, the traveling expenses of Its
investigator and the cost of furniture and
office equipment and supplies for its investigator.
V truly yours,
ney General of Texas
Prepared by Pat Bailey
Assistant Attorney General
-1036-
. .
Hon. Thomas C. Green, page 3 (M-216)
APPROVED:
OPINION COMMITTEE
Hawthorne Phillips, Chairman
Kerns Taylor, Co-Chairman
W. V. Geppert
Arthur Sandlln
John Banks
Wardlow Lane
A. J. CARUBBI, JR.
Executive Assistant
-1037-
|
{
"pile_set_name": "FreeLaw"
}
|
Welcome To The Top "Hunting Beagle" Sites
Connecting Beaglers with Beaglers for the betterment of the "Hunting Beagle" like no other site in the world!
Page 1: Top 100 Sites
Page 2 & 3: New Sites & Sites Contending To Reach The Top 100
NOTE: Do to the Overwhelming interest in "Hunting Beagle Websites" and the large number of websites added {over 300}, we NOW reset hits in and hits out to ZERO every 15 days. This is fair to new sites as it gives them equal opportunity to appear higher on the list; also it shows a more variety and recent popularity ranking.
Sites with "broken links" for more than 72 hours will be deleted. Sites without a contact number, email address, location or the identity of site owner, will be deleted.
By placing your site in the Top "Hunting Beagle" Sites, also submits your site to search engines free of charge.
BEAGLES THAT "HUNT TO JUMP THEN RUN TO CATCH" MARIE, BLUETICK DAUGHTER OF REGGIE BRED TO RUDY, BLUETICK SON OF JASPER.ACCEPTING PUPPY DEPOSITS .731 584-8333. 8 MO TICKLED BLUE PUP GOT 2ND AND 3RD IN CANADA FOR LIN MCVICAR AND GRANDSON KYLE IN ALL AGE TRI
Located in Western Kentucky. Breeding the total package line bred Buckshot Jones gundog beagles. Dark blues with outstanding conformation, extreme hunt, and head down straight up the middle controlled power.
A message board for rabbit hunters, field trialers, and anyone that owns, runs, or hunts with beagles. Pedigree generator. Forums include dogs and puppies for sale, beagle health, bloodlines and pedigrees, kennels, members websites,and more.
Benson's Kennel owner, Tremayne Benson, is known as one of the "Top Beagle Trainers Under 35". People travel across the USA to see his outstanding hounds. He’s known to have hard hunting, extreme drive, smart, and obedient hounds. Experience It Now!
My Beagle's bloodline goes all the way back to the beagle Hall of Fame. "If your looking for hunting beagles that jump & circle there own rabbit. Stay on the trail and aren't afraid to go in the brush, with the desire to hunt...THEN YOU'VE ARRIVED!
Beagles on Fire is located in Holly Springs, MS. We raise straight up upper medium speed quality gun dogs that have great hunt, run the track with good line control, & handle extremely well. We specialize in real good young started dogs sure to impress.
ARHA Hall of Fame Daddy Rabbit, aka Aubrey Holcombe, and the Daddy Rabbit Kennels. Daddy Rabbit Kennel is known all over the USA for producing the hardest hunting, briar busting, run to catch, hall of fame, red beagles.
I will professionally, safely and dependably haul your dogs for you! If you’re tired of: paying airport high rates, buying another crate or driving to the airport after waiting for the perfect temperature call, Buckeye Dog Hauling!
Welcome to Stillwater Valley Kennels (est.1987) home to AKC Reg. hard hunting and line driving gundogs and field trail beagles. Our goal is to breed and raise hounds that do what they are bred for produceng game. Thanks for visiting my site, Don Nash
Rabbit Running AKC Blackheaded Bluetick Beagles & Blackcreeks Beagles Located in Central Florida ,my lines consist of Oakhill,Gunsmoke, Blackcreek and Otis,they are great gun dogs with extreme hunt and desire and will hunt all day ! Take a look around th
Located in Salyersville, Ky. We have Primarily Gundogs. I Run Northway/Branko hounds primarily atm. Have some young Turbo dogs coming up too. Talk, look, buy, sell, trade Beagles. More information added soon.
The site tells info about my kennel, my dogs, and my rabbit hunts. We are located in Wrightsville, Georgia. My kennel is made up of some registered and grade dogs. Registered dogs are AKC and ARHA registered.
Located in Hampton, VA, Bayside Hunting Beagles is primarily based upon the old Yellow Creek strain, with influences such as Dingus Macrae & Gay Baker derived hounds. Our dogs are medium fast with great line control and aggressive hunt.
Our kennels are in Western NC. We raise AKC/UKC registered beagles. We have Branko, Primetime, Northway, and Buddy's Iron Eagle bloodlines. Our dogs are bred for line control, speed, and the brains to put it all together for trial or gun.
We raise, breed and sell AKC field trial quality rabbit hunting beagles. We are located in the upstate of South Carolina, about 25 minutes from the Georgia state line and and under and hour from the North Carolina line.
Swamp Bottom Beagles is a Christian based, rabbit-dog kennel, located in Coats, NC. The most predominate bloodline in our kennel, is a SHORTS PRO bloodline. We try to keep some puppies, and "started" dogs for sale, most of the time.
Southeast Missouri beaglers breeding hard-hitting rabbit dogs. We run several proven bloodlines including Patch, Turbo, and Pine Mt. Chester. We like red dogs, but as long as they will smoke a bunny, we're not picky.
Welcome to our site we are from the mountains of south east Kentucky we promote a medium speed line control gun dog that can run a rabbit under any condition and has the brains and nose to do so
Thanks for checking out our page
My name is Rinard Leonard and I'm located in Titus, AL. My dogs are hard-hunting,smart,
disciplined and run with the intent to catch. My bloodlines consist of Branko,JD,Tadpole,Bramlett,Patch and Weir Creek
Carders Creation Kennels is a small kennel in Greenville KY. We produce hard hitting, run for the front Rabbit hounds. Are beagles are Ukc and Akc registered. Pups available, 2 outstanding stud hounds available.
We raise upper medium 13-14 inch Chocolate and White Weir creek hounds that are bred to bust the briars in the South and run a steady smooth line. These hounds have a good nose with brains to use it. They have good jump ability and can latch on a rabbit.
Missouri Talk To Me Beagles derive from Dingus MacRae, Indian Hills, and Yellow Creek bloodlines. We strive for "complete" gundogs that can run in all conditions and have a balance of desire, nose & intelligence. Located in Kansas City, Missouri.
Denham's Beagles is owned and operated by Kyle and Robert Denham in central Missouri, our goal is to raise hard hunting beagles with alot of nose and brains, we also raise San Juan, Smokey Mountain Cottontail and Belgain Hare cross running rabbits, and ha
At Lowe Valley we put God first and give Him thanks for the little hunting companion that we call beagles. Most of our dogs are stubby/pro but we do have some that are bear bred. All are SPO style hounds that must hunt.
We sell 13 inch bluetick beagle puppies. We are a small family breeder not a puppy mill. Our dogs are of good quality they have champion bloodline. We are located in Lisle, New York. You can call us at (607) 743-1397 or visit us at azandjzbeagles.webs.com
We breed, raise, and train AKC & UKC registered beagles for rabbit hunting. We are a small family run kennel and we breed for medium to upper medium speed gun dogs with excellent line control, close check work, drive, brains, a big nose, and extreme hunt.
My name is Elliott L. Peters and the owner of EP's Northway Iron Jaw Kennels. In my kennel I breed for hounds that have hard hunt, fast foot, correct mouth, and the brains to put it all together and account for their game.
Welcome, this is a discussion forum dedicated to the hunt and pursuit of the snowshoe hare and the hounds, hunters and breeders devoted to this endeavor. The HUNTED HARE is uninterested in trials or titles, just the "average joes" and their dogs.
Our kennel is located on 78 acres of woods and fields in beautiful Northeastern Pennsylvania. It contains a running pen and two puppy pens plus a rabbit breeding pen to maintain our rabbit population in the starting pen.
|
{
"pile_set_name": "Pile-CC"
}
|
We just published this month’s IPv6 network operator measurements and as usual they include exciting developments in IPv6 deployment from all around the globe.
Vodafone Portugal – Comunicações Pessoais, S.A. (AS12353), a full subsidiary of the Vodafone Group, is the second mobile operator in Portugal, both chronologically and in market share. This month’s IPv6 network operator measurements show Vodafone Portugal ranked 92nd with a rapidly growing IPv6 deployment percentage of 4.17%.
You can view the full listing of newly updated IPv6 network operator measurements for this month.
If you’re a network operator deploying IPv6 and would like to join Vodafone Portugal and the other networks that make up the ranks of World IPv6 Launch participants, please register your network for measurement.
|
{
"pile_set_name": "OpenWebText2"
}
|
The last thing that you want to happen when you are on an interview is to lose the job because of the way that you are acting. You will want to be on your best behavior during the interview, so that the people who you are speaking with will be impressed with you, not put off by you.
Be Polite
Make sure that you are always polite and kind when in an interview. Bring out all of your manners and treat the ones you are talking to with respect. Say “ma’am” and “sir”, if appropriate, and make sure to shake their hands and look interviewers in the eyes. By just using these simple tips, you will have a much better chance of getting the job.
Start the Right Way
It doesn’t hurt anything to start everything off on the right foot, either. If you do end up getting the job that you have interviewed for, then you will be on a good page with your new bosses. They will think of you as being someone polite and someone who cares about their job, and that will be a good thing.
Set Expectations from the Interview
Being on your best behavior during the interview will encourage you to keep up with the good behavior once you have started with your new job. This will help you to keep the job. If you always do the right thing for your boss and co-workers, they’ll be invested in your success, too.
If you need help crafting the perfect resume and cover letter, chose our Career Support Bundle for a resume, cover letter, thank you letter, prospect letter and follow-up letter.
|
{
"pile_set_name": "Pile-CC"
}
|
A wellhead Christmas tree has a plurality of outlets where valves are placed in connection with these outlets for control of these outlets. These valves are then mounted externally on a large structure and may be subject to damage, for example in the event of falling loads. If the valve is damaged and/or is also completely knocked off, the outlet from the wellhead Christmas tree will be open to the environment. This will result in spillage to the environment and is therefore not desirable. A possible solution to this is to provide an extra barrier valve internally in the actual bore outlet of the wellhead Christmas tree leading out to the valve which may be damaged. This extra barrier valve will have a fail safe closed position. A solution of this kind is described, for example, in the applicant's own Norwegian application NO20080211, entered as PCT application with publication number WO2009102214. This ensures that in the event of damage due to falling loads there is no spillage to the environment.
In order to install an extra valve in the bore outlet, the external valve normally has to be uncoupled in order to insert a distance piece between the wellhead Christmas tree and the outer valve, which distance piece is intended for use as a hydraulic chamber for the extra internal valve. Furthermore, the pipe system outside the Christmas tree has to be adapted so as to make room for this distance piece. This distance piece must be included in the assembly as the extra valve which has to be placed in the bore outlet has a control unit/actuator extending slightly beyond the valve and therefore outside the Christmas tree. This control unit/actuator may therefore extend into the main channel in towards the main valve and may possibly hinder closing/opening of the main valve. Thus it is necessary to have a distance piece, and in addition it is also necessary to supply control fluid to the actuator which is to be placed in the bore outlet. This process of installing a distance piece is time-consuming and there is therefore a need for a simpler solution for installation of a valve in the bore outlet in a wellhead Christmas tree.
|
{
"pile_set_name": "USPTO Backgrounds"
}
|
Search form
What We Do
ABE is an innovative science education program that introduces secondary school students to the excitement of scientific discovery. ABE provides teachers with the loan of research-grade equipment, supplies, curriculum, and professional development at no cost.
ABE provides high-quality laboratory experiences that promote students’ interest in science and science careers, is aligned with curriculum standards, and explores the development of medicines using technologies employed by the biotechnology industry. The program features a hands-on molecular biology curriculum that introduces students to the excitement of scientific discovery.
Participating teachers receive a loaned kit (free of charge), with research-grade equipment and supplies that allow students to participate in advanced science laboratories. Professional development for ABE teachers enables them to keep pace with scientific advances and technologies and to introduce their students to the latest scientific discoveries. Teachers and students alike experience the excitement of participating in cutting-edge science with significant real-world applications.
Learn more about the exciting advances in the pharmabiotech industry at Amgen Science.
Total Student Participation
~700k
Teacher Participation for 2017-18 School Year
1,495
Amgen Biotech Experience is an international program funded by the Amgen Foundation with direction and technical assistance provided by Education Development Center (EDC).
|
{
"pile_set_name": "Pile-CC"
}
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.