Alaamer/medium-articles-posts-with-content

0

null

0

f5af2b715248

2018-01-02

2018-01-02 10:39:34

2018-01-02

2018-01-02 10:42:13

6

false

en

2018-05-25

2018-05-25 10:48:27

14

19115411af54

5.636792

17

4

0

Today’s companies are racing towards artificial intelligence to make it a big part of their digital strategy. The rise of chatbots in…

5

Important use cases of AI-powered chatbots in the banking industry Today’s companies are racing towards artificial intelligence to make it a big part of their digital strategy. The rise of chatbots in finance and banking sectors is the latest wave of new technologies adopted. Particularly in the banking industry, it is changing the face of the communication interface by adopting Artificial Intelligence. Leveraging the strength of artificial intelligence and increasing popularity of messaging apps, conversational interfaces are enabling unprecedented banking engagement and re-establishing relationship banking. WHY THE BANKING INDUSTRY NEEDS CHATBOT? Introducing chatbots in the banking sector can bring a huge change in customer experience and keep up the pace with changing customer expectations. Chatbot has the potential to automate all the repetitive questions which are time-consuming and has a huge impact on the department’s performance. No matter the use-case, banks are now stepping forward to use chatbots to simplify the overall banking experience for the customers. BANKING CHATBOT USE CASES: PERSONAL BANKING SERVICES Gone are the days of standing in long queues at the bank and filling out paperwork to access general banking services. AI Chatbots for banking Introducing chatbots in the banking industry improves overall customer satisfaction and engagement. Customers can check account balance or simply ask for a statement of the transactions using a simple interface with the help of chatbots. To execute these account-related activities, a unique identifier is provided by the customer to get authorized and access account data. Transferring funds to an inter-bank or third party by certain authentication methods can save customers time and workload to bankers. Customers can also get a quick view of their earnings and expenditure from customers previous data and the plotted graph can show how much they will spend in coming months. UNINTERRUPTED CUSTOMER SUPPORT Customers served with a most personalized approach is the key to growth. Without customer satisfaction, no organization can sustain for long in the market. In the banking industry, it is necessary to provide 24×7 customer support. Chatbots will help with tasks like resolving queries, options to update client KYC, provide information on new schemes and services around the clock. They ensure that the customer’s queries are solved in the shortest period and never let customers feel that they are interacting with a machine. CUSTOMER FEEDBACK & MEASUREMENTS All the branches around the globe will analyze their individual usage rate on the available schemes and obtain feedback from each customer. Based on the overall feedback, management can further consider refining existing schemes or implement new plans. Digital collaboration networks like intranet form the basic building block for gathering meaningful insights to gain business efficiency and smooth workflows. Imagine a chatbot on such intranets helping employees access the information with just a conversation. Intranet-based chatbot can help employees for better internal communication to gather insights from different branches and help the core management to take further innovative steps. Hence, it gives a win-win for both employees in gaining meaningful insights and banks in gaining productivity. “By 2022, 40% of customer-facing employees and government workers will consult daily an AI virtual support agent for decision or process support.” Download the free eBook DELIVERING PERSONALIZED MARKETING The banking industry has a wide range of products and services for its customers. But, not all customers are attentive to every service. Chatbots can deliver personalized offers to customers based on their profile data or life events. Highly-targeted products and services are brought to the customers at the right time by the preferred messaging apps which can intensely increase conversion rates. By using this channel for communication, banks can achieve a higher market value without annoying the customer. EMPLOYEE SELF-SERVICE An employee should login to HRMS and raise a request to update his details or access personal records or payroll details or transact with the Human Resource officials personally. If this process is carried out using chatbots, employees can apply leave, access personal details, payroll details, update contact details, reviewing of the timesheet, requesting for overtime payment, viewing of compensation history and submitting of reimbursement slips without much of human intervention. Employees can chat with the bot and ask to raise a request on their behalf so that they can better utilize their work hours and increase productivity. Employee self-service portal may be operated on an intranet or via a web service. BENEFITS OF USING CHATBOTS IN BANKING INDUSTRY: Chatbots are the new 24/7 customer service tool that can operate without any human interface once they’re set up. Help to automate fraud prevention processes and collect critical information from potentially impacted bank users. Push relevant content to end users and analyze user engagement. Lead an organization’s personalized methodology and create incremental income. Makes your brand identity more consistent with one voice, one message, one tone for each client. Run smoothly during peak traffic times, which means that the response rate will remain consistent, resulting a great user experience. HERE ARE SOME RECENT EXAMPLES OF THE WORLD’S BIGGEST BANKS WHICH ARE USING CHATBOTS TO BOOST THEIR BUSINESS: BANK OF AMERICA: Being one of the biggest banks in the USA, Bank of America (BoA) is riding the tide of AI-powered chatbots in the financial sector. A year ago, the organization introduced Erica, a voice-and-text empowered chatbot for clients. Erica had commended as an advanced virtual assistant to help clients make smarter decisions. Erica helped in sending notifications, suggests ideas how a customer can save money, gives reports on their FICO score, and encourages payment of bills within the banking application. JPMORGAN CHASE: Even though chatbots are often used to automate repetitive tasks, the biggest U.S. bank JPMorgan is utilizing bots to streamline its back-office operations. They recently launched COIN, a bot which can analyze complex legitimate contracts quicker and more proficiently than human lawyers. Since its launch a year ago, the bot has helped JPMorgan spare more than 360,000 hours of labor. This chatbot additionally uses the technology to parse messages for employees, allow access to software systems, and handle basic IT requests like resetting passwords. Later, the organization intended to keep using bots to find a new source of income, decrease expenses and reduce the risks. CAPITAL ONE: Capital One introduced Eno, a text-enabled chatbot that helps clients to deal with their money using their mobiles. Clients can get information from the chatbot about their account balance, transaction history, and credit limit as an instant message. It can even enable clients to pay their credit card bill in no time. Eno is the second of Capital One’s virtual assistants after the organization launched its own Amazon Alexa a year ago. Alexa is a virtual assistant accepting inputs in the form of voice commands. This bot enables Capital One clients to know about upcoming payments, check account balances, and pay their credit card bill using their voice. Customers can even know the transaction history based on it, so that they can manage their future expenditure. Download the free eBook CONCLUSION Unlike the traditional banking methods, chatbots can bring in a better and faster user experience providing 24×7 intelligent customer service. Chatbots helps in streamlining the operations, automate customer support, and provide a more convenient and enjoyable customer experience. Minimal setup, easy integration, and accessibility via a conversation medium are the key drivers in chatbot adoption. Source acuvate.com This story is published in The Startup, Medium’s largest entrepreneurship publication followed by 277,994+ people. Subscribe to receive our top stories here.

Important use cases of AI-powered chatbots in the banking industry

79

important-use-cases-of-ai-powered-chatbots-in-the-banking-industry-19115411af54

2018-06-05

2018-06-05 05:24:31

https://medium.com/s/story/important-use-cases-of-ai-powered-chatbots-in-the-banking-industry-19115411af54

false

1,242

Medium's largest publication for makers. Subscribe to receive our top stories here → https://goo.gl/zHcLJi

null

The Startup

null

swlh

STARTUP,TECH,ENTREPRENEURSHIP,DESIGN,LIFE

thestartup_

Chatbots

chatbots

Chatbots

15,820

Veeranjaneyulu Chettupalli

SEO Consultant | Analytics | Social Media Expert | Big Data enthusiast https://buff.ly/2uQc1ku

2dbcc0c939d4

veeran

66

286

20,181,104

null

0

null

0

d19c41fec9df

2018-04-09

2018-04-09 14:05:29

2018-04-09

2018-04-09 14:11:45

5

false

en

2018-07-13

2018-07-13 17:34:04

10

1912948147a3

5.199371

0

Consumers stick around when their evolving expectations are met.

5

Brand Loyalty is Deeper Than the Product In a world where quality products are widely accessible, brands can no longer count on brand loyalty existing the way it used to. There have been a lot of articles over the past few years discussing the death of brand loyalty as we know it. Forbes uses Tide, which is the perfect example of what brand loyalty used to look like. Just think about it. Moms used Tide not just because it was the best cleaning option, but they saw it as the only option. So, what did their children do? They bought Tide just like mom and dad — a testament to how loyalty and belief can become ingrained in us. What changed? With the evolving tech and media landscape, more products are at consumers’ fingertips. Gone are the days where limited options and lack of reviews kept consumers from trying new products. Brands are now expected to work for the business of their customers. If they’re willing to do that, then customer loyalty is achievable, but it will never be like it once was. It’s a target in constant motion that can only be struck when brands bring Consumer Experience to the forefront. If loyalty doesn’t exist like it used to, then what is it now? Customers expect more from brands than ever before. Top-notch products are no longer an option, they’re an expectation. When brands can’t rely on their products to differentiate themselves in the marketplace, they’re forced to dig deeper and strategically market themselves to attract today’s new tech-savvy and perceptive customers. CAUSE MARKETING Consumers like brands that have a voice, the ones who are openly involved in the community. Large, faceless corporations are now unsavory. How are they bettering the world? Cause Marketing is at its peak and people are responding positively. A poll revealed that 90% of consumers would buy the cause-branded products if given the choice between two similarly priced products. Cause Marketing can be profitable and brand building at the same time, too. Take Walgreen’s Red Nose Day for example — a fun, playful campaign. Ads featured celebrities from all different backgrounds, united by a singular cause of raising money for children in poor communities across the globe. Fans were encouraged to show their support by purchasing red noses and posting pictures of themselves on social media. The campaign has raised over $100 million in the last 3 years, which is both incredible and undeniably life-changing for many. EMOTIONAL CONNECTION Making things personal for consumers can create connections that bridge the loyalty gap. Brands in the electronics, fashion and lifestyle markets are perceived as the most personal. Why? Because goods like these are the ones people spend the most time with every day. In fact, Harvard Business Review’s research shows that these emotionally connected customers can be 50% more valuable on average. When you think of electronics, it’s probably a little strange to think about touch screens, wires, chips and processors being deeply personal. But consider the number one thing that is with consumers every single day, arguably 24 hours a day: cell phones. They’re now an extension of who we are as people, even down to the brand that we use. Apple does a fantastic job appealing to emotion with their exciting pre-release keynotes and storytelling ads. The brand has managed to tap into the basic human needs of connection and belonging to something bigger than themselves. LOYALTY PROGRAMS We know, this is old news and it’s not just about getting customers to sign up for a card that gives them points. It’s about creating an easy, seamless experience that keeps customers coming back. A 2016 retail loyalty study found that customers who are members of loyalty programs, such as frequent flier clubs, generate between 12 to 18% more revenue than non-members. Think about Starbucks and their membership program. It comes at no cost to the customer, and rewards them for doing what they were already going to do: buy coffee. Starbucks capitalizes on this by running promotions to get you to buy more, like the Starbucks for Life instant-win game, and allows you to order through your phone to avoid standing in long lines. Integration with Alexa has taken things even further, letting everyone’s favorite voice-activated companion order coffee on their behalf. Not only is Starbucks reaping the benefits of added sales, but they’re also collecting priceless data on their customers. Much like Amazon and their data collection, it will bring them one step closer to providing the right service at the right time for each individual customer. SEGMENTATION OF 1 The idea of Segmentation of 1 has been floating around for some time now, with many companies starting to answer the need. Essentially, customers no longer want to be treated like everyone else; they have individual needs and expect to be treated as such. This poses a massive problem for brands. How do they do this for every person? The answer comes down to data and the strategy behind using that data. Amazon, Spotify, Netflix, Starbucks and many more are already starting to experiment with personalizing their product experiences. Amazon is the clear leader of the pack here. The dashboard you see after singing into your Prime account is a prime example. Everything is tailored to the actions you take while on and off their site. This is further enhanced with Alexa. Now you don’t even have to be on their website for them to continue to learn more about you and make recommendations. Orson Welles is shaking somewhere in his boots right now, without question. It should come as no surprise that analysts are predicting that Amazon will be the first trillion-dollar company. They continuously look toward the future instead of focusing on the present. Investors might have relentless heartburn over this, but the company has proven time and time again that they get it. What do analysts see as the key to their continued success? Alexa, and the integration of AI into consumers’ everyday lives. If they can continue this momentum into the future, Alexa will be with us 24/7 just like our cell phones — ready at the command of a word to get us whatever we need. WHY SHOULD YOU CARE? Gaining new customers is great for an expanding business model, though you can’t forget about your current customers. Retaining your current customers is 25 times less expensive than obtaining new ones, and they don’t have to be two completely different initiatives. If that doesn’t convince you, customer experience studies have shown that repeat buyers spend 33% more than new ones, and just 20% of existing customers account for 80% of a company’s future profits. Offering the experiences noted in this post, as well as others for new and existing customers, will help you hit the ever-moving target of brand loyalty. Originally published at www.mindstreaminteractive.com.

Brand Loyalty is Deeper Than the Product

0

brand-loyalty-is-deeper-than-the-product-1912948147a3

2018-07-13

2018-07-13 17:34:04

https://medium.com/s/story/brand-loyalty-is-deeper-than-the-product-1912948147a3

false

1,157

The latest Customer Experience insights from Mindstream Interactive.

null

MindstreamInteractive

null

Mindstream Interactive

null

mindstream-interactive

CUSTOMER EXPERIENCE,DIGITAL MARKETING,CONTENT STRATEGY,COLUMBUS OHIO,WEBSITE DESIGN

MindstreamInt

Marketing

marketing

Marketing

170,910

Mindstream Interactive

Mindstream Interactive is a customer experience agency that balances the innovative with the tried-and-true.

d81403b7fe5b

Mindstream_Interactive

12

1

20,181,104

null

0

from statsmodels.tsa.stattools import grangercausalitytests grangercausalitytests(df, maxlag = 1, verbose=True) #where df dataframe with two columns contains time series data you wanna compare Output: Granger Causality number of lags (no zero) 1 ssr based F test: F=0.6931 , p=0.4061 , df_denom=196, df_num=1 ssr based chi2 test: chi2=0.7037 , p=0.4016 , df=1 likelihood ratio test: chi2=0.7024 , p=0.4020 , df=1 parameter F test: F=0.6931 , p=0.4061 , df_denom=196, df_num=1

4

null

2018-06-23

2018-06-23 18:32:13

2018-06-24

2018-06-24 17:54:27

6

false

en

2018-06-26

2018-06-26 19:54:03

11

19130f38e2e

4.070755

0

Hi! Just wanna share some of my thoughts about correlation and causation problem.

3

Correlation and causation problem Correlation coefficient is 0.666004 Hi! Just wanna share some of my thoughts about correlation and causation problem. I’m sure that you probably saw number of different articles that claims that you can follow some of the habits of successful people and then you can became successful too, something like “10 rules of successful people” or “15 habits of highly successful people” or “Bill Gates wakes up every morning at 4 am so you need to start follow his schedule right now to be successful”. Of course that articles suffers from survivor’s bias but additionally after reading this articles you may conclude that this actions (or habits) was cause of the success that they achieved! Unlike correlation, causation is much harder to determine due to incomplete information we often have. For example we may observe that after event A, event B happened. So our brain (not just ours but other animals too) can conclude that A is cause of B. But at the same time there are might be several different hidden reasons that actually causes event B. There is famous experiment with “Skinner box”, where pigeons were trying to find pattern in a random actions. So main rule to remember: correlation does not imply causation In real world event almost always can be caused by multiple factors. We can split those factors into two groups: Necessary - If B cannot be caused unless A is present, then A is a necessary cause of B Sufficient - if the presence of A implies the occurrence of B, then A is a sufficient cause. In Data Science it’s very useful to keep this concept in mind, before running into premature conclusions. There is funny website with ridiculous correlations which is bright illustration that correlation != causation. Do we have statistical tests for that? Despite all of that difficulties with causation, can we actually analyse it? There are several statistical test that can help you with that. 1. Granger causality test For the time series data we can try Granger causality test, which is can help us determine whether one time series is useful in forecasting another. Example of Time Series A following time series B with time lag In order to apply this test two principles must be respected: The cause happens prior to its effect The cause has unique information about the future values of its effect Main problem is that works in specific cases only when A Granger-causes B. Yes, and don’t forget about error “Post hoc ergo propter hoc” Formula proposed by Clive Granger looks like this: where P — refers to probability, A is an arbitrary non-empty set, I(t) and I-x(t) respectively denote the information available as of time t in the entire universe, and that in the modified universe in which X is excluded. If the above hypothesis is accepted, we say that X Granger-causes Y. You can read more about that method in initial paper In python this statistical test can be run with 2. ATT and ATE For example you want to test whether or not new drug helping to cure some disease. So we need to conclude which effect was caused by new drug. ATE -average treatment effect: where Y1- is a response of people who took new drug, Y0- is a response of control group ATE is the average (over the whole population) of the individual level causal effects δ ATT- average treatment effect on the treated, ATT is average of the individual level causal effects for the observations that got treated, and is useful to explicitly measure the effect on those observations for which the treatment was intended. Or in case of A/B testing for example for some commercials, the ATE would the average difference in customer spend between customers that did and customers that did not participate in the promotion. The ATT would be the average effect only for customers that participated. Main problem with that, for a customer that participated in the promotion, we can never simultaneously measure the response in case they did participate, and the response in case they did not participate. You can read more about in this paper Main thing to remember: Even that we have some methods for causation analysis, non of them can provide explicit proof of causation. The best way to determine causality effect is experiment. But even with perfect experimental design we cannot prove causation, but can collect weaker or stronger evidence of causality. Conclusion: Correlation does not imply cauasution Keep in mind two groups of causation factors We have some statistical test but they may not worked Best way to determine causation is experimental prove, but it may not work as well References: Aldrich, John (1995). “Correlations Genuine and Spurious in Pearson and Yule” Holland, Paul W. (1986). “Statistics and Causal Inference” Granger, C. W. J. (1969). “Investigating Causal Relations by Econometric Models and Cross-spectral Methods”. Econometrica. Daniel E Ho, Kosuke Imai, Gary King, and Elizabeth A Stuart. 2011. MatchIt: Nonparametric Preprocessing for Parametric Causal Inference, Journal of Statistical Software, 8, 42.

Correlation and causation problem

0

correlation-and-causation-problem-19130f38e2e

2018-06-29

2018-06-29 13:19:35

https://medium.com/s/story/correlation-and-causation-problem-19130f38e2e

false

827

null

Data Science

data-science

Data Science

33,617

Alexander Osipenko

Data Scientist passioned in Deep Learning and Time Series analysis

fe96afc141b0

subpath

111

3

20,181,104

null

0

null

0

634d4b270054

2018-06-25

2018-06-25 05:06:35

2018-06-25

2018-06-25 05:07:47

1

false

en

2018-06-25

2018-06-25 05:07:47

3

1913cd4bdc33

1.14717

0

Navatics Mito, the underwater drone with a 4K camera opens up ocean exploration and helps create amazing footage. The Mito uses a floating…

5

Navatics Mito Drone Captures Amazing Underwater Footage Navatics Mito, the underwater drone with a 4K camera opens up ocean exploration and helps create amazing footage. The Mito uses a floating teether to transmit signals to the controller, but the teether also uses a built-in solar panel to extend the battery life of the drone from 2 hours to 4 hours run time. This drone uses four thrusters to move underwater and can also tilt up or down up to 45 degrees without changing the depth, and this drone can travel about 6.5 feet per second. “With Navatics Mito, our goal is to create an underwater drone with the ability to stream high-quality video and maneuver with maximum stability to get the best footage,” said Navatics CEO, Andreas Widy. “We want to provide this professional level experience so that anyone from the hobbyist consumer to the professional videographer can experience the marine world without limitation,” added Andreas. Source: https://bit.ly/2IqrN6T About DEEPAERO DEEP AERO is a global leader in drone technology innovation. At DEEP AERO, we are building an autonomous drone economy powered by AI & Blockchain. DEEP AERO’s DRONE-UTM is an AI-driven, autonomous, self-governing, intelligent drone/unmanned aircraft system (UAS) traffic management (UTM) platform on the Blockchain. DEEP AERO’s DRONE-MP is a decentralized marketplace. It will be one stop shop for all products and services for drones. These platforms will be the foundation of the drone economy and will be powered by the DEEP AERO (DRONE) token.

Navatics Mito Drone Captures Amazing Underwater Footage

0

navatics-mito-drone-captures-amazing-underwater-footage-1913cd4bdc33

2018-06-25

2018-06-25 05:07:48

https://medium.com/s/story/navatics-mito-drone-captures-amazing-underwater-footage-1913cd4bdc33

false

251

AI Driven Drone Economy on the Blockchain

null

DeepAeroDrones

null

DEEPAERODRONES

null

deepaerodrones

DEEPAERO,AI,BLOCKCHAIN,DRONE,ICO

DeepAeroDrones

Deepaero

deepaeros

Deepaero

0

DEEP AERO DRONES

null

dcef5da6c7fa

deepaerodrones

277

0

20,181,104

null

0

null

0

null

2017-12-14

2017-12-14 02:09:11

2017-12-14

2017-12-14 02:14:03

1

false

en

2017-12-14

2017-12-14 02:14:03

2

191444550645

3.169811

2

0

Whether you are thinking about a career change or getting ready for a deep learning project, it’s wise to understand the types of data…

4

3 Data Science Roles and How to Choose the Right One for You Whether you are thinking about a career change or getting ready for a deep learning project, it’s wise to understand the types of data science work you are most interested in. Three Types of Data Scientist Roles Most data science and deep learning jobs can be placed in one of three categories. Which one are you, and how should you pursue it? Research Data Scientist Product Data Scientist Data Engineer The Research Data Scientist The Research Data Scientist pushes the boundary of what is possible with creativity, maths and experimentation. Understanding and advancing theory is a must. You need to be comfortable doing what’s never been done before. This professional is mathematical and probably has a higher degree in maths, computer science or machine learning. The Research Data Scientist creates or rewrites algorithms, improves their performance, applicability and range of action. That’s how knowledge is pushed forward. A warning: the jobs for this role are more rare and more competitive than jobs for the other roles. Research Data Scientists might find jobs with Google Brain, Deep Mind, Twitter Cortex team, Facebook AI Research (FAIR) team, Tesla autopilot team, OpenAI and similar research labs. The Product Data Scientist The Product Data Scientist builds models to understand different aspects of how the users interact with the product to inform decision makers such as the executive team, product managers and engineering managers. She may also build models to extract insights from production and internal processes, for example to predict movements in the supply chain or operations. She is able to answer questions like: How are users interacting with the product? What changes can be made for more sales or better safety or faster service? Often, this role involves running experiments, like A/B tests or multi armed bandits in order to determine which feature is better, more engaging or more efficient. The Product Data Scientist works to continuously improve the product using data. In addition, many products today include intelligent components as features. Think for example of intelligent notifications and content recommendation in social media apps. In such cases, the Product Data Scientist drectly contributes to product development. Product Data Scientists are found in all major tech companies like Google, Apple, Microsoft, Airbnb, Optimizely, Uber, Lyft and so on. The Data Engineer It’s almost impossible to handle massive amounts of data and users without a sound infrastructure. More often than not, companies looking for “Data Scientists” are really looking for a data engineer. The Data Engineer designs and builds the architecture for systems that grow and scale. This professional is a developer at heart — one who can think ahead and prepare for pitfalls. In this role, you will prevent and resolve crashes, over-allocated resources, categorize messy data sets, and generally create order out of chaos. Their interest lies downstream of the model training phase. They get excited by the aspect of real-time interaction and serving of predictions to millions of users. You are ready for this role when you can design data architectures. Data Engineers might find jobs with the same groups mentionesd above alongside Uber and Lyft, and Netflix. Explore, then decide. The profiles above are very different and it is important to understand which role is best fitting your skills, your mindset and what you love doing. If you are unsure of where your heart lies, it is never a bad thing to have a little experience in a lot of areas. Rather than picking one, you try them all and here are some ways to do that. Apply for a side gig Short-term side gigs, like those you might find on Upwork.com, are great ways to stretch into unfamiliar areas. Work on a personal project Is there a data problem you’ve always wanted to work on? Dedicating a few hours every week to a personal project could help you expand your toolset. Participate in a data science competition See if you like building models by taking part in a data science competition on Kaggle or at a Hackathon. It is a great way to make friends and test your skills and interests for this profession. Learn online Sites like Udemy, Udacity, and Coursera offer courses to learn at your own pace giving you the chance to upgrade your skills in your spare time. Take a short course or bootcamp The easiest and fastest way to dive into something is with targeted coaching and practice. A short, instructor-led course or bootcamp allows you to quickly dive into material with the benefit of personal assistance. When you’re ready to dive into Deep Learning, check out these courses by Catalit & Data Weekends.

3 Data Science Roles and How to Choose the Right One for You

2

3-data-science-roles-and-how-to-choose-the-right-one-for-you-191444550645

2018-06-12

2018-06-12 23:09:40

https://medium.com/s/story/3-data-science-roles-and-how-to-choose-the-right-one-for-you-191444550645

false

787

null

Data Science

data-science

Data Science

33,617

CATALIT

null

e9ded0b6ba4

catalit

0

1

20,181,104

null

0

null

0

null

2018-05-30

2018-05-30 08:31:30

2018-05-30

2018-05-30 10:58:48

3

false

en

2018-05-30

2018-05-30 12:56:36

0

1914f97f7602

3.534906

10

0

It’s the first week since the GDPR is now in force, and hopefully your inbox can relax after the stream of privacy updates!

5

GDPR BINGO It’s the first week since the GDPR is now in force, and hopefully your inbox can relax after the stream of privacy updates! If you are a citizen of the EU — do you feel different with your new found legal rights? I do. I recently made a career change to become a Data Scientist, and completed an educational program studying Machine Learning, Deep Learning and Artificial Intelligence. Among other things I learned techniques which can be used to extract emotional sentiment from raw text, and happened to be learning these methods at the time the Gaurdian interview of Cambridge Analytica whistleblower, the Christopher Wylie was released. “My name is Christopher Wylie, I am a Data Scientist.” “I made Steve Bannon’s psychological warfare tool” The scope and implications of his and Zuckerberg’s testimony (remember when the internet made funny pictures of Zuckerberg as an Android instead of discussing the issue?) and the overall impact is a complicated and deep subject. For now, let’s say that the timing coincided with my general feeling of unease regarding my personal privacy and how my information was being used for manipulation All of this was then amplified by learning of the unprecedented changes and techniques occurring in the field of Artificial Intelligence. Data Science papers are being published now, which make work last year obsolete. Fast adoption of new techniques and technologies and low barriers to deployment make the future of privacy more uncertain than it has been in my experience. Intent With this as a basis, I wanted to better understand the 99 articles of the GDPR and how they apply to me personally and professionally, and to get a feeling for how companies are handling the new regulatory environment. My intent is not to expose particular companies who are not in compliance with GDPR, as I believe most companies are unsure what compliance looks like, I am prepared to wait a while until we see some cases move through the legal system. I am not here to troll and overload small business owners with administration costs, and will not state any company names or communications. The game During high-school in Ontario, I volunteered several times at a Bingo hall, a smoky room where retirees can pass the afternoon in conversation over a 5 by 5 grid, hoping to connect 5 squares and yell BINGO! I thought of using this as the structure for my investigation. I have begun the process of submitting GDPR Subject Access Requests to 24 companies with whom I have an online account. These companies are randomly distributed over the BINGO card. A square is punched when the company complies with my request. Subject Access Requests In making my Subject Access Requests (SAR), I used the following overall structure for emails, forms, and letters; [DATA] What personal data is stored by the controller or processor? [PROCESSING] How is my data used to i.e. target ads? [SECURITY] Is the data stored securely? Were there any breaches? [THIRD PARTIES] Provide a list of 3rd parties Each SAR was tailored to each company and communication format. The game so far To date, I have contacted less than 10 companies, each with varying degrees in ease of contact. Here are my darts (bad!) and laurels (good!) in my experience thus far; Impossible to find a contact email, even after chatting with customer support :( Responding with the same link to an online portal, even though I am requesting more information :( Some clear non-compliance, one company failed to provide third party company information until I pressed them :( An email address readily provided to contact their Data Protection Officer (i.e. in the privacy email or policy) :) Response within 24 hours by personal email :) An online data portal, where you can browse your information :) Generally the challenge of the game right now is dealing with the big corporations. Some have a policy of not providing any direct email contact information, likely because responding to individual requests is expensive. In these instances, I will be sending letter to their corporate addresses, just like in the days of yore. Let’s have a discussion What crosses your comfort boundary in the new age of internet privacy? Will anyone care about GDPR in a few months when the hype dies down? Which types of companies will bust due to costs and liabilities of GDPR (not just claiming that’s the reason, but really making their business impossible)? When will the first big legal case reach the ECJ, and what will it look like? When will we have the first Deep Fake scandal? How do deep fakes impact the free press? Stay tuned for an update in a few weeks and hopefully a BINGO!

GDPR BINGO

91

gdpr-bingo-1914f97f7602

2018-06-18

2018-06-18 15:50:03

https://medium.com/s/story/gdpr-bingo-1914f97f7602

false

791

null

Privacy

privacy

Privacy

23,226

Marcus Jones

Marcus is a philosopher, engineer, and data scientist. Passionate about blockchain, AI, and robotics!

4b33182d7b97

jones.0bj3

20

12

20,181,104

null

0

null

0

null

2018-02-15

2018-02-15 21:15:31

2018-02-15

2018-02-15 21:23:54

1

false

en

2018-02-15

2018-02-15 21:35:40

0

1914fea520b6

2.298113

3

0

Gone are the days when Artificial Intelligence seemed to appear only in science fiction movies and seemed to be nothing more than…

2

The impact of Artificial Intelligence Gone are the days when Artificial Intelligence seemed to appear only in science fiction movies and seemed to be nothing more than imagination. With the advancement in technology, today, we are making great progress towards making AI a big reality. Machine learning, a field under artificial intelligence has made tremendous efforts in enabling machines to learn from data using various neurological algorithms inspired from human brains. I personally believe that AI hasn’t have had great Impact on people until now but I believe that in coming future it will probably transform the way we humans perform most of our jobs. Let us take a look at those areas which AI is expected to impact the most. Transportation is one of those areas which I expect will undergo a major transformation. Currently self-driving cars are not completely automated and are dependent on drivers for safety purposes. Looking at the pace of development of automation of cars, I believe that very soon we will see automated cars on our streets. Since these automated cars are being highly trained for all possible scenarios with higher accuracy, I believe that they will relatively be a little more skilled than humans Replacing Dangerous Jobs like bomb defusing, manufacturing toxic substance is expected to be replaced by robots. There are few robots that are already performing these jobs in USA. Hence, if robots take up these kinds of dangerous and hazardous jobs with a lot more accuracy, than we humans can lead a better and a safe life. Friends are one of those groups of people who play a very important role in our life. Unfortunately, not all of us have friends, with whom; we can share our feelings, problems, emotions. Using AI, artificial friends have also been developed though not most of them have any sort of emotions. There are few robots in China which do have emotional intelligence and can understand a human’s emotions and feelings. This way technology can be used to help those people who don’t have a great friend circle and those who are basically introverts. This is one of my favorite products of AI. One of the areas that is expected to be hit hard due to transformations that AI will make, is Employment sector. I believe that the landscape of jobs will undergo a major transformation. Quality of jobs will be changed. Smart and intelligent people who would have adapted to these changes will be pushed towards higher income and the rest will be pushed towards lower income. But it is important to realize that if one part of job is eaten away by machines than we do have another part of the job that machines are not good at and this part of the job is meant for humans. So if we adapt to the changes, employment for humans will always be there. AI has the potential to transform the world and if it does, it will automate most of the jobs that we perform and will impact our everyday lives. So to conclude, I would like to put forward the opinion that, if we adapt and accept the changes and change ourselves for the change, than, there are high chances that this technology that is growing at fast rate will probably have more positive impact than negative.

The impact of Artificial Intelligence

3

the-impact-of-artificial-intelligence-1914fea520b6

2018-02-16

2018-02-16 07:51:19

https://medium.com/s/story/the-impact-of-artificial-intelligence-1914fea520b6

false

556

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Honey Khandelwal

null

e1ccee5f74cf

honeykhandelwal

9

13

20,181,104

null

0

null

0

ec73089917ed

2017-11-18

2017-11-18 13:34:06

2017-11-18

2017-11-18 15:47:12

1

false

en

2017-11-18

2017-11-18 16:29:02

9

19155b93c870

5.660377

3

0

In this post I want to discuss a question of (to me at least) paramount importance — data quality. New machine learning algorithms, their…

4

It all Starts with Data Quality In this post I want to discuss a question of (to me at least) paramount importance — data quality. New machine learning algorithms, their evaluation and optimisation are of course useful and great fun, but if you apply a great algorithm to noisy, dirty or corrupt data, don’t expect it to give great results back. Two episodes from brilliant poscasts discuss this issue among other things, like public awareness of machine learning, how to be a good CDO (chief data officer) and even how gradient descent works. One of the episodes came out already a month ago, but I was waiting for a worthy companion. First, Talking Machines, and I quote them, “talk with Peter Donnelly in his capacity with the Royal Society’s Machine Learning Working Group about the work they’ve done on the public’s views on AI and ML.” Second, in a more recent episode, IBM Big Data & Analytics Hub, talk about ML in production environment, data scaling and governance, data science teams management and, yes, data quality and data corruption (starting around min 20:40). Poor quality data comes in different flavours, none of which taste well. Essentially you want your dataset to represent the phemonenon you want to model as closely as possible. So, what can go wrong? Several things: You sources of data were not built to capture the phenomenon — e.g., you want to model weather patterns (a very ambitious project!) but you use stock exchange as your main input. That you should not do this sounds obvious, right? But you will be surprised what data gets fed to the models sometimes, often due to an honest mistake or based on erroneus assumtions. Example from my own experience: Once I was analysing motion trajectories of left and right wrists from motion capture data. And the first results looked… weird. Turns out in my code I have given a wrong index and was instead analysing motion patterns of left and right knees! Your sources of data can capture the phenomenon in theory but are not placed in the right environment to do so — e.g., you want to use outside temperature records to model the weather but your thermometer was placed indoors all this time. Now, if this sensor was placed in a poorly insulated house (don’t even get me started on poorly insulated houses), your input data will be still capturing weather patters to some extent, but you have to agree it would be much better to record from a sensor outside. Example from my own experience: About a year ago I had access to a bunch of public instagram images and their tags. I was building a deep learning model that would recognise food categories from an image and hoped the tags would give me a nice shortcut to obtained a labeled dataset. After all, if a tag says “#pizza”, there should be pizza in the image, right? Wrong. The tags were all over the place because users don’t use tags to just describe what’s on the image. Your sensor is biased. This is a tricky one. Biases, in humans, are almost inevitable, because they help us to take shortcuts to decisions and save time. They can be helpful in many situations but in many others they are undoubtedly harmful, especially when we are dealing with social situations and decisions that can influence people’s lives. Example from my own experience: I wanted to detect people in Twitter profile pictures. After running a few hndred of them though a model I found online I started to evaluate the results. There were not many false positives — when there was no human on the picutre, the model was saying “no human”. But there were tons on fasle negatives on… Afro-american profiles! That is, there was a person in the picture, clearly, but the model was not detecting them! The researcher who built the model were not evil racists of course, they simply used a bad dataset to train the model, the one that contained mostly white faces in it. Needless to say, I ended up not using the model. So let’s consider this third case in a bit more detail. Let’s say you want to help your HR department to save time and are building a model that will evaluate CVs of applicants and produce two suggestions — invite for an interview or not. You have a dataset with the CVs of previous applicants and whether or not they were invited for an interview. Now let’s pretend we have some kind of amazing CV format that captures very well all the skills and experiences of the applicants (which is not always the case with CVs), so that the first two scenarios are not relevant here. But what you don’t know is that previous applicants’ successes were evaluated by people and one or more of those people had very strong prefences for one or another aspect or skill that actually does not reflect that applicant’s suitability for the role they are applying for. The most glaring candidates for such biases are of course gender, race, age (regardless whether they are stated directly or expressed in hidden ways such as place of birth, first employment record, name, even postcde), but there can be more subtle situations. Say, the person evaluating a CV has very fond memories of a univerisy they graduated from and when they see that an applicant has graduated from the same institution they immediately think that the applicant is a lovely person and is a better fit for the job than another applicant with an identical record but for the university they had graduated from. First of all, please don’t assume I think that such CV evaluators are evil or stupid — no one is free of biases, all we can do is become aware of them and actively work to correct our thought processes. I am by far not the only one to point the importance of data quality and the problem of bias in data. If you want to learn more about this phenomenon, there are many voices you can lend your ear to: Weapons of Math Destruction book ProPublica’s series on Machine Bias John Giannandrea, who leads AI at Google … and many many more What I often ask myself is — why to ML models tend to pick up human biases present in data so easily, especially sexism and racism? So far the only answer I have is that such biases are exactly what they are — superficial but salient. The properties they are based on are very easy to spot — one’s skin color, gender-determined facial features, age, but they don’t correlate with properties what should really matter in many social contexts, unlike such things as resilience, honesty, trustworthiness, etc. There are many studies that claim the latter group can be captured from ones photograph or voice but the accuracy is rather low and again, the models are based on data coming from human perception and human perception is biased. E.g., we tend to think beautiful (a social construct in many ways) people are kind, smart and trustworthy. But are they? And if indeed yes, then why? (Actually, the scientific studies themselves would report on, e.g., “perceived trustworthiness” and later newpapers will over-simplify and distort such reports claiming outraegously that good looks in an interlocutor guarantee you can trust them.) So, what can you do if you have a dataset on your hands and you want to check it for hidden bias? Alas, there is no silver bullet, and the more complex your data the harder it is to audit it. But you can certainly at least have a decent go at it — plot and analyse your data! Visualise it in such ways that would allow you to see whether certain features strongly correlate to your classes and then ask yourself whether they actually should. If you are doing clustering look at the resulting items that end up in the same cluster and see which feaures seem to unite them and whether, again, they should in fact, be the uniting features. If you have features that can promote biases and should not be relevant to the outcome — remove them, but be aware than some seemingly innocent features like postcode can drive biased outcomes. (Of course, sometimes even gender is a perfectly valid feature, for instance if we are dealing with prostate cancer diagnosis, but that’s kind of obvious.) Even if your model seems to be performing very well, it is your responsibility to interrogate it and make sure it is not reinforcing human biases. Image Source

It all Starts with Data Quality

13

it-all-starts-with-data-quality-19155b93c870

2018-03-03

2018-03-03 13:24:52

https://medium.com/s/story/it-all-starts-with-data-quality-19155b93c870

false

1,447

KPD stands for “Kati’s Podcast Digest” and captures the purpose of this publication with 100% accuracy — I’m subscribed many dozens of podcasts on {data, neuro-, popular} science and some episodes are just too good to stay unshared and undiscussed.

null

100 percent KPD

null

100-percent-kpd

PODCAST,SCIENCE,DATA SCIENCE,DIGEST

null

Machine Learning

machine-learning

Machine Learning

51,320

Kati Volk

Data Scientist in Switzerland. All opinions expressed in my posts are my own.

e70f76bf9994

kati_volk

16

20

20,181,104

null

0

null

0

9ff8c2a594cf

2018-03-25

2018-03-25 03:45:14

2018-03-25

2018-03-25 04:37:01

1

false

zh-Hant

2018-03-25

2018-03-25 05:03:11

3

1915b2a4b41f

0.283019

5

0

連續兩周緊鑼密鼓地參加創業發表會、創業須知課程、大師講座，投資與財務課程，這些種種因素交織在我這兩週"進化爆炸"的腦袋裡，記錄下來除了再次整理、產出提醒自己2018年曾有過這些觀察與思維以外，也是幫助自己日後回顧並且記錄自己思維進化的證據。

5

【將軍趕路，莫追小兔】 Photo by Dana Critchlow on Unsplash 連續兩周緊鑼密鼓地參加創業發表會、創業須知課程、大師講座，投資與財務課程，這些種種因素交織在我這兩週"進化爆炸"的腦袋裡，記錄下來除了再次整理、產出提醒自己2018年曾有過這些觀察與思維以外，也是幫助自己日後回顧並且記錄自己思維進化的證據。綜合現場感覺氛圍，及觀察諸位創業者的思考邏輯與我自身的心得: 台灣創業家們不缺好產品與好人才，缺的是好思維與商業模式這點我認為跟台灣曾經身為全球代工產業鏈的角色定位有很大的關係，你身處的角色定位會建構出你所處的思維邏輯，思維邏輯則會再"限制"且呈現你所認知的世界，進而強化你對於自身角色的定位認同。這次一個雞生蛋的循環內外整合。我發現參賽團隊各個背景厲害，有海歸派、有從優秀大學出來、理想性很高的學生，聚集優秀人才是很好，但是他們都是一群能力相近的優秀者，譬如設計師與設計師們。看不到團隊成員裡有互相互補且共同成長得能力。相同性質的人建構企業不論是商業模式、初衷皆都有盲點，會呈現出一種"自high"的一廂情願(有共同目標與精神是很好，但是因為每人皆只做自己擅長的事(例如設計師們共同創業只做設計與銷售，最重要的財務方面反而因為沒有人願意去規劃而荒廢) “創業者只找同溫層”，卻沒有去多方面拓展自己團隊的能力，以不同專業領域人的角度與眼光來思辨與省視如何讓自己企業進化的更好，如何強力存活的現象。而且很可怕的是，幾乎現場很少有人聽得懂評審對於自身的疑問，回答時會陷入一種連外人都聽得出來在強辯的氛圍(評審的提問我認為是有指標性的，這代表一種市場上的疑問，在進入市場燒錢以前發掘到問題都是好的) 2. 務必蓋好財務結構模式，這是你創業夢想的基礎回到前篇來看，務必替自己尋找一位懂財務專業的人來幫自身的企業做規劃與隨時檢視(自己來做會更好)，我推薦此書，看完之後會破除很多以前"自以為"的商業觀念，且對於提升自身、不要再陷入代工思維的經濟模式會有很多突破盲點的幫助: 數字力就是賺錢力：看不懂財報沒關係，只要牢記宅經濟必學的二個獲利數據，創業、開店、提升業績，完全搞定數字力就是賺錢力：看不懂財報沒關係，只要牢記宅經濟必學的二個獲利數據，創業、開店、提升業績，完全搞定書名：數字力就是賺錢力：看不懂財報沒關係，只要牢記宅經濟必學的二個獲利數據，創業、開店、提升業績，完全搞定，原文名稱：「数字」が読めると本当に儲かるんですか？，語言：繁體中文，ISBN：9789573281825，頁數：224，出版社：遠流…www.books.com.tw 3.投資跟創業最好一起學，省時間 2017下半年時我開始學投資，那時還沒有創業概念，但我發現一件事 "看懂財務報表這件事就大略看得懂別人家企業在做蝦米!"，於是我反推思考，如果學會看懂財務報表的種種要素與數字之間的聯帶關係，學會看得懂別人建構優良企業的數字根據的邏輯是哪些，是否代表我也可以以此邏輯去建構出讓別人(投資者)看得懂並且也想投資的"自身企業" 這是一種"數學基礎好"，你就有基礎可以往化學或物理去選擇你想要的道路"，因為數學是科學之母，學好這個觀念就可以往自己想要的路上更堅實的邁進。自從出社會工作以來，除了在大公司工作被束縛的感覺之外，我就一直有一種隱性感覺，台灣想當老闆的人很多，但是有能力當得起老闆的人很少(這點身為員工就可以清楚地感覺得到，從老闆吩咐你做事到老闆叫你採買東西的邏輯就可一窺略知) 創業是個需要持續自身偏執的，這是一個沒有公式的遊戲，以自己的方式去建構出讓孩子生成他未來生存所必需的器官與內臟循環系統各位創業者，咱們共勉之~

【將軍趕路，莫追小兔】

29

將軍趕路-莫追小兔-1915b2a4b41f

2018-03-30

2018-03-30 11:18:33

https://medium.com/s/story/將軍趕路-莫追小兔-1915b2a4b41f

false

22

那些你在閱讀/創業/戀愛/所學到的事 What I learned from those things

null

katherlin3

null

《情商小姐Lady Emotions的時間銀行》

k90894@gmail.com

壞捲毛凱特的-時間銀行

TIME,ROMANCE,MANAGEMENT,AWARENESS,INSIGHTS

null

Creators

creators

Creators

799

Lin ChiaYi

#Social Enterprise Investor明明出身廣告行銷科系、但時常被老闆叫去做【營運管理】的人，想不到因此學會用不同的角度看待事情。關注資源整合，集中力量做聰明懶的事。因為口誅會與人吵不喜歡的架，所以選擇邏輯性地筆伐。身體教育工作者，關注【女性角色、自我身體教育覺知、情緒覺察】

77618e8c3099

ladyemotions

240

155

20,181,104

null

0

null

0

null

2018-02-06

2018-02-06 20:44:18

2018-02-06

2018-02-06 21:09:15

1

false

es

2018-02-07

2018-02-07 17:30:01

2

1915c22bc2cc

1.332075

1

0

Dentro del laboratorio del Dr. Frankestein, no trabajará solamente su ayudante Igor para desarrollar el “diabólico plan”; ahora también…

5

EL CINE Y LA INTELIGENCIA ARTIFICIAL CREAN UN NUEVO FRANKESTEIN foto de The Frankenstein AI director Lance Weiler Dentro del laboratorio del Dr. Frankestein, no trabajará solamente su ayudante Igor para desarrollar el “diabólico plan”; ahora también contará con nuevos colaboradores contemporáneos: los espectadores participarán en la construcción de la escena. Es que con la incursión de la inteligencia artificial (IA) y machine learning en el cine; la audiencia puede transmitirle a la película sus miedos, horror, sentimientos y respuestas a preguntas sensibles -difícilmente olvidables-. Los nuevos directores ofrecen participación e interacción directa al público para el adn del futuro “monstruo”. En Sundance 2018 se presentó el experimento The Frankenstein AI http://frankenstein.ai/ del director Lance Weiler, y apoyado por la Columbia University School of the Arts’ Digital Storytelling Lab (DSL) http://www.digitalstorytellinglab.com/; la nueva narrativa que dura unos 45 minutos, invita a los espectadores a una habitación oscura donde interactúa con el “amigo monstruo”, generándose una serie de preguntas y respuestas íntimas existenciales, comienza así un nuevo recorrido de comunicación instantánea entre los humanos y esta reciente invención virtual, el protagonista querrá aprender de los humanos y viceversa; al final del experimento, es creado un nuevo Frankestein producto de un algoritmo producido entre la máquina, y las respuestas y sentimientos compartidas de la atrevida audiencia activa. Si bien la inteligencia artificial ya había sido utilizada en películas de Steven Spielberg y otras, para estos desarrollos eran necesario presupuestos millonarios; a partir de ahora estas tecnologías son posibles de adquirir a muy bajos costos, lo que naturalmente podría transformarse en un incipiente arte de producción masiva para los directores de proyectos cinematográficos. Quizás sea el futuro de un nuevo paradigma del cine y formas de contar historias, que podrá depender de un alto nivel de creatividad y avanzados algoritmos humanizados. Nosotros los espectadores activos, expectantes de lo que sucederá.

EL CINE Y LA INTELIGENCIA ARTIFICIAL CREAN UN NUEVO FRANKESTEIN

10

el-cine-y-la-inteligencia-artificial-crean-un-nuevo-frankestein-1915c22bc2cc

2018-05-11

2018-05-11 06:13:47

https://medium.com/s/story/el-cine-y-la-inteligencia-artificial-crean-un-nuevo-frankestein-1915c22bc2cc

false

300

null

Frankenstein

frankenstein

Frankenstein

220

Mauricio Viola

Especialista en búsquedas de información en internet y buscadores. busquedas.net / Co-Fundador Search Engine Hacks

a889677cca8c

busquedasnet

62

274

20,181,104

null

0

null

0

634d4b270054

2018-04-13

2018-04-13 06:21:47

2018-04-13

2018-04-13 06:23:07

1

false

en

2018-06-05

2018-06-05 08:51:23

3

1917c305c39a

1.256604

0

The evolution of drones has helped professionals, industrialists, scientists and researchers, to capture, survey, analyze and map a variety…

5

‘Drone Or A Manned Aircraft’ What Should You Opt For? The evolution of drones has helped professionals, industrialists, scientists and researchers, to capture, survey, analyze and map a variety of spaces and places. Talking about the drones and the manned aircraft, the basic function revolves around which aircraft is to be used for the specific part of the project. Unmanned aerial vehicles (UAV’s) have been referred to a disruptive technology because of their potential to replace expensive and time-consuming tasks. Drones are the best alternatives when it comes to capture information from large areas of land. Where to use manned aircraft and drone? To cover a large areas quickly, or when it’s required by the RFP (requests for proposals) opt for manned aircraft. “Right now in the industry, there are still a lot of requirements that call for flying an entire country or state,” said Matt Coleman, VP of Solutions Engineering for PrecisionHawk. “Drones come in where there are small areas; areas in remote locations that are hard to fly. You can get higher resolution with drones, as compared to manned aircraft surveying,” said Matt Coleman. Source: https://bit.ly/2qsdzMN About DEEPAERO DEEP AERO is a global leader in drone technology innovation. At DEEP AERO, we are building an autonomous drone economy powered by AI & Blockchain. DEEP AERO’s DRONE-UTM is an AI-driven, autonomous, self-governing, intelligent drone/unmanned aircraft system (UAS) traffic management (UTM) platform on the Blockchain. DEEP AERO’s DRONE-MP is a decentralized marketplace. It will be one stop shop for all products and services for drones. These platforms will be the foundation of the drone economy and will be powered by the DEEP AERO (DRONE) token.

‘Drone Or A Manned Aircraft’ What Should You Opt For?

0

drone-or-a-manned-aircraft-what-should-you-opt-for-1917c305c39a

2018-06-05

2018-06-05 08:51:24

https://medium.com/s/story/drone-or-a-manned-aircraft-what-should-you-opt-for-1917c305c39a

false

280

AI Driven Drone Economy on the Blockchain

null

DeepAeroDrones

null

DEEPAERODRONES

null

deepaerodrones

DEEPAERO,AI,BLOCKCHAIN,DRONE,ICO

DeepAeroDrones

Deepaero

deepaeros

Deepaero

0

DEEP AERO DRONES

null

dcef5da6c7fa

deepaerodrones

277

0

20,181,104

null

0

import os import json import requests BLS_API_KEY = os.environ.get('BLS_API_KEY') BLS_ENDPOINT = "http://api.bls.gov/publicAPI/v2/timeseries/data/" def fetch_bls_series(series, **kwargs): """ Pass in a list of BLS timeseries to fetch data and return the series in JSON format. Arguments can also be provided as kwargs: - startyear (4 digit year) - endyear (4 digit year) - catalog (True or False) - calculations (True or False) - annualaverage (True or False) - registrationKey (api key from BLS website) If the registrationKey is not passed in, this function will use the BLS_API_KEY fetched from the environment. """ if len(series) < 1 or len(series) > 25: raise ValueError("Must pass in between 1 and 25 series ids") # Create headers and payload post data headers = {'Content-Type': 'application/json'} payload = { 'seriesid': series, 'registrationKey': BLS_API_KEY, } # Update the payload with the keyword arguments and convert to JSON payload.update(kwargs) payload = json.dumps(payload) # Fetch the response from the BLS API response = requests.post(BLS_ENDPOINT, data=payload, headers=headers) response.raise_for_status() # Parse the JSON result result = response.json() if result['status'] != 'REQUEST_SUCCEEDED': raise Exception(result['message'][0]) return result $ export BLS_API_KEY=yourapikey >>> series = ['LNS14000000', 'LNU05026645', 'LNS12032194', 'LNS14027689'] >>> data = fetch_bls_series(series, startyear=2000, endyear=2015) >>> print(json.dumps(data, indent=2)) { "Results": { "series": [ { "seriesID": "LNS14027689", "data": [ { "year": "2009", "period": "M12", "periodName": "December", "footnotes": [ {} ], "value": "8.7" }, { "year": "2009", "period": "M11", "periodName": "November", "footnotes": [ {} ], "value": "8.8" } [… snip …] ]}]}} import pandas as pd info = pd.read_csv('../data/bls/series.csv') info.head() def series_info(blsid, info=info): return info[info.blsid == blsid] # Use this function to lookup specific BLS series info. series_info("LAUST280000000000003") info[info.source == 'LAUS'] info[info.blsid.apply(lambda r: r.startswith('LNS14'))] import csv from itertools import groupby from operator import itemgetter # Load each series, grouping by BLS ID def load_series_records(path='../data/bls/records.csv'): with open(path, 'r') as f: reader = csv.DictReader(f) for blsid, rows in groupby(reader, itemgetter('blsid')): # Read all the data from the file and sort rows = list(rows) rows.sort(key=itemgetter('period')) # Extract specific data from each row, namely: # The period at the month granularity # The value as a float periods = [pd.Period(row['period']).asfreq('M') for row in rows] values = [float(row['value']) for row in rows] yield pd.Series(values, index=periods, name=blsid) series = pd.concat(list(load_series_records()), axis=1) def plot_single_series(blsid, series=series, info=info): title = info.get_value(info[info.blsid == blsid].title.index[0], 'title') series[blsid].plot(title=title) >>> plot_single_series("LNS12300000") def plot_multiple_series(blsids, series=series, info=info): for blsid in blsids: title = info.get_value(info[info.blsid == blsid].title.index[0], 'title') series[blsid].plot(label=title) plt.title("BLS Time Series: {}".format(", ".join(blsids))) plt.legend(loc='best') >>> plot_multiple_series(["LNS14000025", "LNS14000026"])

35

f4f1e49a4f74

2017-12-31

2017-12-31 17:16:57

2017-12-31

2017-12-31 17:19:42

6

false

en

2017-12-31

2017-12-31 17:19:42

15

1917d79bb01f

11.24434

1

0

By Benjamin Bengfort

4

Exploring Bureau of Labor Statistics Time Series By Benjamin Bengfort Machine learning models benefit from an increased number of features — “more data beats better algorithms”. In the financial and social domains, macroeconomic indicators are routinely added to models particularly those that contain a discrete time or date. For example, loan or credit analyses that predict the likelihood of default can benefit from unemployment indicators or a model that attempts to quantify pay gaps between genders can benefit from demographic employment statistics. The Bureau of Labor Statistics (BLS) collects information related to the labor market, working conditions, and prices into periodic time series data. Moreover, BLS provides a public API making it very easy to ingest essential economic information into a variety of analytics. However, while they provide raw data and even a few reports that analyze employment conditions in the United States, the tables they provide are more suited towards specialists and the information can be difficult to interpret at a glance. In this post, we will review simple data ingestion of BLS time series data, enabling routine collection of data on a periodic basis so that local models are as up to date as possible. We will then visualize the time series using pandas and matplotlib to explore the series provided with a functional methodology. At the end of this post, you will have a mechanism to fetch data from BLS and quickly view and explore data using the BLS series id key. The BLS API The BLS API currently has two versions, but it is strongly encouraged to use the V2 API, which requires registration. Once you register, you will receive an API Key that will authorize your requests, ensuring that you get access to as many data sets as possible at as high a frequency as possible. The API is organized to return data based on the BLS series id, a string that represents the survey type and encodes which version or facet of the data is being represented. To find series ids, I recommend going to the data tools section of the BLS website and clicking on the “top picks” button next to the survey you’re interested in, the series id is provided after the series title. For example, the Current Population Survey (CPS), which provides employment statistics for the United States, lists their series; here are a few examples: Unemployment Rate — LNS14000000 Discouraged Workers — LNU05026645 Persons At Work Part Time for Economic Reasons — LNS12032194 Unemployment Rate — 25 Years & Over, Some College or Associate Degree — LNS14027689 The series id, in this case, starts with LNS or LNU: LNS14000000, LNU05026645, LNS12032194, and LNS14027689. There are two methods to fetch data from the API. You can GET data from a single series endpoint, or you can POST a list of up to 25 ids to fetch multiple time series at a time. Generally, BLS data sets are fetched in groups, so we'll look at the multiple time series ingestion method. Using the requests.py module, we can write a function that returns a JSON data set for a list of series ids: This script looks up your API key from the environment, a best practice for handling keys which should not be committed to GitHub or otherwise saved in a place that they can be discovered publicly. You can either change the line to hard code your API key as a string, or you can export the variable in your terminal as follows: The function accepts a list of series ids and a set of generic keyword arguments which are stored as a dictionary in the kwargs variable. The first step of the function is to ensure that we have between 1 and 25 series passed in (otherwise an error will occur). If so, we create our request headers to pass and receive JSON data as well as construct a payload with our request parameters. The payload is constructed with the keyword arguments as well as the registration key from the environment and the list of series ids. Finally, we POST the request, check to make sure it returned successfully, and return the parsed JSON data. To run this function for the series we listed before: You should see something similar to the following result: From here it is a simple matter to operationalize the routine (monthly) ingestion of new data. One method is to store the data in a relational database like PostgreSQL or SQLite so that complex queries can be run across series. As an example of database ingestion and wrangling, see the github.com/bbengfort/jobs-report repository. This project was a web/D3 visualization of the BLS time series data, but it utilized a routine ingestion mechanism as described in the README of the ingestion module. To simplify data access, we’ll use a database dump from that project in the next section, but you can also use the data downloaded as JSON from the API if you wish. Loading Data into Pandas Series For this section, we have created a database of BLS data (using the API) that has two tables: a seriestable that has information describing each time series and a records table where each row is essentially a tuple of (blsid, period, value) records. This allows us to aggregate and query the timeseries data effectively, particularly in a DataFrame. For this section we've dumped out the two tables as CSV files, which can be downloaded here: BLS time series CSV tables. Querying Series Information The first step is to create a data frame from the series.csv file such that we can query information about each time series without having to store or duplicate the data. Working backward, we can create a function that accepts a BLS ID and returns the information from the info table: I utilize this function a fair amount to check if I have a time series in my dataset or to lookup seemingly related time series. In fact, we can see a pattern starting to emerge from function and the API fetch function from the last section. Our basic methodology is going to be to create functions that accept one or more BLS series ids and then perform some work on them. Unifying our function signatures in this way and working with our data on a specific key type dramatically simplifies exploratory workflows. However, the BLS ids themselves aren’t necessarily informative, so like the previous function, we need an ability to query the data frame. Here are a few example queries: This query returns all of the time series whose source is the Local Area Unemployment Statistics (LAUS) program, which breaks down unemployment by state. However, you’ll notice from the previous section that the prefixes of the series seem to be related but not necessarily to the source. We could also query based on the prefix to find related series: Combining queries like these into a functional methodology will easily allow you to explore the 3,368 series in this dataset and more as you continue to ingest series information using the API! Loading the Series The next step is to load the actual time series data into Pandas. Pandas implements two primary data structures for data analysis — the Series and DataFrame objects. Both objects are indexed, meaning that they contain more information about the underlying data than simple one or two-dimensional arrays (which they wrap). Typically the indices are simple integers that represent the position from the beginning of the series or frame, but they can be more complex than that. For time series analysis, we can use a Period index, which indexes the series values by a granular interval (by month as in the BLS dataset). Alternatively, for specific events you can use the Timestamp index, but periods do well for our data. To load the data from the records.csv file, we need to construct a Series per time series data structure, creating a collection of them. Here's a function to go about this: In this function we use the csv module, part of the Python standard library, to read and parse each line of our opened CSV file. The csv.DictReader generates rows as dictionaries whose keys are based on the header row of the csv file. Because each record is in the format (blsid, period, value) (with some extra information as well), we can groupby the blsid. This does require the records.csv file to be sorted by blsid since the groupby function simply scans ahead and collects rows into the rows variable until it sees a new blsid. Once we have our rows grouped by blsid we can load them into memory and sort them by time period. The period value is a string in the form 2015-02-01, which is a sortable format; however, if we create a pd.Period from this string the period will have the day granularity. For each row, we create a period using the .asfreq('M') to transform the period into the month granularity. Finally, we parse our values into floating points for data analysis and construct a pd.Series object with the values, the monthly period index, and assign a name to it — the string blsid, which we will continue to use to query our data. This function uses the yield statement to return a generator of pd.Series objects. We can collect all series into a single data frame, indexed correctly as follows: If you’re using our data, the series data frame should be indexed by period, and there should be roughly 183 months (rows) in our dataset. There are also 3366 time series in the data frame represented as columns whose column id is the BLS ID. If any of the series did not have a period matched by the global period index, the concat function correctly fills in that value as np.nan. As you can see from a simple head: the data frame contains a wide range of data, and the domain of every series can be dramatically different. Visualizing Series with Matplotlib Now that we’ve gone through the data wrangling hoops, we can start to visualize our series using matplotlib and the plotting library that comes with Pandas. The first step is to create a function that takes a blsid as input and uses the series and info data frames to create a visualization: The first thing this function does is look up the title of the series using the series info data frame. To do this it uses the get_value method of the data frame which will return the value of a particular column for a particular row by index. To look up the title by blsid, we will have to query the info data frame for that row, info[info.blsid == blsid], then fetch the index, and then we can then use that to get the 'title' column. After that we can simply plot the series, fetching it directly from the series data frame and plotting it using Pandas. Warning: Don’t try series.plot(), which will try to plot a line for every series (all 3366 of them); I've crashed a few notebooks that way! This function is certainly an enhancement of the series_info function from before, allowing us to think more completely about the domain, range, and structure of the time series data for a single blsid. I typically use this function when adding macroeconomic features to datasets to decide if I should use a simple magnitude, or if I should use a slope or a delta, or some other representation of the data based on its shape. Even better though would be the ability to compare a few time series together: In this function instead of providing a single blsid, the argument is a list of blsid strings. For each series, we plot them but add their title as a label. This allows us to create a legend with all the series names. Finally, we add a title that indicates the series blsid for reference later. We can now start making visual series comparisons: One thing to note is that comparing these series worked because they had the approximately the same range. However, not all series in the BLS data set are in the same range and can be orders of magnitude different. One method to combat this is to provide a normalize=True argument to the plot_multiple_series function and then use a normalization method to bring the series into the range [0,1]. Conclusion The addition of macroeconomic features to models can greatly expand their predictive powers, particularly when they inform the behavior of the target variable. When instances have some time element that can be mapped to a period, then the economic data collected by Census and BLS can be easily incorporated into models. This post was designed to equip a workflow for ingesting and exploring macroeconomic data from BLS in particular. By centering our workflow on the blsid of each time series, we were able to create functions that accepted an id or a list of ids and work meaningfully with it. This allows us to connect exploration both on the BLS data explorer, as well as in our data frames. Our exploration process was end-to-end with respect to the data science pipeline. We ingested data from the BLS API, then stored and wrangled that data into a database format. Computation on the timeseries involved the pd.Period and pd.Series objects, which were then aggregated into a single data frame. At all points we explored querying and limiting the data, culminating with visualizing single and multiple timeseries using matplotlib. Helpful Links BLS Timeseries CSV Tables BLS Developer Documentation Jobs Report BLS Database Ingestion Interactive Exploration of the Employment Situation Report Acknowledgements Nicole Donnelly reviewed and edited this post, Tony Ojeda helped wrangle the datasets. District Data Labs provides data science consulting and corporate training services. We work with companies and teams of all sizes, helping them make their operations more data-driven and enhancing the analytical abilities of their employees. Interested in working with us? Let us know!

Exploring Bureau of Labor Statistics Time Series

1

exploring-bureau-of-labor-statistics-time-series-1917d79bb01f

2018-05-13

2018-05-13 13:59:04

https://medium.com/s/story/exploring-bureau-of-labor-statistics-time-series-1917d79bb01f

false

2,728

Data science tutorials, thought pieces, and other awesome content.

null

DistrictDataLabs

null

District Data Labs

tojeda@districtdatalabs.com

district-data-labs

DATA SCIENCE,MACHINE LEARNING,ARTIFICIAL INTELLIGENCE,ANALYTICS,BIG DATA

DistrictDataLab

Data Science

data-science

Data Science

33,617

District Data Labs

Data science consulting firm, research lab, and open source collaborative.

96c976e31f28

DistrictDataLabs

921

471

20,181,104

null

0

null

0

32881626c9c9

2018-05-21

2018-05-21 22:15:37

2018-05-22

2018-05-22 14:43:01

3

false

en

2018-05-22

2018-05-22 14:43:01

6

1917e2d3c63d

3.142453

9

2

0

What started as a simple test of my company’s machine learning model builder quickly became an ethical dilemma.

5

Ethics of A.I. — what I learned building a gender detector What started as a simple test of my company’s machine learning model builder quickly became an ethical dilemma. Originally, this post was meant to be a step-by-step guide on building your own machine learning classifier using images you can easily download from Google. First find 100 pictures of men, then 100 pictures of woman, run this handy tool that automates building and training a machine learning model, and then post the results. Right? How do I write a technical post about using machine learning to classify gender without offense or misstating norms? According to this article by CBS news, I can use the term sex to describe the classification of a person as male or female at birth. Infants are assigned a sex, usually based on the appearance of their external anatomy Ok — that is a start. The value of detecting someone’s sex or gender in an image applies to marketing. Marketers target men and women with different ads, so if they can detect the perceived gender of someone walking up to a kiosk, for example, they can show those people more relevant ads. Facebook lets anyone target someone’s sex or stated gender in their profile, among many other things, so they’re incentivized to continue to help marketers go after their targets. And advertisers show no signs of slowing their targeted approach Does that mean its right, fair, or ethical? I’m not sure. Another vertical that is interested in quickly determining as much information as they can (from video and images of people) is security agencies. If you’ve ever listened to a police scanner, you’ve certainly heard law enforcement refer to people as either male or female. Automating the tagging of faces detected in surveillance footage is in use today, and is the continual subject of research into improvements in computer vision and visual recognition. Once again, I don’t know what that means ethically. What I do know is that even going about simply building this model for this post, I learned some potentially interesting challenges all of us in technology and business face with regards to bias in AI. As I mentioned before, my first step was to download 100 photos of women from Google. Screen shot of searching to “women” on Google Images Right away I felt confident my classifier would work. I could tell that the way women were photographed; the hair, the makeup, the poses etc. would be so different to the way men were photographed, that training a neural net would be a piece of cake. I’ll throw in a side note on some good practices of data cleaning here. When building a training set, make sure you have consistency. Pick photos of only one face, and try to avoid photos with anything other than the subject in the foreground or background. But was I really building a gender detector? Or was I simply building a tell-the-difference-between-the-way-the-two-most-culturally-accepted-genders-are-photographed model? And is that ethical or right? This is a screenshot of search results for men. Google Images search results for “men” We can visually tell from this dataset that the way these men are photographed is very different (with some exceptions). When I trained the model I expected good results, but didn’t quite expect an accuracy of 97%. Granted, the model is probably a bit overfit, but the difference in the two classes are significant enough that an image recognition system had no problem telling the difference with only 100 examples (actually, 80 examples, as it uses the remaining 20 for validation). I’m not really sure what this all means. Is there a nexus of bias and ethical problems in AI or is it simply just all about the training data? Whomever decides what the training data is may be the one who has to address any concerns related to ethics or bias, or maybe I have to address it when writing a post about building a gender classifier with machine learning. Either way, its good to get it out there and to have conversations about it since AI isn’t going anywhere.

Ethics of A.I. — what I learned building a gender detector

157

ethics-of-a-i-what-i-learned-building-a-gender-detector-1917e2d3c63d

2018-06-20

2018-06-20 20:05:13

https://medium.com/s/story/ethics-of-a-i-what-i-learned-building-a-gender-detector-1917e2d3c63d

false

687

Data Driven Investor (DDI) brings you various news and op-ed pieces in the areas of technologies, finance, and society. We are dedicated to relentlessly covering tech topics, their anomalies and controversies, and reviewing all things fascinating and worth knowing.

null

datadriveninvestor

null

Data Driven Investor

info@datadriveninvestor.com

datadriveninvestor

CRYPTOCURRENCY,ARTIFICIAL INTELLIGENCE,BLOCKCHAIN,FINANCE AND BANKING,TECHNOLOGY

dd_invest

Machine Learning

machine-learning

Machine Learning

51,320

Aaron Edell

b9a723f6b2a0

aaronedell

1,935

39

20,181,104

null

0

null

0

null

2018-02-23

2018-02-23 02:13:42

2018-02-23

2018-02-23 02:16:48

2

false

en

2018-03-31

2018-03-31 06:38:55

0

19186363959

2.304088

70

0

This is the first of a brand new series to get a closer look at MATRIX. We think communication is important and we want to share all of the…

4

Inside the MATRIX — #1: How does our AI assisted cancer diagnosis and treatment system work? The National Health and Welfare Commission invited us to share details This is the first of a brand new series to get a closer look at MATRIX. We think communication is important and we want to share all of the great things we’re achieving with our supporters. Recently we were invited to The Pathology Quality Control Evaluation Center to visit the remote disease diagnosis and quality control platform, as well as the Center for Disease Diagnosis, to carry out technical exchanges and discuss cooperation. The Pathology Quality Control Evaluation Center is a digital remote diagnosis and quality control platform for digital pathology established by the National Health and Welfare Commission of the People’s Government of the People’s Republic of China. The Center integrates various resources and the use of modern network technologies and digital slice image processing systems to achieve remote diagnosis goals. The Center informed us that the current medical and disease research is developing rapidly, and that although pathologists are experienced, it is difficult to diagnose the diseases encountered in a short period of time. It is hoped that the introduction of new technologies will provide more space for quick and accurate pathological diagnoses. Clinical diagnoses are made according to identifying common signs and patterns based on past information and images based on accumulated experience. AI can copy the whole process and learn quickly and more efficiently, whilst being able to recall the information more readily when it’s needed. Through deep learning technology, AI can continue to learn X-rays, medical books, essays, electronic medical records, and things like the formation of tumor tissue and normal histopathology; anything which is already available in huge databases. When a digital pathological section needs to be diagnosed, the most similar sections are quickly retrieved from the database by image recognition technology and diagnosed to help the doctor formulate a treatment plan. AI-assisted disease diagnosis and treatment is an inevitable future trend of medical development. It is the perfect combination of pathological image recognition technology and machine learning methods, and the deep learning facilities provided by the our bayesian Proof of Work algorithm is a perfect training ground. Not only will it help with Pathologists’ day-to-day work which requires high degrees of concentration and an extremely wide knowledge base, it also facilitates a greatly increased accuracy and efficiency of pathological diagnosis. The diagnosis and treatment solution we are currently building is one to aid in the diagnosis of thyroid cancer and liver cancer. We are working with Beijing Cancer Hospital, the 302 Hospital and a number of other first-line hospitals in the China to realize this goal — our solution will be in place in each of these hospitals when it’s ready. In the next phase, we will continue to strengthen communication with The Center, work to improve the AI technology used in digital pathological diagnosis with the hospitals, optimize the cooperation details and promote the common development of both parties.

Inside the MATRIX — #1: How does our AI assisted cancer diagnosis and treatment system work?

1,379

inside-the-matrix-1-how-does-our-ai-assisted-cancer-diagnosis-and-treatment-system-work-19186363959

2018-06-11

2018-06-11 03:21:25

https://medium.com/s/story/inside-the-matrix-1-how-does-our-ai-assisted-cancer-diagnosis-and-treatment-system-work-19186363959

false

509

null

Cancer

cancer

Cancer

17,070

MATRIX AI NETWORK

An open source public intelligent blockchain platform

ad51c60ef692

matrixainetwork

1,003

0

20,181,104

null

0

null

0

93da5952636b

2018-05-10

2018-05-10 10:55:20

2018-05-10

2018-05-10 10:55:21

0

false

en

2018-05-10

2018-05-10 10:55:21

14

191898eb0dcd

1.332075

4

0

null

5

Data Scientist versus Data Engineer: How are they different? Data Science is an approach to merge data analysis, business analytics, deep learning with other related methods. With the advent of digital technology, data has gained momentum in a variety of work areas as more minds are driven towards it. Discover the distinction between a data scientist and data engineer, to get a better understanding which is the best fit as a possible career option says, Jane Thomson, Content Marketing Manager, GreyCampus Data exists everywhere. In the current era of technology, digital data is expanding exponentially in the digital universe. Lately, much has been written about the different areas of data science roles, particularly the difference between data engineers and data scientists. The upsurge is specifically due to the fact that the new business models have replaced old ones. Qualitative data has superimposed quantitative data and the results are pretty mind-boggling. The role of a data engineer has moderately gained momentum along with a data scientist. Let’s discuss the key differences between data scientists and data engineers, primarily focusing on roles and responsibilities, educational qualifications, tools, languages and software, pay structure, job perspective etc. Generally speaking, data scientists arrange for cleaning, maneuvering and organizing big data. They use statistical measures and machine learning programs to assemble data used in prognostic modeling. They conduct industry research and rigorous data analysis to answer business requirements effectively. The data engineer is a person who fosters, builds, investigates and supports architectural databases and maintains enormous processing systems. Data engineers handle raw data that might consist of human, machine or instrument errors. They should recommend solutions and implement ways to improve data dependability, efficiency, and quality. Apparently, both parties are essential to crunch the data and provide valuable insights for making crucial business decisions. Considerably, the efforts of both parties are imperative to design and present the data differently in the most functional manner. One basic similarity between a data scientist and a data engineer is their Computer Science backgrounds. Although data scientists are more likely to have studied statistics, operations research, economics, and mathematics. Posted on 7wData.be.

Data Scientist versus Data Engineer: How are they different?

28

data-scientist-versus-data-engineer-how-are-they-different-191898eb0dcd

2018-05-17

2018-05-17 07:04:45

https://medium.com/s/story/data-scientist-versus-data-engineer-how-are-they-different-191898eb0dcd

false

353

Insights in the People, Process, Technology and Visualisations of the Data Landscape

null

7wdata

null

The Data Intelligence Connection

yves@7wdata.be

the-data-intelligence-connection

DATA,INNOVATION,AGILITY

7wdata

Big Data

big-data

Big Data

24,602

Yves Mulkers

BI And Data Architect enjoying Family, Social Influencer , love Music and DJ-ing, founder @7wData, content marketing and influencer marketing in the Data world

1335786e6357

YvesMulkers

17,594

8,294

20,181,104

null

0

null

0

634d4b270054

2018-06-06

2018-06-06 12:05:00

2018-06-06

2018-06-06 12:05:42

1

false

en

2018-06-06

2018-06-06 12:06:26

3

191901382e8d

0.920755

3

0

Hailo develops chips capable of processing artificial intelligence applications on edge devices installed in autonomous vehicles, drones…

5

Artificial Intelligence Chipmaker Hailo Raises $12.5 Million Hailo develops chips capable of processing artificial intelligence applications on edge devices installed in autonomous vehicles, drones, internet of things (IoT) devices and smart home appliances Tel Aviv-based artificial intelligence chip startup Hailo Technologies Ltd. has raised $12.5 million in a series A funding round, the company announced Tuesday. Investors include Jerusalem-headquartered equity crowdfunding platform OurCrowd, as well as two venture capital funds dedicated to automotive technologies — Maniv Mobility, and Next Gear Management Ltd. The latest round brings the company’s total funding raised to $16 million. See More @ https://bit.ly/2sGysFh About DEEPAERO DEEP AERO is a global leader in drone technology innovation. At DEEP AERO, we are building an autonomous drone economy powered by AI & Blockchain. DEEP AERO’s DRONE-UTM is an AI-driven, autonomous, self-governing, intelligent drone/unmanned aircraft system (UAS) traffic management (UTM) platform on the Blockchain. DEEP AERO’s DRONE-MP is a decentralized marketplace. It will be one stop shop for all products and services for drones. These platforms will be the foundation of the drone economy and will be powered by the DEEP AERO (DRONE) token.

Artificial Intelligence Chipmaker Hailo Raises $12.5 Million

101

artificial-intelligence-chipmaker-hailo-raises-12-5-million-191901382e8d

2018-06-16

2018-06-16 15:59:41

https://medium.com/s/story/artificial-intelligence-chipmaker-hailo-raises-12-5-million-191901382e8d

false

191

AI Driven Drone Economy on the Blockchain

null

DeepAeroDrones

null

DEEPAERODRONES

null

deepaerodrones

DEEPAERO,AI,BLOCKCHAIN,DRONE,ICO

DeepAeroDrones

Deepaero

deepaeros

Deepaero

0

DEEP AERO DRONES

null

dcef5da6c7fa

deepaerodrones

277

0

20,181,104

null

0

null

0

null

2018-01-03

2018-01-03 13:28:48

2018-01-03

2018-01-03 00:00:00

0

false

en

2018-01-03

2018-01-03 13:33:39

2

191a3efd494c

1.003774

0

Pre-hire screening is the process of examining the backgrounds of possible employees and is commonly used to verify the authenticity of an…

5

Pre-Hire Screening Will Save You From Post Hiring Hassles Pre-hire screening is the process of examining the backgrounds of possible employees and is commonly used to verify the authenticity of an applicant’s claims as well as to determine any likely illegal history, workers compensation claims, or employer sanction. It is one of the crucial processes before you hire a candidate for the organization. This thought must have strike a lot of people that hiring itself is a lengthy process why to make it more hectic. But this process can help you get rid of lot of troubles and could discourage applicants with a troubled history from applying for an open position. Some of the benefits of having a pre-hire screen are as follows: Helps you chosen the right hire It verifies education and certifications It highlights dishonesty Keeps employees and customers safe Gives you a full picture of your applicant The process should take place after initial applicant screening procedures are complete and an employer has made a conditional job offer. Once the process is underway, employers should not permit the new employee to begin working until the background check is complete. Importance and Value of Pre-Employment Background Some business owners might wonder if including a pre-employment background check during the hiring process is worth the additional time and expense. But when seriously considering a candidate for an open position, conducting a background check can save you from later hassles. Do conduct the pre-hire screen to avoid the post-hiring hassles! Originally published at www.human-telligence.com on January 3, 2018.

Pre-Hire Screening Will Save You From Post Hiring Hassles

0

pre-hire-screening-will-save-you-from-post-hiring-hassles-191a3efd494c

2018-01-03

2018-01-03 13:33:40

https://medium.com/s/story/pre-hire-screening-will-save-you-from-post-hiring-hassles-191a3efd494c

false

266

null

Hiring

hiring

Hiring

16,840

Siddhant Saxena

null

6c4ddee95499

sid_94916

2

15

20,181,104

null

0

null

0

null

2017-10-27

2017-10-27 22:06:56

2017-10-27

2017-10-27 22:47:45

0

false

en

2017-10-27

2017-10-27 22:52:13

1

191ad96b8d45

0.898113

0

Exciting!! The new product will be coming next week!

4

Boostinsider New Product Will Be Coming Exciting!! The new product will be coming next week! Boostinsider is the influencer marketing intelligence platform that reveals influencers’ full cycle insights on Youtube, Instagram, Facebook, Tumblr and Twitter. To efficiently improve influencer marketing performance, Boostinsider has created data-driven solutions of over 350K of the most engaged influencers on social media. Boostinsider created a series of the world’s first AI-based influencer marketing tools including: Social SaaS, the influencer search engine that helps marketers to best identify the right influencer for their brands. This search engine works together with the Campaign Management Platform to track and manage the entire influencer campaign cycle. SocialBook, the lighter version of Social SaaS, is an AI-powered data analytics tool for any YouTube Influencer. SocialBook provides real-time influencer channel insights, influencer market value, and audience insights at scale. Social Adwords is an automatic campaign management platform for CPC performance based campaigns, which leverages influencers on link-friendly platforms like Twitter, Facebook and Tumblr. BoostSDK encourages performance-based campaigns by making the revenue share model easy for influencers and mobile apps/game developers. Boostinsider, established in 2014, headquartered in Silicon Valley and has offices in Beijing, Chengdu and Shenzhen. Boostinsider has worked with over 300 brands, including Machine Zone, Cheetah Mobile, EA, Alibaba, IGG, and FitTea. Powerful Influencer Marketing Boostinsider is an influencer marketing platform where top social media influencers promote brands they love. Launch a…www.boostinsider.com

Boostinsider New Product Will Be Coming

0

boostinsider-new-product-will-coming-191ad96b8d45

2017-10-27

2017-10-27 22:52:14

https://medium.com/s/story/boostinsider-new-product-will-coming-191ad96b8d45

false

238

null

Influencer Marketing

influencer-marketing

Influencer Marketing

8,618

Sunny

Boostinsider Marketing Associate

7fcce2e54c

sunny_68369

0

3

20,181,104

null

0

null

0

null

2018-04-29

2018-04-29 14:52:54

2018-04-29

2018-04-29 14:57:55

2

false

en

2018-04-29

2018-04-29 14:57:55

1

191b3db6d61a

2.12673

1

0

Did you know that the quote “If only Hp knew what Hp knows, we would be three times more productive” by HP’s CEO Lew Platt can be subjected…

5

How To Preserve Knowledge Within An Organization Did you know that the quote “If only Hp knew what Hp knows, we would be three times more productive” by HP’s CEO Lew Platt can be subjected to several interpretations? It could be from the angle that Hp as a whole isn’t harnessing the several promising potentials that Hp has or Hp is not properly managing the premise of structure and organization in which it was built upon which is restricting them from being three times more productive. In essence, this writes up would be tied to both interpretations in reference to the preservation of knowledge within an organization. The following are the ways to efficiently preserve knowledge within an organization which would boost productivity; Sensitive Information Should Remain Within Boundaries Of The Organization In an organization where inter-communication cannot be controlled, dissemination of sensitive information should be carefully mapped out. A new invention or modification of a product or project should not be announced publicly. This means that when groundbreaking developments have been made, only some selected individuals who are key players in the organization should know about such developments and ensure it stays within the boundaries of the organization. The Key Players In The Organization Should Ensure That People Hear Only What They Want Them To Let’s face the fact, people would always talk to people. This is a factor that can’t be controlled but can be managed well. In a situation where it is an established fact that information tends to slip past the intended boundary then the blind spots should be played on. If the organization has discovered that several arms of the organization are responsible for the profit of the whole organization then it shouldn’t be presented as such. Only the key players should know the actual numbers because the future of the organization relies on such vital information and it’s better if the information remains restricted to only a few. Codified Messages Could Also Be Adopted In circumstances where the level of productivity of an organization rests on the circulation of information, all efforts should be dedicated to protecting such information. Adopting encrypted means of sending messages the organization would not be a bad idea. If a message is to be circulated within the organization with the knowledge of the risks surrounding the message then encryption is advised. Access to such messages could be through retina scan, thumbprints and other forms of encryption. To ensure productivity within an organization certain information which determines the productivity of the organization should be highly guarded against unauthorized persons to avoid any form of compromise and decline of expectancy. Tuqqi: Smarter Organizations for Brilliant People Tuqqi, your organization's personalized news platform, built from your internal & external working tools, to automate…www.tuqqi.com

How To Preserve Knowledge Within An Organization

1

how-to-preserve-knowledge-within-an-organization-191b3db6d61a

2018-04-29

2018-04-29 20:10:47

https://medium.com/s/story/how-to-preserve-knowledge-within-an-organization-191b3db6d61a

false

462

null

Design

design

Design

186,228

Rose Brewer

Freelance writer, Believes that using knowledge can be easier. Works at “Tuqqi” — the Founder of the enterprise knowledge management solution. . www.tuqqi.com

aa887f16f465

Tuqqi

3

27

20,181,104

null

0

null

0

8d070fd82f53

2017-08-04

2017-08-04 16:16:33

2017-09-24

2017-09-24 15:26:05

1

false

en

2017-09-24

2017-09-24 15:26:05

17

191b3eed5558

2.166038

4

0

How can fitness trackers aid medical research? GDPR: competitive advantage or compliance? Why do companies buy AI startups?

5

Do tech companies really need all that data? How can fitness trackers aid medical research? GDPR: competitive advantage or compliance? Why do companies buy AI startups? All included in this week’s curated data digest below. 👇 Do tech companies really need all that data? Walter Frick reports on a new working paper, by Lesley Chiou of Occidental College and Catherine Tucker of MIT, which explores whether “larger quantities of historical data affect a firm’s ability to maintain market share in Internet search”. This, in turn, measures the competitive advantage of data. The findings show that “historical data may be less valuable than fresher data” and that anonymising data “didn’t appear to impair the search experience”. Frick contextualises this in the current AI rush and questions whether the tech giants will have a clear advantage due to their “massive data troves” or whether “newcomers” can acquire “enough data to train intelligent systems”. Frick also questions whether the “trade-off” of giving up personal data “in exchange for products that are free and easy to use” is a correct assumption: Yes, people benefit from the many excellent and free tech products out there. Yes, they’ll probably benefit in countless ways from new AI-powered solutions. But they don’t always need to completely give up their privacy to get them. Nesta published their first report for DECODE (DEcentralised Citizen Owned Data Ecosystem) on the future of the personal data economy. Notably, the report points out the need for accessible anonymous data for the common good. How can fitness trackers aid medical research? Megan Molteni explores how FitBit data is being used to study “how exercise can help people deal with disease”. FitBit launched FitBase in 2012, and it has since “collected over 3.5 billion minutes of Fitbit data on behalf of research customers at places like John Hopkins, MD Anderson Cancer Center, and the Dana Farber Cancer Institute”. At current, FitBase data has been used to publish 457 research studies. One of the key findings has been how exercise can help to prevent cognitive decline amongst breast cancer survivors. GDPR: competitive advantage or compliance? Nice blog post by SSC on how the GDPR can position the UK as “one of the safest and most secure places worldwide in which to trust and conduct commerce”. The post references the “confidence-sapping effect of cyber breaches” and how this new regulation should drive innovation: Visionary tech leaders will be those who create precedent and make GDPR principles a badge of differentiation and trust. Who become champions of security and transparency to drive increased customer loyalty. Apparently, nearly a half million small businesses in the UK don’t have a clue about the upcoming GDPR. → Hogan Lovells explains how the GDPR may impact DeepMind’s use of patient data going forward. Miscellaneous Why do companies buy AI startups? 🤖 Why is Python growing so quickly? 🐍 The ten fallacies of data science. 🔮 A Wikipedia for Data Viz. 💯 Data Stories -> the “best worth viz”. 🎧 Developing a brainlike device for learning & computation (image above). 👾 Enjoy reading this? Join my free weekly data digest. 💌

Do tech companies really need all that data?

15

do-tech-companies-really-need-all-that-data-191b3eed5558

2018-06-15

2018-06-15 01:12:13

https://medium.com/s/story/do-tech-companies-really-need-all-that-data-191b3eed5558

false

521

Weekly digest on the world of data science, privacy and security — curated by Nick Halstead.

null

datascandigest

null

DataScan

null

datascan-digest

DATA,PRIVACY,DATA PROTECTION,DATA SECURITY,DATA PRIVACY

nik

Big Data

big-data

Big Data

24,602

Nick Halstead

Founder of InfoSum and DataSift

c17465ce895b

nik

4,835

293

20,181,104

null

0

null

0

null

2018-06-02

2018-06-02 16:55:52

2018-06-02

2018-06-02 18:48:25

8

false

en

2018-06-02

2018-06-02 18:48:25

1

191e09de786f

5.333333

5

0

For most of us who are new to AI neural network seems to be a very hi-fi thing but in reality a simple neural network is nothing else but…

5

Neural Network( Logistic Regression) for newbies! For most of us who are new to AI neural network seems to be a very hi-fi thing but in reality a simple neural network is nothing else but Logistic Regression. So in order to understand logistic regression lets first learn computational graph. A computational graph is just a graphical way of representing mathematical expressions. For example in the above given graph a and b are being fed to the square circles and their resultant is then fed to the square root circle from where it is stored in c. Now lets look at the computational graph of logistic regression- A logistic regression basically works on the principle of the equation - y=ax+b which we all have studied. in our schools. In the case of logistic regression the equation becomes Y=WX + B where X represents the input array, W represents the array of the weights which will have to be multiplied with the input array and B represents the intercept or bias(compare W and B with a and b of y=ax+b). Y is then send to a sigmoid function. A sigmoid function is basically an activation function which gives us a probability depending upon the value of Y. A sigmoid function- In the given graph you can see that as the value of z increases the probability of it being a 1 gets even more… Why we use sigmoid function? It gives probabilistic result It is derivative so we can use it in gradient descent algorithm (we will see as soon.) Till here it was a basic working principle of a simple neural network but the question arises that what if we are given a set of inputs(an input array) then how are we gonna decide: The correct value of weights for multiplying with the inputs Correct value of bias Is it the perfect model or do we need more layers(we will discuss them later) Forward Propogation: For that we will need to first do forward propogation. Forward propogation means assigning random or pre-selected values to weight(W) and bias(B) and then feeding them to the computational graph network. Then passing the result through the sigmoid function and finally calculating the loss(error) function on the output value. Lets learn what is loss(error) function Mathematical expression of log loss(error) function is that: where y is the actual value and y_head is the predicted value by the network. We won’t be delving too much into loss function, you can know about it more from here, wonderful lecture series for studying the maths behind ML and AI. Gradient Descent: Well, now we know what is our cost that is error. Therefore, we need to decrease cost because as we know if cost is high it means that we make wrong prediction. Lets think first step, every thing starts with initializing weights and bias. Therefore cost is dependent with them. In order to decrease cost, we need to update weights and bias. In other words, our model needs to learn the parameters weights and bias that minimize cost function. This technique is called gradient descent. Lets make an example: We have w = 5 and bias = 0 (so ignore bias for now). Then we make forward propagation and our cost function is 1.5. It looks like this. (red lines) As you can see from graph, we are not at minimum point of cost function. Therefore we need to go through minimum cost. Okey, lets update weight. ( the symbol := is updating) w := w — step. The question is what is this step? Step is slope1. Okay, it looks remarkable. In order to find minimum point, we can use slope1. Then lets say slope1 = 3 and update our weight. w := w — slope1 => w = 2. Now, our weight w is 2. As you remember, we need to find cost function with forward propagation again. Lets say according to forward propagation with w = 2, cost function is 0.4. Hmm, we are at right way because our cost function is decrease. We have new value for cost function that is cost = 0.4. Is that enough? Actually I do not know lets try one more step. Slope2 = 0.7 and w = 2. Lets update weight w : = w — step(slope2) => w = 1.3 that is new weight. So lets find new cost. Make one more forward propagation with w = 1.3 and our cost = 0.3. Okey, our cost even decreased, it looks like fine but is it enough or do we need to make one more step? The answer is again I do not know, lets try. Slope3 = 0.01 and w = 1.3. Updating weight w := w — step(slope3) => w = 1.29 ~ 1.3. So weight does not change because we find minimum point of cost function. Everything looks like good but how we find slope? If you remember from school, in order to find slope of function(cost function) at given point(at given weight) we take derivative of function at given point. Also you can ask that okay well we find slope but how it knows where it go. You can say that it can go more higher cost values instead of going minimum point. The asnwer is that slope(derivative) gives both step and direction of step. Therefore do not worry :) Update equation is this. It says that there is a cost function(takes weight and bias). Take derivative of cost function according to weight and bias. Then multiply it with α learning rate. Then update weight. (In order to explain I ignore bias but these all steps will be applied for bias) Now, I am sure you are asking what is learning rate that I mentioned never. It is very simple term that determines learning rate. Hovewer there is tradeoff between learning fast and never learning. For example you are at Delhi(current cost) and want to go Mumbai(minimum cost). If your speed(learning rate) is small, you can go Mumbai very slowly and it takes too long time. On ther other hand, if your speed(learning rate) is big, you can go very fast but maybe you make crash and never go to Mumbai. Therefore, we need to choose wisely our speed(learning rate). So at the end we know how to initialize weights and bias and how to calculate the loss function and how using that loss function re-calculate our weights and bias to find the perfect values for a simple neural network. Please do post your comments if you have any doubt somewhere or if you wish to know more about something. P.S.- Ignore the spelling mistakes😂

Neural Network( Logistic Regression) for newbies!

152

neural-network-logistic-regression-for-newbies-191e09de786f

2018-06-07

2018-06-07 05:01:39

https://medium.com/s/story/neural-network-logistic-regression-for-newbies-191e09de786f

false

1,113

null

Machine Learning

machine-learning

Machine Learning

51,320

Vaibhav Shukla

Student, IIIT Kalyani | 💻 Machine Learning | 🌀 Deep Learning Enthusiast

77263b9537e

vaibhavshukla182

8

12

20,181,104

null

0

null

0

31f4f88d6548

2017-09-07

2017-09-07 14:07:35

2017-09-07

2017-09-07 14:40:23

1

false

en

2017-09-07

2017-09-07 15:28:55

1

191e15d16423

1.301887

2

0

Learn more about Hollis Nymark, an M.S. Data Science student and one of four DeepMind Fellows at CDS this year

5

DeepMind Fellow Profile: Hollis Nymark Learn more about Hollis Nymark, an M.S. Data Science student and one of four DeepMind Fellows at CDS this year DeepMind is a leading artificial intelligence research company that tackles real-world challenges ranging from climate change to healthcare. They awarded scholarships to four M.S. in Data Science students at the NYU Center for Data Science (CDS) for the 2017–18 academic year. Hollis Nymark I grew up and currently reside in New York City. My interest in mathematics began when I was a child solving math puzzles with my father. After graduation from the Brearley School in Manhattan, I attended Boston College where I earned my B.A. in Mathematics. One of my first positions after graduation was working as a programmatic media analyst at Vivaki. In this role, I began to learn about how data has the power to shape one’s everyday life, and contributed several research papers to Vivaki’s thought leadership program. Currently, I am an Associate Director of Analytics at VM1 (a Zenith agency dedicated to Verizon). Some of the projects I have led include conducting an analysis to inform Zenith leadership of the accuracy and authenticity of ad verification vendors. Additionally, I designed and coached my team to conduct log-level based time lag to conversion analyses. It has been exciting and gratifying to not only carry out my own analyses, but also to lead and mentor a team on how to do so as well. At NYU, I look forward to exploring data outside of the advertising industry to solve new kinds of problems. As a DeepMind Fellow, I hope to not only learn new methods of analysis, but also contribute to projects focused on creating tools and products to improve everyday life.

DeepMind Fellow Profile: Hollis Nymark

3

deepmind-fellow-profile-hollis-nymack-191e15d16423

2018-05-07

2018-05-07 04:44:07

https://medium.com/s/story/deepmind-fellow-profile-hollis-nymack-191e15d16423

false

292

This is the official research blog of the NYU Center for Data Science (CDS). Established in 2013, we are a leading data science training and research facility, offering a MS in Data Science and, as of 2017, one of the nation’s first universities to offer a Ph.D. in Data Science.

null

nyudatascience

null

Center for Data Science

ab4829@nyu.edu

center-for-data-science

DATA SCIENCE,DATA MINING,TECHNOLOGY,ARTIFICIAL INTELLIGENCE,MACHINE LEARNING

NYUDatascience

Data Science

data-science

Data Science

33,617

NYU Center for Data Science

Official account of the Center for Data Science at NYU, home of the Master’s and Ph.D. in Data Science.

880781a85c2

NYUDataScience

3,530

9

20,181,104

null

0

plot(c(1,2,3,4),c(4,5,6,7)) import matplotlib.pyplot as plt plt.plot([1, 2, 3, 4], [1, 4, 9, 16]) plt.show()

2

1cae9dc6fddf

2018-03-21

2018-03-21 14:55:31

2018-03-21

2018-03-21 17:26:52

8

false

id

2018-03-22

2018-03-22 02:19:44

13

1922194707d1

4.061635

5

1

0

Saya mengenal bahasa pemrograman R sejak saya kuliah semester 5, sekitar 3 atau 4 tahun yang lalu, pada mata kuliah Aplikasi Pemrograman…

5

Alasan Mengapa R Jauh Lebih Baik dari Python Untuk Pemula Saya mengenal bahasa pemrograman R sejak saya kuliah semester 5, sekitar 3 atau 4 tahun yang lalu, pada mata kuliah Aplikasi Pemrograman Matematika di Universitas Pendidikan Indonesia. Pada mata kuliah itu awalnya kita diajarkan menggunakan aplikasi dekstop yang dapat digunakan untuk menyelesaikan masalah matematika, mulai dari MATLAB, SPSS (it’s for statistics, anyway), MAPLE, Geogebra dan yang terakhir R. Sepanjang yang saya ingat saya tidak begitu menikmati mempelajari pemrograman MATLAB dan R karena saya sendiri tidak terbiasa melakukan pemrograman. Walaupun saya cukup baik dalam pemrograman PASCAL, namun saya tidak begitu menikmati mata kuliah pemrograman. Karena menurut saya, saya tidak membutuhkannya ketika akan mengambil konsentrasi paling gaib di jurusan matematika: aljabar. Well, bagi yang ingin tau apa yang saya kerjakan mungkin bisa melihat sekilas isi skripsi saya: Ehh… Tapi, sejak menghadapi dunia nyata (red: dunia dimana lulusan baru plenga-plengo mencari kerjaan), saya menyadari bahwa kemampuan programming sangat dicari oleh perusahaan-perusahaan, baik kecil maupun besar, terutama bagi mereka yang ingin menjadi data scientist (the better coder among statistician, and the better statistician among programmer). Mau tidak mau saya harus belajar bahasa pemrograman dan sains data, karena jelas tidak ada satupun perusahaan yang menganggap bahwa aljabar graf ultramatricial atau Complex Kumjian-Pask algebras bisa meningkatkan ROI perusahaan. Dan yang dari jurusan Ilmu Komputer nanya: ini topologi jaringan? Saat itu ada dua bahasa pemrograman yang umum dipelajari untuk menjadi data scientist, yaitu R dan python. Saya sudah mempelajari R sejak kuliah sedangkan python termasuk baru saya pelajari. Keduanya menjanjikan hal yang sama: kemudahan. Namun, setelah saya mempelajari keduanya saya memilih R sebagai bahasa yang harus saya kuasai. Berikut adalah alasan mengapa saya lebih memilih R dibanding python. R lebih mudah dan intuitif untuk dipelajari dan digunakan untuk analisis dan manipulasi data. Dalam sejarahnya, R didesain sebagai bahasa pemrograman yang digunakan untuk melakukan analisis dan manipulasi data, sedangkan python merupakan bahasa pemrograman yang sifatnya umum (disebut juga general-purpose programming language). Oleh karena itu, R memiliki banyak built-in function yang dapat digunakan untuk melakukan analisis data dibandingkan dengan python. Misalnya untuk mencari mean atau rerata dari sebarisan data, di R kita cukup menggunakan fungsi mean() ketika para pengguna python harus membuat fungsi khusus atau import library numpy dan menggunakan method numpy.mean(). Bahkan untuk sekedar menjumlahkan dua buah vektor akan jauh lebih intuitif di R dibandingkan dengan melakukannya di python. Jumlah dua buah vektor di R Sweet! ‘Jumlah’ (?) dua buah vektor di python Ehh… what? Bagi yang ingin tau bagaimana cara menjumlahkan dua buah vektor di python bisa merujuk ke link berikut. Setidaknya agar bisa menjumlahkan kedua buah vektor dengan mudah kita harus menggunakan numpy. Tidak hanya untuk analisis data, bahkan untuk melakukan plotting data, R jauh lebih simpel dibandingkan dengan python, cukup dengan menggunakan fungsi `plot()`, masukkan nilai x dan y (jika dua dimensi), maka plot pun jadi: Sedangkan jika menggunakan python harus memanggil library matplotlib dengan cara menggunakan method plt sebagai berikut: R lebih simple kan? Umumnya hasil riset terbaru selalu diimplementasikan di bahasa pemrogramman R R menerbitkan jurnal open-access yang berisi implementasi teori-teori statistika atau matematika tertentu dalam bahasa pemrograman R. Oleh karenanya, R memiliki package yang up-to-date mengikuti perkembangan ilmu pengetahuan. Tidak heran jika R memiliki lebih dari 12.336 packages (dan mungkin terus tumbuh) yang siap digunakan untuk melakukan analisis yang anda inginkan. Mulai dari packages yang berkaitan dengan statistika sampai machine learning. Bahkan TensorFlow dan Keras yang dibanggakan para pengguna python juga ada di R. R nyaris bisa digunakan untuk apa saja: mulai dari membuat web sampai robotik “Tapi kan python lebih general, python jauh lebih hebat dong dari R!”. Wait: Yes! Bahkan tidak perlu menggunakan atau mempelajari HTML, CSS dan javascript! Dengan menggunakan package Shiny, kita bisa membuat web hanya dengan menggunakan bahasa pemrograman R, baik untuk layout sampai business logic. Anda bisa melihat contoh aplikasi web yang dibuat dengan shiny pada situs berikut. Contoh aplikasi web dibuat menggunakan shiny “Tapi kan Shiny terbatas tampilannya cuman gitu-gitu doang, sedangkan kalau ngoding pakai python kita bisa bikin tampilan lebih dinamis…” Let me introduce you OpenCPU and Plumber. Dengan menggunakan OpenCPU kita bisa membangun aplikasi web dengan memanfaatkan kemampuan R, sedangkan dengan menggunakan Plumber kita bisa membuat REST API. Dengan bantuan ROS (Robotic Operating Systems), R dapat digunakan untuk melakukan pemrograman robotik. Bahkan dengan R kita bisa membuat labirin untuk game Minecraft I mean, for real…?! Kesimpulan R memang memiliki kelebihan yang menurut saya jauh lebih banyak dibandingkan python dalam hal kemudahan penggunaan, terutama dalam analisis data. Walaupun begitu R juga memiliki beberapa kekurangan yang perlu dipertimbangkan seperti performance atau kecepatan dalam mengeksekusi program yang masih kalah dibandingkan python. Walau begitu, dengan mempertimbangkan kemudahan dan fleksibilitas, R tetap yang terbaik.

Alasan Mengapa R Jauh Lebih Baik dari Python Untuk Pemula

6

alasan-mengapa-r-jauh-lebih-baik-dari-python-1922194707d1

2018-04-17

2018-04-17 05:00:26

https://medium.com/s/story/alasan-mengapa-r-jauh-lebih-baik-dari-python-1922194707d1

false

776

Indonesia R Enthusiast

null

IndoR

null

indor

R PROGRAMMING,STATISTICS,MACHINE LEARNING,DATA SCIENCE,ANALYTICS

null

Data Science

data-science

Data Science

33,617

Hadi Suryo (hadimaster65555)

A statistical consultant in StatsMaster and Web Developer. Meet him at hadimaster65555.github.io

2dad1c6c3695

hadimaster65555

8

13

20,181,104

null

0

null

0

null

2017-09-30

2017-09-30 19:38:48

2017-09-30

2017-09-30 19:50:37

1

false

en

2017-09-30

2017-09-30 19:51:53

0

1923e4a07ac1

5.316981

5

0

It doesn’t matter what industry you’re in, you’ve likely heard, or even had talks, about using personalization in your online marketing…

5

2017 Best Marketing Segmentation Tools Best Marketing Segmentation Ideas It doesn’t matter what industry you’re in, you’ve likely heard, or even had talks, about using personalization in your online marketing. Segmentation is nothing new in general marketing, but it has definitely transformed. Today, we are using various factors to segment and personalize the experience of customers and prospects. Particularly, we’re beginning to see the importance of gender marketing segmentation and how it relates to improving conversion rates. As you may already know, today’s consumers are more savvier than previous generations. This means we have to get savvier with our marketing. Why Should Marketing Segmentation Matter to You? Now, there are multiple reasons why marketing segmentation is important in every industry. In most cases, you have more than one type of customer. And it’s very likely you attract both male and female clientele. After all, there are women working in just about every industry these days. And more women are taking the role as the one who makes purchasing decisions for the household. So having the ability to attract and appeal to both markets is essential. But to put it for you plainly, here is a quick list of the benefits gender marketing can offer your business: Precise audience targeting Better understanding of consumers Increased conversion rates Improved customer outreach How Are Gender Roles Changing? We know of all the different details we can collect about consumers based on analytics data we gather. But why hone in on the gender tidbit? Maybe because gender is no longer just an identifier of your physical sex. And it’s likely because we’re finally seeing the psychology behind marketing to these sets of people. Even Facebook has found gender to be of importance. This is why it now offers 71 options for gender. Plus, it has a gender option, so people can identify themselves how they want to on their platform. But that’s just political correctness. Needless to say, it’s a very smart move. It’s very important that as a business, you roll with the changes society throws. Otherwise, your brand could end up being placed on the back burner by consumers. Now, although we are witnessing changing gender roles and how we market to these individuals, we still have to take gender into account. Otherwise, your marketing campaigns aren’t going to be as effective. Will Changing Gender Roles Affect Marketing? Most of us grew up with the concept of man and woman, which had clear and more defined gender roles. In a nutshell, gender was determined by the sex you were born. So you were either a male or a female. However, society is now adopting new concepts to this — some people who are born one sex may psychologically identify as another sex. The strict gender norms, stereotypes and roles we’ve grown accustomed to are no longer written in stone. According to the American Psychological Association, your true gender is determined by the interrelationship between gender expression, gender identity and gender biology. So in other words, your physical sex, what you identify yourself as and how you behave or present yourself is what determines your gender. This can play out in multiple ways. For example, you could have a person who was born a woman, who identifies as a man, but still dresses and acts feminine. Or you could have a male, who identifies as a woman and dresses as one. So the gender spectrum is pretty broad, which creates a bit of a conundrum for marketers. In the past, it was simple to identify a person by their gender, but today, it is completely up to the individual. As we continue down the path to gender equality and diversity, these concepts will shift more. This is especially true for the younger generations, who are being brought up with this new way of thinking and self-expression. For instance, you will now find gender-neutral pronouns and uniforms being inserted in elementary schools. Young children are even being given the option to identify as male or female. And, not to mention, the gender-neutral restrooms being added to schools and businesses. It’s a new world and a new day, so it’s important for your business to keep up. How Can Marketers Adapt? This is what marketers do best — adapt to the needs of the consumers they’re trying to attract. While this is a tricky situation, it’s not impossible. As always, it’s about knowing your customers by learning everything you can about them. What this means is identifying your audience’s emotional needs, aspirations, desires, hopes and functional needs. Once you know this, you can make strategic decisions on how to proceed with your marketing campaigns. The key is backing away from gender stereotypes as much as possible. Shifting your products and services away from focusing on gender-specific and transitioning them to be gender-neutral. This is easier in certain industries, such as insurance and personal finance. But for industries that are gender-specific, such as makeup and hair styling tools, this is different. The key would be to focus on the needs of the customer versus their gender. Step away from seeing your customer as male or female and just focus on solving their problems. Does this mean you should ignore knowing the gender of your audience? Absolutely not. If anything, you should reach out to them based on their perceived gender, and then allow them to choose the gender they prefer to be addressed as in your communications with them. And that brings me to the gender database tools you can use to determine the sex of your audience in advance. Let’s review which are the best of the best. We’ll start with the #1 database on the market. #1 Name Gender Pro You’ll find this database mentioned several times throughout my e-book and for good reason. You can find a great collection of gender-based names gathered from the top English-speaking countries around the world — not just the USA. For instance, these names derive from Canada, United Kingdom and Australia as well. These names are gathered from publicly available data sources. So if you’re selling products or services on a global scale, it’s good to have a tool like this. This way, you can determine the gender of your customers and prospects before reaching out to them. You won’t find a better price for a name gender database. And even better, you get to keep the database forever. You only have to worry about the upfront cost and that’s it. The other tools require a monthly or annual rate to maintain access to their databases. With Name Gender Pro, you’re getting a reliable database that’s simple to install. Once you do, you get instant access to the database in the form you want — csv or sql. In just 10 minutes, you can have your databased updated with names based on gender, allowing you to start segmenting your marketing campaigns immediately. Hands down, it’s the best tool on the market today. #2 Gender API In this database, you get access to hundreds of thousands of names collected from 178 different countries. However, not all of the countries are English-speaking. This may not particularly cater to your audience. Like Name Gender Pro, you can integrate it with your own database. But the cost is higher and you won’t own the data. Also, the database is smaller in comparison to Name Gender Pro. #3 US Census This is the source that NameGenderPro.com uses for its database. With this option, you get to own the data, but you’ll have to do all of the work to put together all the data. A scraper is needed to parse the data. Also, the database is much smaller, but it is free. But again, the amount of time you’ll have to spend on it may not be worth it. After all, you have to separate the unisex first names on your own. On that note, having a tool you can trust will come in handy when you begin personalizing your marketing campaigns. So consider using one to help improve your conversion rates!

2017 Best Marketing Segmentation Tools

109

2017-best-marketing-segmentation-tools-1923e4a07ac1

2018-03-23

2018-03-23 03:00:14

https://medium.com/s/story/2017-best-marketing-segmentation-tools-1923e4a07ac1

false

1,356

null

Growth

growth

Growth

22,187

Gender Marketing & Consumer Behavior

Group of academic scholars and marketing analytics experts with in-depth knowledge in gender, marketing and consumer behavior. Visit NameGenderPro.com for more.

70dfa10a1cd6

namegenderpro

5

1

20,181,104

null

0

null

0

null

2018-08-31

2018-08-31 18:13:51

2018-08-31

2018-08-31 18:24:36

5

false

es

2018-09-07

2018-09-07 14:14:59

0

1924f81d746e

1.338994

0

¿Que es arte conceptual?

5

Arte conceptual ¿Que es arte conceptual? Según mi entendimiento es un tipo de arte que va mas allá de lo que esta siendo demostrado o expuesto en físico y lleva una idea representada de forma abstracta. Obras: Mi obra: Utilizando la pagina web deep dream generator pude escoger dos fotografías y esta con la ayuda de inteligencia artificial creaba una nueva imagen con la primera imagen como base pero le atribuye características y colores de la segunda imagen. La inteligencia artificial creo una imagen a mi parecer muy interesante ya que tomo la imagen de mi carro en una curva y creo a un monstruo reptil de tres cabezas persiguiéndolo. Crea una imagen sacada de películas de ciencia ficción en la cual yo estoy escapando de un monstruo en una realidad o dimensión distorsionada.

Arte conceptual

0

arte-conceptual-1924f81d746e

2018-09-07

2018-09-07 14:14:59

https://medium.com/s/story/arte-conceptual-1924f81d746e

false

134

null

Art

art

Art

103,252

Sebastian Patricio Bustamante Dellepiane

null

e694dbb8e0b4

sebastian.bustamante

0

1

20,181,104

null

0

null

0

3155ee25db64

2018-07-30

2018-07-30 17:25:42

2018-07-30

2018-07-30 17:28:06

1

false

en

2018-07-30

2018-07-30 17:28:06

0

1926e3d31dec

1.237736

0

PALO ALTO (March 1, 2018) — ModuleQ, the provider of AI-powered delivery solutions for business information providers, names Norman Tsang…

5

Norman Tsang Named As ModuleQ’s CFO Norman Tsang named as ModuleQ’s CFO PALO ALTO (March 1, 2018) — ModuleQ, the provider of AI-powered delivery solutions for business information providers, names Norman Tsang, CFO. Tsang brings broad and relevant experience as a Financial and Operating Executive at multiple companies with successful exits. Dr. David Brunner, ModuleQ Founder & CEO said “I’m extremely pleased to welcome Norman to our team. His financial, operational and strategic skills are critical for our continued expansion plans.” Tsang added “ModuleQ’s AI technology is truly game-changing. I am excited to join such a team deep of technology executives and PhDs.” Tsang has been successful as both an operating and financial executive. He was on the Board and Interim COO for ENSO Financial Analytics, a SaaS based information provider that raised multiple investment rounds and was later acquired by ICAP. He Co-Founded and was a Managing Partner at Alerion Partners which raised 2 successive Funds. He invested in and was a Board member for Alerion’s most successful portfolio companies including ScentAir Technologies, Convergence Marketing, Enviroscent and True Citrus. Norman has been an executive at VC/PE backed companies including VP Sales/Marketing, PSI self-checkout (acquired by IBM) and VP Marketing, Star Market (acquired by Sainsbury). Earlier in his career he also worked at McKinsey, Citi and IBM. Norman has an MBA from Stanford and BS from NYU. ModuleQ’s intelligent agent platform helps information providers drive revenue by reaching professionals directly with automated delivery of personalized content. ModuleQ’s unique Personal Data Fusion AI uses Office 365 data to identify the business information that professionals need to make better decisions and have more impactful business conversations. Contact ModuleQ Sherry@ModuleQ.com

Norman Tsang Named As ModuleQ’s CFO

0

norman-tsang-named-as-moduleqs-cfo-1926e3d31dec

2018-07-30

2018-07-30 17:28:07

https://medium.com/s/story/norman-tsang-named-as-moduleqs-cfo-1926e3d31dec

false

275

With ModuleQ's People-Facing AI insights, you don't need to find critical information. It finds you. Business Answers, Before You Ask.

null

moduleq

null

ModuleQ

null

moduleq

AI,ARTIFICIAL INTELLIGENCE,DATA

moduleq

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

ModuleQ

Business Answers, Before You Ask. With ModuleQ's People-Facing AI insights, you don't need to find critical information. It finds you.

3b32c10d111e

ModuleQ

0

1

20,181,104

null

0

null

0

null

2017-10-04

2017-10-04 10:33:40

2017-10-04

2017-10-04 10:37:39

1

false

en

2017-10-04

2017-10-04 10:54:50

0

19279ac9bde9

1.679245

1

0

What’s the difference between Machine Learning and Deep Learning?

5

Artificial Intelligence Q&A What’s the difference between Machine Learning and Deep Learning? Machine Learning is a field of research, whereas Deep Learning is a technique used in Machine Learning. The Machine Learning method of data analysis works extremely well and was invented 20 years ago, but only now has it begun to really show its potential. Self-driving cars, online recommendations, social media listening, fraud detection — all modern uses of Machine Learning. Will AI replace humans? Absolutely not, while there are situations at present wherein humans have already been replaced by computers/algorithms/robotics, etc., those replacements were made in areas where there was an extremely limited input and output. That’s why it’s commonplace to have robotics working in areas such as car manufacturing, because the production lines are easy to automate. However, human input is still very much required in a variety of other areas. AI struggles massively while attempting to tackle tasks involving creativity or education for example — therefore it’s safe to say that AI will not replace humans. What’s the next challenge in AI? The next big challenge is simply to create and appropriately implement AI in real-world operations so our lives can be further improved. Where do you see AI in 5–10 years? It’s certain that there’ll be more data available. Simply put, an abundance of data will lead to better AI. This is because the data allows for better precision, better performance and ultimately better real-world applications. Are you or will you be afraid of AI? Ultimately, it depends on the application of AI. There’s a lot of good things that it can help with, but obviously there are darker applications too. I think it comes down to what the AI is being used to do — there may be a need in the future for AI regulation but for now, there is no reason to fear AI. Can AI ever be conscious or self-aware? From a researcher’s point of view, that’s the dream: to have machines that think. Having machines that not only understand the world around them, but also also understand their own inner state. As a field of study, we’re just not there yet. Will we reach that point in 5–10 years? It’s not clear. As a society, we will have to tentatively cross that bridge when we come to it.

Artificial Intelligence Q&A

1

artificial-intelligence-q-a-19279ac9bde9

2018-06-08

2018-06-08 20:32:21

https://medium.com/s/story/artificial-intelligence-q-a-19279ac9bde9

false

392

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

beautifeye

Automating market research with Artificial Intelligence

92061bf597c4

beautifeyelabs

94

157

20,181,104

null

0

null

0

null

2017-12-04

2017-12-04 19:05:09

2017-12-04

2017-12-04 19:14:26

0

false

en

2018-09-02

2018-09-02 22:03:29

14

1928813394c7

23.532075

2

0

When considering robots, intelligent agents, and intelligent digital assistants, questions of autonomy and agency arise. This informal…

5

What Are Autonomy and Agency? When considering robots, intelligent agents, and intelligent digital assistants, questions of autonomy and agency arise. This informal paper attempts to define these key concepts more clearly and explore questions of what are they, how they are different, and how they are related. Synthesized definitions for autonomy and agency will be provided after discussing all the relevant aspects of these concepts. A companion paper, Intelligent Entities: Principals, Agents, and Assistants, will build on these concepts, but essential concepts of intelligent entities, principals, agents, and assistants will be introduced here as well since these two sets of concepts are closely related and intertwined. You can’t have one without the other. Note that this paper is more focused on people, robots, and software agents rather than on countries or autonomous regions within countries (e.g., Catalonia in Spain or Kurdistan in Syria and Iraq), although the basic definition of autonomy still applies to those cases as well. Also note that sociology, philosophy, and agent-based modeling and simulation use the terms agency and agent as the terms autonomy and autonomous entity are used in this paper (freedom of choice and action, unconstrained by any other entity.) For quick reference, see the section entitled Definitions of autonomy and agency for the final definitions of these terms. Dictionary definitions A later section of this paper will come up with synthesized definitions for autonomy and agency that are especially relevant to discussion of intelligent agents and intelligent digital assistants, but the starting point is the traditional dictionary definitions of these terms. Definition entries from Merriam-Webster definition for autonomy: self-directing freedom and especially moral independence the state of existing or acting separately from others the quality or state of being independent, free, and self-directing the quality or state of being self-governing Definition entries from Merriam-Webster definition for agency: the relationship between a principal and that person’s agent the capacity, condition, or state of acting or of exerting power — operation a person or thing through which power is exerted or an end is achieved — instrumentality a person or thing through which power is used or something is achieved a consensual fiduciary relationship in which one party acts on behalf of and under the control of another in dealing with third parties the power of one in a consensual fiduciary relationship to act on behalf of another general agency — an agency in which the agent is authorized to perform on behalf of the principal in all matters in furtherance of a particular business of the principal special agency — an agency in which the agent is authorized to perform only specified acts or to act only in a specified transaction the law concerned with the relationship of a principal and an agent Three related terms are entity, principal, and agent. Definition entries from Merriam-Webster definition for entity: independent, separate, or self-contained existence something that has separate and distinct existence and objective or conceptual reality There are other meanings for entity, but those are the senses relevant to this paper. Definition entries from Merriam-Webster definition for principal: a person who has controlling authority or is in a leading position a chief or head man or woman the chief executive officer of an educational institution one who engages another to act as an agent subject to general control and instruction the person from whom an agent’s authority derives the chief or an actual participant in a crime the person primarily or ultimately liable on a legal obligation a leading performer — star Definition entries from Merriam-Webster definition for agent: one that acts or exerts power something that produces or is capable of producing an effect a means or instrument by which a guiding intelligence achieves a result one who is authorized to act for or in the place of another a computer application designed to automate certain tasks (such as gathering information online) a person who does business for another person a person who acts on behalf of another a person or thing that causes something to happen something that produces an effect a person who acts or does business for another someone or something that acts or exerts power a moving force in achieving some result a person guided or instigated by another in some action a person or entity (as an employee or independent contractor) authorized to act on behalf of and under the control of another in dealing with third parties Intelligent entities Autonomy and agency are all about intelligent entities and their freedom to make decisions and take actions, and their authority, responsibilities, and obligations. Generally, an entity is any person, place, or thing. In the context of autonomy and agency, an intelligent entity is a person or thing which is capable of action or operation and at least some fraction of perception and cognition — thought and reason, coupled with memory and knowledge. More specifically, an intelligent entity has some sense of intelligence and judgment, and is capable of making decisions and pursuing a course of action. Whether or not an intelligent entity has autonomy or agency is not given: Some entities may have autonomy, but not agency. Some entities may have agency but not autonomy. Some entities may have both autonomy and agency. Some entities may have neither agency nor autonomy. Computational entities An intelligent entity can be a person or a machine or software running on a machine. The latter are referred to as computational entities or digital entities. They include: Robots Driverless vehicles Smart appliances Software agents Intelligent agents Digital assistants Intelligent digital assistants Apps Web services How much autonomy or agency a given computational entity has will vary greatly, at the discretion of the the people who develop and deploy such entities based on needs, requirements, desires, preferences, and available resources and costs. Sometimes people want more control over their machines, and sometimes they value greater autonomy, agency, or automation to free themselves from being concerned over details. Entities As a convenience and for conciseness, this paper will sometimes use the shorter term entity as implicitly referring to an intelligent entity, either a person or a computational entity. Actions and operations Definitions: Action. Something that can be done by an entity. An observable effect that can be caused in the environment. Operation. Generally a synonym for action. Alternatively, an action that persists for some period of time. For example flipping a switch to turn on a machine is an action, while the ongoing operation of the machine is an operation. The flipping of the switch was an operation too, only of a very short duration. If a machine would operate only while a button was depressed, the pressing and holding of the button as well as the operation of the machine would both be actions and operations. Tasks, objectives, purposes, and goals Definitions: Task. One or more actions or operations intended to achieve some purpose. Purpose. The reason or desired intent for something. Goal. A destination or state of affairs that is desired or intended, but without a plan for a set of tasks to achieve it. Objective. Synonym for goal. Subgoal. A portion of a larger goal. A goal can be decomposed into any number of subgoals. Motivation. The rationale for pursuing a particular objective or goal. Intentions. Desired objective or goal. What is desired, not why or how. Principals and agents A separate paper, Intelligent Entities: Principals, Agents, and Assistants, will delved deeper into principals and agents, but definitions for the purposes here: Principal. An intelligent entity which has the will and desire to formulate an objective or goal. Agent. An intelligent entity which has the capacity and resources to pursue and achieve an objective or goal on behalf of another intelligent entity, its principal. A given entity may be either: A principal but not an agent. Does all actions itself, without any delegation to agents. An agent but not a principal. Both a principal and an agent. A principal for subgoals. Neither a principal nor an agent. Possibly an assistant for specific tasks, but not any goals. Delegation of responsibility and authority The essence of the relationship between principal and agent is delegation. The principal may delegate responsibility and possibly even authority for one or more objectives or goals to one or more agents. Principal as its own agent Some intelligent entities may act as both principal and agent, doing its own work, rather than delegating its work to one or more agents. Agent as principal for subgoals For more complex objectives, an agent may decompose a larger goal into subgoals, with each subgoal delegated to yet another agent for whom this agent acts as principal. Authority The authority of an intelligent entity is the set of actions that the entity is permitted to take. A principal would have unlimited authority. An agent would have limited authority related to the goal(s) that the principal is authorizing the agent to pursue. In the real world, many principals are in fact agents since they act on behalf of other principals. A company has a board of directors, investors, and shareholders. Robots have owners. Responsibility, expectation, and obligation The responsibility of an intelligent entity is the set of expectations and obligations of the entity in terms of actions. A principal has no responsibility, expectations, or obligations per se. A principal may act as it sees fit. An agent has responsibility, expectations, and obligations as set for it by its principal. An agent may act as it sees fit, provided that its actions satisfy any limitations or constraints set by its principal. In the real world, many principals are in fact agents since they act on behalf of other principals. A company has a board of directors, investors, and shareholders. Robots have owners. So a company or robot may have responsibilities, expectations, and obligations set by somebody else. General obligations Regardless of obligations which result from autonomy and agency, all intelligent entities will have general obligations which spring from: Physics. Obey the laws of physics. Reality. The real world. Natural law. For example, gravity, entropy, and the capacity of batteries. Limited resources and their cost. For examples, the availability and cost of electricity, storage, computing power, and network bandwidth. Laws. Obey the laws of man. Including regulations and other formalized rules. Ethics. Adhere to ethical codes of conduct. Including professional and industry codes of conduct. Ethics Just to reemphasize from the previous section, that intelligent entities will have to adhere to ethical considerations in the real world. Liability A principal may be exposed to liability to the extent that it enlists the aid of an agent and that agent causes harm or loss or violate laws or rules while acting on behalf of the principal. Requested goals might have unintended consequences which incur unexpected liability. An agent may be exposed to liability if it naively follows the guidance of its principal without carefully reviewing whether specified goals, expectations, or obligations, might cause harm or loss or violate laws or rules when carried out. Elements of a goal A goal must be: Formulated. Clearly stated. Planned. A strategy developed. A plan developed. Resources allocated. Tasks identified. Pursued. Individual tasks performed. Decisions may need to be made or revised and the original plan adapted based on results of individual tasks. Achieved or not achieved. The results or lack thereof. Relationship between principal and agent Power, action, control, and responsibility are involved in formulating a plan for setting and pursuing objectives and goals. Power. The principal has the power to set the objectives and goals to be pursued. The agent has only the delegated power to select tasks to achieve the objectives and goals set by the principal and to pursue them through actions, but no power to change the objectives or goals themselves. Action. The agent is responsible for performing the actions or tasks needed to achieve the objectives and goals set by the principal. The agent is also responsible for deciding what tasks and actions must be performed to achieve the objectives and goals, and for coming up with a plan for performing them Control. The principal controls what objectives and goals are to be pursued. The agent controls what tasks and actions must be performed to achieve the objectives and goals and how to perform them. The principal is in charge. The principal is the boss. The agent is subservient to the principal. The principal delegates to agents or assistants. Contracts Generally there is a contract of some form between a principal and its agent, which clearly sets out the objectives and goals, responsibilities, expectations, and obligations of both parties, both the principal and the agent. The contract details what is expected of the agent, what the agent is expected to deliver, what the agent needs to pursue the specified goals, including resources, and what compensation the agent will receive in exchange for achieving the goals. The contract also details any limitations or restrictions that will apply to the agent and its work. The contract authorizes and empowers the agent. Contracts are needed both for human entities and for computational entities. Capacity for agency There are really two distinct senses of agency: The capacity to act or exert power. The relationship between a principal and an agent that empowers the agent to operate on behalf of the principal. The latter requires that there is a principal involved, doing the empowerment, the authorization to act on its behalf. The former can exist even if there is no principal present. An intelligent entity can act on its own interests, on its own behalf, being its own principal. An entity can be self-empowering. That’s what it means for an entity to have agency in a traditional, sociological or philosophical sense. The first sense is true in both instances, where either a principal is present as an external entity, and when no principal is present. In the context of intelligent agents and intelligent digital assistants, agency usually refers to the latter sense, that the agent is acting on behalf of the principal, which is commonly a human user, but may also be some other computational entity, such as another intelligent agent or a robot. Assistants A separate companion paper, Intelligent Entities: Principals, Agents, and Assistants, will introduce the concept of an assistant, which is quite similar to an agent in the sense that it is capable of performing the tasks needed to achieve goals, but can only perform specific tasks as dictated by its principal without any sense of any larger goal or objective that the task is needed to achieve, and with much less room for discretion as to how to perform the tasks. An assistant has limited agency in that it performs tasks on behalf of a principal but it lacks the authority or capacity to decide which tasks to perform in the context of a goal or objective. Full autonomy of a principal A principal has full autonomy or complete autonomy, the full freedom to formulate, choose, and pursue goals and objectives. An agent does not have such full autonomy. Limited autonomy or partial autonomy of agents Generally speaking, agents do not have autonomy in the same sense as the full autonomy of a principal, but agents do have limited autonomy or partial autonomy in the sense that they are free to choose what tasks to perform to achieve the goals or objectives chosen by the principal, and how to pursue those tasks. Assistants have no autonomy Unlike principals and agents, assistants have no autonomy whatsoever. They don’t get to choose anything. Their only job is to perform the tasks given to them by an agent or their principal. Okay, technically, assistants do have a modest degree of autonomy, but very modest and very minimal. Any system that doesn’t require a principal to be directly controlling every tiny movement by definition is delegating at least a small amount of autonomy. But not enough for the term autonomy to have any significant relevance to the freedom of action of such a system. That’s the point of distinguishing assistants from agents — to indicate the almost complete lack of autonomy. Assistants have responsibility but no authority A principal can delegate to both agents and assistants. Both will have responsibilities, but only agents have even a limited sense of authority, the authority to decide how to turn an objective or goal into specific tasks or actions. An assistant has no authority, simply the responsibility for a specified task or action, as specified, with little or no room for discretion or decision. Control A principal always has control over agents to which it has delegated responsibility for goals, and control over assistants to which it has assigned specific tasks. A principal could change or revise or even cancel goals, instructions which agents would be obligated to comply with. A principal can at any time request a status report on progress that an agent is making on a goal or objective. Robots Superficially, robots would seem to be fully autonomous, but in reality they have the more limited autonomy or partial autonomy of agents. After all, robots are owned and work on behalf of their owners, performing tasks and pursuing goals as their owners see fit, and dictate. That said, as with an agent, a robot can be granted a significant level of autonomy and be given fairly open-ended goals, so that they could actually be fairly autonomous even if not absolutely fully autonomous. Robots and computers out of control with full autonomy? That would make for a fairly scary science fiction story, a world in which robots and computers could be granted complete autonomy and not have to answer to anybody. But I wouldn’t expect that reality anytime soon. But it’s also possible that someone might mistakenly grant a robot complete autonomy and it might be difficult to regain control over the robot. Although, it would be possible to make it illegal to grant a robot full autonomy. The HAL computer in the 2001: A Space Odyssey movie and the Skynet AI network of computers and machines in the Terminator movies were in fact machines which somehow gained full autonomy — with quite scary consequences. It would be interesting to see a science fiction movie in which fully autonomous robots have a strictly benign and benevolent sense of autonomous responsibility. But maybe that violates the strict definition of autonomy — if they act as if to serve people, then they aren’t truly autonomous. Maybe robots would need to exist in colonies or countries or planets or space stations of their own, with full autonomy there, rather than coexisting within our human societies. Robot societies and human societies could coexist separately and could interact, but respecting the autonomy of each other, with neither in charge or dominating the other. Maybe. Mission and objectives A mission is a larger context than discrete goals. Think of the mission of an enterprise or organization. It’s purpose. It’s market or area of interest. The mission will break down into objectives, which will break down into discrete goals. The enterprise or organization may periodically review and adjust, revise, or even radically change its mission and objectives. At its own discretion. That’s autonomy. An agent is given a discrete goal to pursue. A small part of a larger mission and its objectives. An agent does indeed have a mission and objective, but they are set by its principal. An agent has no control over its mission or objective. A principal has a larger mission and associated objectives for which discrete goals are periodically identified and assigned to discrete agents. A principal sets its own mission and objectives. For more discussion of mission and objectives, see the companion paper, Intelligent Entities: Principals, Agents, and Assistants. Mission and operational autonomy There are two categorical distinctions concerning the autonomy of an entity: Mission autonomy. The entity can choose and control its own missions and objectives rather than be constrained to pursue and follow a mission or objective set for it by another entity, a principal. This is closer to true autonomy. Operational autonomy. The entity can decide for itself how to accomplish operational requirements. This is independent of control of the overall mission and objectives. This is characteristic of an agent, although an autonomous entity would tend to also have operational autonomy as well. So: Principals have mission autonomy. And generally operational autonomy as well. Agents have operational autonomy. But no mission autonomy. Independence — mission and operational Autonomy is roughly a direct synonym for independence. We can speak of two categorical distinctions concerning the independence of an entity: Mission independence. The entity can choose and control its own missions rather than be constrained to pursue and follow a mission set for it by another entity, a principal. This is closer to true autonomy. Operational independence. The entity can decide for itself how to accomplish operational requirements. This is independent of control of the overall mission. This is characteristic of an agent, although an autonomous entity would tend to also have operational independence as well. So: Principals have mission independence. And generally operational independence as well. Agents have operational independence. Luck and Mark d’Inverno: A Formal Framework for Agency and Autonomy Michael Luck and Mark d’Inverno published a paper back in 1995 entitled A Formal Framework for Agency and Autonomy which examined agency and autonomy as this paper does but focused strictly on software agents and multi-agent systems in particular: Abstract: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.51.4431 PDF: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.25.8941&rep=rep1&type=pdf The abstract: With the recent rapid growth of interest in MultiAgent Systems, both in artificial intelligence and software engineering, has come an associated difficulty concerning basic terms and concepts. In particular, the terms agency and autonomy are used with increasing frequency to denote different notions with different connotations. In this paper we lay the foundations for a principled theory of agency and autonomy, and specify the relationship between them. Using the Z specification language, we describe a three-tiered hierarchy comprising objects, agents and autonomous agents where agents are viewed as objects with goals, and autonomous agents are agents with motivations. The relevant phrases: agents are viewed as objects with goals autonomous agents are agents with motivations I cite this reference here neither to blindly accept it nor to quibble with it, but simply to provide a published foundation which readers can reference. That said, I’ll offer a couple of relatively minor quibbles, more along the lines of how to define terms: I’d prefer to use the term entity or even intelligent entity rather than object. To my mind, objects include trees, rocks, and mechanical machines, but generally not include peoples and animals per se. Technically, intelligent entities are indeed objects, but the term object doesn’t capture the essential meaning of an intelligent entity. The cited paper defines an object as an entity that comprises a set of actions and a set of attributes. This notion of having a set of actions or capable of acting is a bit more than the traditional, real-world, non-computer science sense of the meaning of the concept of an object. A machine is capable of acting in some sense, but unless it has some sort of robotic brain, it has no sense of sensing its environment and making decisions about how to interact with its environment. A sense of agency is needed. A washing machine or refrigerator would fit the meaning of an object in the sense of the cited paper, although I would refer to them as assistants rather than mere objects in a real-world sense. They have no agency, with no ability to choose how to pursue a goal rather than to blindly perform a specific task. That ability to perform tasks does fit the definition of an assistant used in this paper. A driverless car would be a good fit for what I would call an intelligent entity and would fit the concept of agent used in the cited paper. You tell the car where you want to go and it figures out the rest, coming up with a plan and figuring out what tasks are needed to get you to your objective. Said driverless car would superficially seem to have a sense of autonomy, in that it can move around without a person at the controls, but it lacks the ability to set its goals. It can pursue and follow goals given to it, but not set them. In that sense, both mine and the cited paper, said driverless car does not have autonomy. Driverless cars did not exist back in 1995, but I think even now the authors of the cited paper would likely agree that a driverless car lacks the motivation or ability to set goals that is required to meet the definition for autonomy. As the paper would seem to agree, goals are set from motivations. As the paper would seem to agree, being an agent does not automatically confer the presence of motivations. Agents don’t need to be motivated. They just need to be able to pursue and achieve goals. In the context of software agents, which was indeed the context of that 1995 paper, I’d refer to degree of autonomy, meaning the extent to which the agent is free to make its own choices, as opposed to the degree to which the agent’s principal has already made choices and has decided to constrain the choices or autonomy of the agent. An upcoming companion paper, Intelligent Entities: Principals, Agents, and Assistants, will explore this notion of principal with respect to autonomy. The cited paper uses the term motivation to essentially mean that the agent has the ability to set its own goals. I agree with the cited paper that agents are all about goals. The open issue is who sets the goals for a given agent. In my terms, it is the principal which sets goals. That could be a person, or some piece of software or even a robot. And this paper does allow for the prospect of subgoals so that an agent can act as principal for a subgoal. In the terms of the cited paper, an autonomous agent would correspond to my concept of principal. A key difference between the terminology of the cited paper and of this paper, is that this paper first seeks to ground the terms in the real world of human entities or people before extending the terms and concepts to the world of machines and software. Motivation Motivation is a greater factor in autonomy, but can be relevant to agency as well. A principal should clear have some good reason for its choices in setting objectives and goals. Its motivation. An agent might have some minor motivation for its choices as to what tasks to perform to pursue and achieve the goals given to the agent by its principal, but those minor motivations pale in significance to the larger motivation for why the goal should be pursued at all, something only the principal can know. The contract between principal and agent may likely express the motivation for each goal or objective, although that expression may have dubious value to the agent. One exception is when the specification for the objectives might be technically weak and too vague, incomplete, or ambiguous, leaving the agent with the job of deducing the full specification of objectives by parsing the motivation. That’s not the best approach, but may be the only viable or sane approach. Sociology and philosophy The concept of agency takes on a different meaning in sociology and philosophy — it is used as a synonym for what is defined as autonomy in this paper and in the context of robots, intelligent agents, and intelligent digital assistants. The relevant dictionary sense is: the capacity, condition, or state of acting or of exerting power With no mention of any principal or other external intelligent entity setting objectives for the agent to follow. That would be more compatible with the sense of principal as agent used in this paper, where the agent is indeed setting its own objectives and goals. That’s an unfortunate ambiguity, but that’s the nature of natural language. For more information on these usages, consult the Wikipedia: Wikipedia article on Agency (sociology). Wikipedia article on Agency (philosophy). Agent-based modeling (ABM) and agent-based simulation (ABS) One other field in which agency is defined as being synonymous with autonomy is agent-based modeling (ABM), also known as agent-based simulation (ABS), in which agents have a distinct sense of independence, autonomy. These agents are more like the principals defined in this paper. ABM/ABS is a hybrid field, a mix of computer science and social science, and not limited to computer science or social science, either. In fact, it can be applied to other fields as well. Anywhere that there are discrete, autonomous entities that interact and can have some sort of aggregated effect. ABM/ABS is more of a tool or method than a true field per se. For all intents and purposes, ABM/ABS could be considered part of social science and sociology. Definitions As promised, here are the synthesized definitions of autonomy and agency as used in this paper: Autonomy. Degree to which an intelligent entity can set goals, make decisions, and take actions without the approval of any other intelligent entity. The extent to which an entity is free to exert its own will, independent of other entities. Can range from the full autonomy of a principal to the limited autonomy or partial autonomy of an agent to no autonomy for an assistant. The entity can decide whether to take action itself or delegate responsibility for specific goals or specific tasks to other intelligent entities, such as agents and assistants. Agency. Ability of an intelligent entity, an agent, to plan, make decisions, and take actions or perform tasks in pursuit of objectives and goals provided by a principal. The agent has limited autonomy or partial autonomy to decide how to pursue objectives and goals specified by its principal. A contract between principal and agent specifies the objectives and goals to be pursued, authorizing action and obligations, but leaving it to the agent to decide how to plan, define, and perform tasks and actions. The agent may decompose given objectives and goals into subgoals which it can delegate to other agents for whom this agent is their principal. Note: In sociology and philosophy agency refers to autonomy or the extent to which an entity is free to exert its own will, independent of other entities. Some derived terms: Degree of autonomy. Quantification of how much autonomy an entity has. Limited autonomy. Partial autonomy. Some degree of autonomy short of full autonomy. Weak autonomy. Entity with limited autonomy, constrained by goals set by other entities. Roughly comparable to agency. Autonomous intelligent entity. Intelligent entity that has some degree of autonomy. Autonomous entity. Synonym for autonomous intelligent entity. Or any entity which acts autonomously, even if not strictly intelligent. Full autonomy. Complete autonomy. Absolute autonomy. True autonomy. Unlimited, unrestricted autonomy. No other entity is able to exert any meaningful control over such an autonomous entity. Mission autonomy. The entity can choose and control its own missions rather than be constrained to pursue and follow a mission set for it by another entity, a principal. This is closer to true autonomy. Operational autonomy. The entity can decide for itself how to accomplish operational requirements. This is independent of control of the overall mission. This is characteristic of an agent, although an autonomous entity would tend to also have operational autonomy as well. Limited agency. Some degree of agency short of full agency. Some degree of autonomy short of full autonomy. Full agency. Unlimited, unrestricted agency, limited only by the contract between the agent and its principal. Still only a limited degree of autonomy, constrained by its contract with its principal. Degree of agency. Quantification of how much agency an entity has. Agent. Any entity with some degree of agency, but lacking full autonomy. Autonomous agent. Improper term, in the view of this paper. An agent would, by definition, not be fully autonomous. Nonetheless, the term is somewhat commonly used in computer science to indicate an agent with a relatively high degree of autonomy. These definitions should apply equally well to human and computational entities, or at least be reasonably compatible between those two domains. Terms used within those definitions are defined elsewhere in this paper, including: Entity Intelligent entity Principal Agent Assistant Objective Goal Task Action Subgoal Responsibility Authority Delegation Contract Autonomous systems Generally and loosely speaking, people speak of autonomous systems, whether it be a robot, a software application, a satellite, a deep space probe, or a military weapon. This is not meant to imply that such systems are fully, completely, and absolutely autonomous, but simply that they have a high degree of autonomy. Or what we call limited autonomy or partial autonomy in this paper. And to draw a contrast to directly or remotely controlled systems such as drones where every tiny movement is controlled by a human operator. Lethal autonomous weapons (LAWs) A very special case is what is called a lethal autonomous weapon or LAW. These weapons are of significant ethical concern since they largely take human judgment, human discretion, and human compassion out of the equation. As noted for autonomous systems in general, even so-called lethal autonomous weapons will not typically be fully, completely, and absolutely autonomous. They may have a significantly higher degree of autonomy, but not true, full autonomy. There is some significant effort to assure that at least some minimal degree of human interaction occurs, what they call meaningful human control. That’s still a somewhat vague term, but the concept is still in the early stages. Even an automatic rifle or machine gun has a trigger, causing it to stop firing when a person decides to stop holding the trigger. That’s meaningful human control. Even before we start getting heavily into artificial intelligence (AI), there are already relatively autonomous systems such as the Phalanx CIWS close-in weapon system gun for defense against anti-ship missiles. It is fully automated, but with oversight by a human operator. It can automatically detect, track, and fire on incoming missiles, but the operator can still turn it off. A big ethical concern for lethal autonomous weapons is the question of accountability and responsibility. Who is responsible when an innocent is harmed by such an autonomous weapon when there is no person pulling the trigger? A practical, but still ethical, concern is the technical capability of discriminating between combatants and civilians. Granted, even people have difficulty discriminating sometimes. Technical capabilities are evolving. They may still be too primitive today by today’s standards, but further evolution is likely. In fact, there may come a day when autonomous systems can do a much better job of discrimination than human operators. The only truly fully autonomous lethal weapon I know of is the minefield. Granted it has no AI or even any digital automation, and the individual mines are not connected, but collectively it acts as a system and is fully, absolutely autonomous. It offers both the best and worst of military and ethical qualities. It has no discrimination. It is fully autonomous. It is quite reliable. It is quite lethal. It is quite humane. It has absolutely no compassion. It has no accountability. No responsibility. And no human operator can even turn it off other than by laboriously and dangerously dismantling the system one mine at a time. Somebody put the mines there, but who? Now, take that rather simple conception of a minefield and layer on robotics, digital automation, and even just a little AI, and then you have mountains of technical, logistical, and ethical issues. That’s when people start taking about killer robots and swarms. Sovereignty Another related term which gets used in some contexts as a rough synonym for both autonomy and independence is sovereignty. From the Merriam-Webster definition of sovereignty: freedom from external control One can refer to an entity as being sovereign if it is autonomous or independent. But generally, it won’t be necessary to refer to sovereignty rather than autonomy. Summary To recap: Autonomy refers to the freedom of an intelligent entity to set its own objectives and goals and pursue them, either by acting directly itself or delegating goals to agents. An autonomous intelligent entity (principal) controls its own destiny. Agency refers to the freedom of an intelligent entity (agent) to pursue goals delegated to it by its principal as it sees fit, although subject to expectations and obligations specified by its principal in the contract which governs their relationship. An agent owes its allegiance to its principal. Although in sociology, philosophy, and agent-based modeling and simulation the terms agency and agent are used and defined as the terms autonomy and autonomous entity are in this paper. One can also refer to degree of autonomy, so that an agent has some limited degree of autonomy and so-called autonomous systems have a fair degree of autonomy even though they do no have full, complete, and absolute autonomy. Lethal autonomous weapons? Coming, but not here yet, and not in the very near future. For more of my writings on artificial intelligence, see List of My Artificial Intelligence (AI) Papers.

What Are Autonomy and Agency?

12

what-are-autonomy-and-agency-1928813394c7

2018-09-02

2018-09-02 22:03:29

https://medium.com/s/story/what-are-autonomy-and-agency-1928813394c7

false

6,236

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Jack Krupansky

Freelance Consultant

6254b66338f2

jackkrupansky

51

1

20,181,104

null

0

null

0

null

2018-01-11

2018-01-11 13:09:16

2018-01-11

2018-01-11 13:20:58

2

false

en

2018-03-28

2018-03-28 18:24:13

2

1928aac482ab

2.745597

3

0

Marshall McLuhan once said, “Man becomes, as it were, the sex organs of the machine world, as the bee of the plant world, enabling it to…

3

Making Music With Watson Beat Marshall McLuhan once said, “Man becomes, as it were, the sex organs of the machine world, as the bee of the plant world, enabling it to fecundate and to evolve ever new forms. The machine world reciprocates man’s love by expediting his wishes and desires, namely, in providing him with wealth.” I have recently been afforded that wealth in the form of inspiration. On January 8th IBM released Watson Beat, their cognitive music platform on GitHub. What follows was originally posted on T3, and is the summary of my first hours with Watson Beat — my new muse. Have you ever found yourself in a creative rut? As a musician, this happens to me all the time. After IBM released the Watson Beat code to the public on GitHub yesterday, I have a new muse to turn to for inspiration. This Artificial Intelligence (AI) tool is designed to compose music. Traditionally, people think of AI as simply data and analytics. But Watson Beat leverages AI to turn data into a creative expression. It’s like nothing you’ve ever seen (or heard) before. End composition visualized. A new evolution in music composition Watson Beat is IBM’s music arm of Watson. Its AI-based software understands how to make music based on two filters: emotion and style. But it’s smart enough to understand the entire composition of a music file — from the melody to the chorus to the time signatures. Watson Beat’s machine learning technology takes those inputs and completely rearranges them, giving you a fresh way to look at the same chords. Take a song like “Mary had a Little Lamb” and assign it the emotion of romantic and the style of reggae. Watson Beat will take those parameters and return a track that completely reimagines this well-known nursery rhyme. The T3 Innovation team decided to test out this new tool and see what we could create in a matter of hours. Crafting a track in less than two hours When IBM released the Watson Beat code our team wasted no time getting to work. We gave Watson Beat a MIDI file comprised of three instruments (vibraphone, synthetic strings, and synthetic bass), four bars and 60 beats per minute. In Logic, it looked like this: Random 4 bar MIDI input to Watson Beat In less than an hour and a half, we produced a 1:40 second track based entirely on that original MIDI file and Watson Beat’s deconstruction and ultimate reconstruction of it. The result sounds nothing like the original input. It shows how Watson’s AI filter can completely reimagine the version of what we provided. Watson Beat Output plus percussion tracks. As a musician, this was mind blowing. Watson was able to give me a composition faster than I could have ever gotten it out of my head. And that’s just scratching the surface of what Watson Beat can do. Based on the guide rails you put in place, Watson Beat will continue to spit out composition after composition, giving musicians endless inspiration and an entirely new way to look at the creative process. Using AI to amplify inspiration So why is all this important? Aside from the obvious benefits of being able to crank out tunes fast, Watson Beats is another example of why we shouldn’t be afraid of AI. Many people see AI as the replacement of humans. But at T3, we think AI is about the augmentation of humans. Through Watson Beat, I was able to amplify my natural human abilities to take my music further than I thought possible. Through this marriage of technology and humans, we can get to the next evolution in music, inspiration, and who knows what else. To be continued… Inspiration and composition by IBM Watson Beat. Produced by Brandon Gredler. Engineered by Joshua Brewer. Percussion by Austin Hegarty.

Making Music With Watson Beat

25

making-music-with-watson-beat-1928aac482ab

2018-03-28

2018-03-28 18:24:14

https://medium.com/s/story/making-music-with-watson-beat-1928aac482ab

false

626

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

BG

Director of Product Design @T3thinktank

908c19c32e5c

BrandonGredler

223

684

20,181,104

null

0

null

0

null

2017-11-14

2017-11-14 08:45:03

2017-11-15

2017-11-15 08:14:35

0

false

en

2017-11-15

2017-11-15 08:14:35

0

1929410eec6c

0.660377

3

0

Well, it’s been a month since our Tokensale finished and I am anxious to deliver our latest news and inform our community about Anryze…

5

Anryze Distributed Speech Recognition Platform. One month after Tokensale ends Well, it’s been a month since our Tokensale finished and I am anxious to deliver our latest news and inform our community about Anryze progress. So here it is. We have tested our neural network on different businesses and have found out that it needs to be educated on several more datasets, so we did purchase several and collected some on our own. Those datasets are focused on sales, medicine, legal, construction, etc. For NN improvement we hired two more data scientists and two more developers. We have created a department for developing Distributed Miners Backend and attracted several consultants for this process. Expected alfa testing will start on December 12th. Anryze platform website will be ready and prepared for personal use (our B2C direction) at the beginning of December. That means that you will be able to upload audio/video files and receive the text from them in one click. For businesses, it will be ready, as promised, at the beginning of the next year.

Anryze Distributed Speech Recognition Platform. One month after Tokensale ends

9

anryze-distributed-speech-recognition-platform-one-month-after-tokensale-ends-1929410eec6c

2018-06-14

2018-06-14 05:06:11

https://medium.com/s/story/anryze-distributed-speech-recognition-platform-one-month-after-tokensale-ends-1929410eec6c

false

175

null

Machine Learning

machine-learning

Machine Learning

51,320

Anton Gera

Founder and CEO at @Anryze

89d0fd2de2d0

anton.gera

600

611

20,181,104

null

0

null

0

null

2018-08-25

2018-08-25 16:41:22

2018-08-25

2018-08-25 16:41:36

0

false

en

2018-08-25

2018-08-25 16:41:36

1

1929ab7d8c68

2.067925

0

[Download] [PDF] Life 3.0: Being Human in the Age of Artificial Intelligence EPUB By Max Tegmark Link…

1

Read Online Life 3.0: Being Human in the Age of Artificial Intelligence By Max Tegmark Full PDF #EPUB [Download] [PDF] Life 3.0: Being Human in the Age of Artificial Intelligence EPUB By Max Tegmark Link https://reviewskindlenew.icu/?q=Life+3.0%3A+Being+Human+in+the+Age+of+Artificial+Intelligence . . . . . . . . . . . . . . . . . . . Read Online PDF Life 3.0: Being Human in the Age of Artificial Intelligence, Download PDF Life 3.0: Being Human in the Age of Artificial Intelligence, Download Full PDF Life 3.0: Being Human in the Age of Artificial Intelligence, Download PDF and EPUB Life 3.0: Being Human in the Age of Artificial Intelligence, Read PDF ePub Mobi Life 3.0: Being Human in the Age of Artificial Intelligence, Reading PDF Life 3.0: Being Human in the Age of Artificial Intelligence, Read Book PDF Life 3.0: Being Human in the Age of Artificial Intelligence, Read online Life 3.0: Being Human in the Age of Artificial Intelligence, Download Life 3.0: Being Human in the Age of Artificial Intelligence Max Tegmark pdf, Download Max Tegmark epub Life 3.0: Being Human in the Age of Artificial Intelligence, Read pdf Max Tegmark Life 3.0: Being Human in the Age of Artificial Intelligence, Download Max Tegmark ebook Life 3.0: Being Human in the Age of Artificial Intelligence, Read pdf Life 3.0: Being Human in the Age of Artificial Intelligence, Life 3.0: Being Human in the Age of Artificial Intelligence Online Download Best Book Online Life 3.0: Being Human in the Age of Artificial Intelligence, Read Online Life 3.0: Being Human in the Age of Artificial Intelligence Book, Read Online Life 3.0: Being Human in the Age of Artificial Intelligence E-Books, Read Life 3.0: Being Human in the Age of Artificial Intelligence Online, Read Best Book Life 3.0: Being Human in the Age of Artificial Intelligence Online, Read Life 3.0: Being Human in the Age of Artificial Intelligence Books Online Download Life 3.0: Being Human in the Age of Artificial Intelligence Full Collection, Download Life 3.0: Being Human in the Age of Artificial Intelligence Book, Read Life 3.0: Being Human in the Age of Artificial Intelligence Ebook Life 3.0: Being Human in the Age of Artificial Intelligence PDF Read online, Life 3.0: Being Human in the Age of Artificial Intelligence pdf Download online, Life 3.0: Being Human in the Age of Artificial Intelligence Read, Download Life 3.0: Being Human in the Age of Artificial Intelligence Full PDF, Read Life 3.0: Being Human in the Age of Artificial Intelligence PDF Online, Read Life 3.0: Being Human in the Age of Artificial Intelligence Books Online, Read Life 3.0: Being Human in the Age of Artificial Intelligence Full Popular PDF, PDF Life 3.0: Being Human in the Age of Artificial Intelligence Read Book PDF Life 3.0: Being Human in the Age of Artificial Intelligence, Read online PDF Life 3.0: Being Human in the Age of Artificial Intelligence, Download Best Book Life 3.0: Being Human in the Age of Artificial Intelligence, Read PDF Life 3.0: Being Human in the Age of Artificial Intelligence Collection, Read PDF Life 3.0: Being Human in the Age of Artificial Intelligence Full Online, Read Best Book Online Life 3.0: Being Human in the Age of Artificial Intelligence, Download Life 3.0: Being Human in the Age of Artificial Intelligence PDF files

Read Online Life 3.0:

0

read-online-life-3-0-1929ab7d8c68

2018-08-25

2018-08-25 16:41:36

https://medium.com/s/story/read-online-life-3-0-1929ab7d8c68

false

548

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

sdfjgfhdg

null

5a50624b8caa

sdfjgfhdg

0

1

20,181,104

null

0

null

0

null

2018-05-15

2018-05-15 01:26:19

2018-05-15

2018-05-15 03:37:16

10

false

en

2018-05-15

2018-05-15 03:37:16

27

192b1b6417f1

7.993396

8

0

Dear DeepBrain Chain Community Members:

3

DeepBrain Chain Weekly Report #11 Dear DeepBrain Chain Community Members: Thanks for your unwavering support! DeepBrain Chain will continue to share the latest progress in tech development, marketing and community building. Below is a recap of what we have achieved last week: 1. Development Last week, our dev team were hard at work on reviewing & optimizing code of iteration 1 for AI training task & problem shooting and solving. We also reviewed the bottom code of iteration 0 one more time. Framework Design We assessed several alternative concepts for data encryption security and will conduct further analysis this week; We analyzed several network threat models and network security concepts under the environment of blockchain and will finish first draft by mid May; Testing We finished the third round of testing for iteration 0 with focus on abnormal performance testing, memory resources leakage testing and abnormal pressure testing; We are able to support more than 110,000 concurrent connections on a single PC/windows platform; We also conduced the first round of testing for iteration 1; Development We corrected the problems found in the third round of testing for iteration 0; We optimized log exception catching when handler resources exceed system limits; We optimized hangup message catching; We solved compatibility issue when docker restful api upgrades; We solved the problem of downloading blockage for decentralized storage data; We optimized code for data synchronization among nodes; We optimized log prompt message during node initialization; We optimized the compiler script for MAC platform and separate platform and business level into different libraries; We built seed nodes for the testnet. 2. Marketing North America Last week, DeepBrain Chain’s EVP of Silicon Valley AI+Blockchain Center and Chief AI Officer, Dr. Dongyan Wang attended several conferences in North America, DeepBrain Chain was the talk of the space. Dr. Dongyan Wang Gave Keynote Speech at GDIS Chief AI Officer, Dr. Dongyan Wang attended several conferences in North America, DeepBrain Chain was the talk of the space. On 10th of May, Silicon Valley Business Institute published the video of Dr. Dongyan Wang’s speech at Global Disruptive Innovation Summit. In the 20 minutes keynote speech, Dr. Wang went from his experience at Cisco to the two AI winters he endured and the challenges to come. He believes that start-ups and companies without raw data face extremely high cost if they want to develop an AI business; and in blockchain industry, mainstream projects based on POW consensus mechanism such as Bitcoin have caused a huge waste of resource and computing power, and have been heavily criticized for it. DeepBrain Chain’s innovation on combining AI and blockchain focuses on solving the issues of high AI computing cost and the resource wastage in blockchain. Here is the Video. South East Asia DeepBrain Chain At Indonesia’s Blockchain Conference Block Jakarta, 2018. BlockJakarta Conference Block Jakarta 2018: Indonesia Blockchain Conference took place on the 9th of May at Ritz Carlton hotel in Jakarta. Eric Yin, Marketing Director of DeepBrain chain, was invited to attend the conference on behalf of DBC as the keynote guest to introduce and analyze the existing relationship between AI and block chain and industry pain points.The conference gathered well-known business leaders and the Indonesia’s biggest local exchange, INDODAX. The conference was also co-organized by Indonesian government departments. The conference focused on blockchain technology development and the government’s regulation on the technology. At the meeting, the practical commercial and industrial value of the participating enterprises and the most concerned DBC AI x Blockchain model in various fields were explained in detail after the combination of the traditional industries. 3. Ecosystem OneGame, a project based on DeepBrain Chain’s AI public chain has been launched. OneGame, a project based on DeepBrain Chain’s AI public chain has been launched. OneGame — — a decentralized virtual world built on DeepBrain Chain’s AI public chain has launched officially. The powerful platform of deep brain chain provides OneGame with the necessary computing resources and the support of artificial intelligence algorithms and data. Jinse Finance and Tuoniao Blockchain Media reported the news. OneGame helps users to generate new models and scenarios through in-depth learning; by enhancing learning, constantly training and enhancing the game intelligence of non-player characters on the platform; through genetic algorithms, applying gene arrangement and combination, let the platform constantly evolve itself. Artificial intelligence algorithm is the core competitiveness of OneGame platform. In this regard, they have worked with teams in the DeepBrain Chain in a number of ways. 4. Media Reports YAHOO Finance put spotlight on DeepBrain Chain’s ‘AI’+‘Blockchain’ platform. Last week, the renown YAHOO Finance published an article titled “Deepbrain Chain, the first artificial intelligence computing platform driven by blockchain ‘’. The article described DeepBrain Chain beingthe first AI computing platform powered by blockchain, and stated that this is a significant innovation on coming AI and cryptocurrency. Article link: https://finance.yahoo.com/news/deepbrain-chain-first-artificial-intelligence-230900369.html?soc_src=social-sh&soc_trk=tw Inc, Coin Journal and other media’s reports Inc.com wrote in their article “A.I. Is Awesome, Blockchain Is a Powerhouse”, and quoted Dr. Wang in his views on AI development. They also expressed the view that ‘’ DeepBrain Chain as the front runner on AI+Blockchain is setting a new milestone’’. A.I. Is Awesome, Blockchain Is a Powerhouse. But Here’s What Combining Them Could Do You know artificial intelligence from your Alexa, Tesla or Netflix account. And now Blockchain is generating serious…www.inc.com Well known American media Coin Journal published an article titled “DeepBrain Chain To Launch AI, Blockchain Research Center In Silicon Valley”, quoting Dr. Wang on how the AI+blockchain ecosystem will link computing power, AI models and data across the globe to greatly lower the threshold to enter AI industry. DeepBrain Chain To Launch AI, Blockchain Research Center In Silicon Valley The DeepBrain Chain Foundation, the organization overseeing the DeepBrain Chain artificial intelligence (AI) computing…coinjournal.net Bitswin and other well known media also reported on us. Bitswin published in-depth analysis on DeepBrain Chain This is a comprehensive analysis on the whole project, the video is accompanied with PPT, along with subtitles. The video starts with explaining what AI is, and the obstacles AI industry is facing, then moves on to how blockchain could solve these issues. The author then introduce DeepBrain Chain’s team background, mining mechanism, exchanges we are on, our blockchain bottom layer structure , throughput amount, consensus mechanism and community development. English Video :https://drive.google.com/open?id=153YX8mDUQyQzpPGB67ZJUg3FEOCdCm9s Chinese Video :https://mp.weixin.qq.com/s/f4T29_OhvN3AxfVYWWs8Zw More than ten Chinese media has reported on DeepBrain Chain’s AI mining machine DeepBrain Chain’s AI mining machine news release shocked the space and more than ten Chinese media including Soho, Interface, Jinse Finance, Gongxiang Finance, Geek Park and Yesky has published the news. 5. KOL focus Last week, DeepBrain Chain CEO, Yong He, and Director of the Silicon Valley AI+Blockchain Center, Dr. Wang Dongyan, was interviewed by a number of Internet and blockchain industry KOL. On May 9th, CEO Yong He presented the plan of the DeepBrain Chain AI miner to the community in a live interview with Tony Tian on YouTube. DeepBrain Chain AI miner is an important part of the ecosystem. It can translate computing power into means for highly efficient AI deep learning and blockchain calculation mining. For participants, entering the main network and “mining” through AI mining machines, AI deep learning and machine learning training can be carried out at considerably small cost, and earn DBC as rewards; for DeepBrain Chain ecosystem, the AI computing power dispersed in the blockchain network can not only help AI enterprises save computational power cost, but also increase the DeepBrain Chain ecosystem’s value with the increase of nodes, so that all the participants in the ecosystem can share AI resources. The founder of DeepBrain Chain and CEO He Yong’s interview with Tony Tian on YouTube On May 11, Dr. Wang was interviewed by YouTube celebrity Crypto Beadles. Dr. Wang said that he would lead the team in the following studies: parallel training of large-scale neural networks, efficient joint learning of deep neural networks, energy consumption reduction of DeepBrain Chain network through reinforcement learning, integration of AI and blockchains and distributed killer AI applications. The interview had more than 133,000 views and more than 3,500 likes in one day. With further promotion of the global community, the market shows great enthusiasm for DeepBrain Chain, and is highly praised by the blockchain and cryptocurrency market. Dr. Wang is interviewed by YouTube celebrity Brad In an interview with Brad last week, Dr. Dongyan Wang introduced the feasibility of DeepBrain Chain’s platform landing, indicating that the development of DeepBrain Chain ecosystem will be promoted through AI hardware / AI mine machine), AI software and AI application. 6. Community Happy Mother’s Day On the eve of Mother’s Day, DeepBrain Chain sent a thank you note to community volunteers expressing sincere gratitude. The rapid development of DeepBrain Chain cannot be separated from the firm support of the community and from the dedication of early volunteers. With our passionate and creative communities by our side, we can face the upcoming challenges with confidence. Current status of our communities: Telegram (English Official Group): 12,183; Telegram (Korean Official Group): 683; Telegram (Indonesia Official Group): 1,744; Telegram (Vietnam Official Group): 1,072; Telegram (Thailand Official Group): 2,147; Twitter: 32,670; Reddit community: 7,971; Facebook Page: 508; We have set up Telegram AI miner group @DeepBrainChainAIminers. We look forward to seeing you in the group! 7. Talent Recruitment Our team members with more than 15 years of experience in R & D, from BAT, Huawei, NetEase, Mobile, Ericsson and other companies joined us and we successfully completed the overall architecture design of iteration 1. We are still recruiting blockchain development engineers, smart contract engineers, virtual machine development engineers, security engineers (penetration), distributed storage development engineers and Java development engineers. We welcome external referrals. Successful referrals will be rewarded with an equivalent of RMB 20,000 in token form. Contact: erin@deepbrainchain.org. Please indicate your resume recommendation and attach the referee’s contact information, so that we can give out the bounty in the future. Chinese H5:http://u6716916.viewer.maka.im/k/RYFDRGSY?from=singlemessage&isappinstalled=0 8. Things to Come Russian Blockchain Week From May 21 to May 25, DeepBrain Chain team will participate in the Russian blockchain week. More than 1,500 industry guests will take part in the five-day event, with 70+ blockchain industry experts bringing cutting-edge presentations. Six in-depth thematic events will tap global trends. DeepBrain Chain will give a keynote speech at the event. http://blockchainweek.moscow/eng/ Amsterdam Blockchain Expo, Netherlands From June 27 to June 28, Dr. Wang and DeepBrain Chain team will attend the Amsterdam Block Chain Expo in the Netherlands and actively seek cooperation with AI enterprises, universities and research institutions. Blockchain Conference & Exhibition Event | Blockchain Expo Europe Blockchain Expo Europe — 27–28 June — RAI, Amsterdam. Blockchain Conference & Exhibition exploring blockchain and…blockchain-expo.com Exchanges we’re currently on: Huobi.pro ,Kucoin ,LBank,gate.io, Allcoin, Bitbns, Switcheo Wallets: NEO-GUI、NEON、NEOtracker 、 O3、Hyper Pay 、ROOTOKEN、Morpheus If you have any suggestions or feedback, feel free to contactinfo@deepbrainchain.org, or DeepBrain Chain Twitter. Learn more about us: Go to Official Website here Download the White Paper here

DeepBrain Chain Weekly Report #11

67

deepbrain-chain-weekly-report-11-192b1b6417f1

2018-06-03

2018-06-03 20:43:46

https://medium.com/s/story/deepbrain-chain-weekly-report-11-192b1b6417f1

false

1,787

null

Blockchain

blockchain

Blockchain

265,164

DeepBrain Chain

AI Computing Platform Driven By Blockchain

379a9e7edef2

DeepBrain_Chain

960

4

20,181,104

null

0

git clone https://github.com/bloomsburyai/cape-slack-tutorial.git pip3 install -r requirements.txt ./get-id.py your_slack_key you_bot_name

3

ed77e237c4b9

2018-01-09

2018-01-09 16:44:52

2018-01-15

2018-01-15 11:50:05

8

false

en

2018-01-15

2018-01-15 11:50:05

22

192bd0d75814

6.027673

12

0

Create a Slack bot using Bloomsbury AI’s machine reading API

5

The smartest member of my team is a Slack bot (or, how to build an AI powered assistant) Cape is an AI that reads documents and then answers questions about them. It does this by finding the bit of text in a document that answers your question. It allows you to leverage state-of-the-art tech to build your own AI with a few lines of code. In this tutorial, we’ll show you how you can build an AI powered slack bot in 10 minutes. Join our public Slack channel and get to know our own bot, capebot, here. The above video shows two of the bots we use internally at Bloomsbury AI. The first, capebot, makes use of the saved reply feature to answer questions about a wide range of topics, e.g. What’s the office WiFi password? Where are certain documents kept? What are our best practices? It uses a sentence similarity model to recognise questions however they’ve been phrased (for example, “What’s our website?” vs “Where’s our web page?”) The second, pythonbot, answers questions about our python API. It does this using Cape’s machine reading feature to automatically extract answers from our API docs. Both bots use exactly the same code, they’ve just had different information uploaded into Cape. Getting Set Up To follow this tutorial you will need: Git Python 3.5 or later First clone our tutorial repository. This contains example code at various stages of development: Inside the new cape-slack-tutorial directory you’ll find the following files: step-1.py The skeleton of a Slack bot which just responds with “Hello!” whenever a user sends it a message. step-2.py A bot capable of answering questions about documents that have been uploaded through the Cape admin interface. step-3.py A bot with additional functionality to allow users to add saved replies from directly within Slack. get-id.py A helper script for finding out your bot’s Slack ID. requirements.txt A list of all the python module dependencies required to complete the tutorial. To install the python dependencies first run: This will fetch the slackclient andcape-client modules. Step 1: A Basic Slack Bot To begin with we’ll create a simple slack bot that just responds with “Hello!” whenever a user sends it a message. First visit the Slack bot page and click the “Install” or “Add configuration” button to create a new bot and add it to your workspace. If you don’t see the button you’ll either need to ask permission from your slack workspace owner or simply create your own workspace. You’ll then be presented with your Slack bot’s API key: You’ll also need to discover your Slack bot’s ID. Included in the tutorial repository is a script called get-id.py which can help with this. Simply run: This will output your bot’s ID, which will look something like: U123A4BCD Next edit step-1.py to include your details: Whenever someone sends a message that includes our bot’s username the handle_question function will be run: Which currently responds to all messages by sending the string “Hello!” to the channel it was contacted on. In the next section we’ll modify this function to send the user’s message to Cape and get a more intelligent response. We’ve only covered the bare essentials required for creating a basic Slack bot here, for a more in depth tutorial focusing purely on creating Slack bots check out Matt Makai’s “How to Build Your First Slack Bot with Python.” Step 2: Answering Questions Training our bot Before our bot can start replying to users we need to feed it some useful information. We can do this through the Cape admin dashboard, if you don’t already have a Cape account you can sign up for free. You can then train the bot in two different ways: Documents — After uploading documents Cape will use machine reading to find answers to a user’s question within them. Saved replies — These are simple pairs of questions and answers. Cape uses a sentence similarity model to match user questions with the saved questions and then provide the corresponding answer. You can make use of both documents and saved replies within the same bot. If a saved reply already exists which matches the user’s question then Cape will respond with this, otherwise it will read your documents and attempt to find the answer within them. Adding our Cape API key To retrieve your API key visit the Cape dashboard and click the settings icon in the top right: Settings This will then reveal a drop down showing your API key: Your user token Copy this and add it to step-2.py alongside your Slack settings: Responding to users Now when we receive a question from a user we can pass it along to Cape using CapeClient’s answer method: This returns a list of answers ordered by confidence, we then send the user the text of the first answer that we found. If Cape doesn’t find an answer above our requested threshold we let the user know that we weren’t able to answer the question. You can set the confidence threshold for your bot in the Cape dashboard settings. Now when we talk to the bot we get a response back based on any saved replies we entered or documents we uploaded: Step 3: Allowing users to teach the bot Adding our admin token To allow our bot to do more than just answer questions we need to provide it with access to our admin token. The admin token gives the client access to API methods which can make modifications to our bot, such as adding saved replies and uploading documents. You’ll find your admin token in the settings panel just underneath your question answering token: Your admin token This can then be added to the settings at the top of step-3.py: Handling user input We then modify the client so that if a message starts with .add-saved-reply we call the add_saved_reply method, otherwise we treat it as a question and answer it normally: We can then make use of the add_saved_reply method in CapeClient to create a new saved reply based on the user’s input: If we encounter any exceptions from Cape (e.g. if the supplied question already exists) we report this back to the user via a Slack message. Now when this question (or a similar phrasing of it) is asked our bot responds with the answer we supplied: Next Steps For a more advanced bot check out our cape-slack repository, this is the Slack bot we use ourselves at Bloomsbury AI. In addition to the features in this tutorial it can also: Run multiple bots simultaneously. Allow users to add paraphrases to questions. Provide multiple possible answers to a question. Explain why it has answered a question the way it has. Show the context surrounding an answer in a document. Wrap Up In this tutorial we showed you how simple it is to use Cape to create your own AI Slack bot. You could also use Cape to: Build a super-powered ctrl+f that finds the answer to questions like ‘Who is the CFO?’, rather than just all the occurrences of a keyword. See our tutorial and demo. Build an expressive query tool for textual data — e.g. find the number of product reviews where a customer mentions they made a complaint. Build an add-on to your private search that mimics Google’s “direct answers”. This is our mission at Bloomsbury AI (the company behind Cape) — to make the expertise stored in documents and in people’s heads as accessible as possible. We’re constantly trying to improve our documentation and support developers using our AI so please reach out with any feedback!

The smartest member of my team is a Slack bot (or, how to build an AI powered assistant)

204

the-smartest-member-of-my-team-is-a-slack-bot-or-how-to-build-an-ai-powered-assistant-192bd0d75814

2018-04-03

2018-04-03 14:22:19

https://medium.com/s/story/the-smartest-member-of-my-team-is-a-slack-bot-or-how-to-build-an-ai-powered-assistant-192bd0d75814

false

1,297

The Blog Of Bloomsbury AI

null

Bloomsbury AI

guillaume@bloomsbury.ai

bloomsbury-ai

null

Slack

slack

Slack

5,843

Mike Sheldon

Software developer at Bloomsbury AI

3e1d03b68b54

mike_3449

7

20,181,104

null

0

null

0

null

2018-01-05

2018-01-05 17:13:43

2018-01-08

2018-01-08 20:20:30

3

false

en

2018-01-08

2018-01-08 20:20:30

16

192e8ba73346

5.674528

4

0

Starting this week, the newest gadgets will glisten in all their glory at CES in Las Vegas. While the latest curved TV is beautiful, the…

5

New Year, New Thinking: Seven strategic considerations for electronics executives Starting this week, the newest gadgets will glisten in all their glory at CES in Las Vegas. While the latest curved TV is beautiful, the IBM Institute for Business Value would like to focus on strategies that can help your business before your feed is filled with shiny objects. This year, IBM’s industry leading thought leadership organization, the Institute for Business Value has a packed research agenda for electronics executives and analysts. We’re kicking the year off with two new reports: “Navigating the Cloud Continuum,” which focuses on hybrid cloud for innovation, and “Data by Design,” which takes a look at IBM’s Digital ReinventionTM method in action. We’re leveraging one of the richest current global research data sets in electronics, developed as part of IBM’s long-running global C-Suite study, from which the datapoints below are all drawn. We have cross industry and economic data as well as critical insights on IoT, AI and Blockchain. It’s a data geeks dream and I can tell you, we are just getting started. From our research, we’ve identified seven initial insights electronics executives should consider to help increase organizational intelligence, innovation, relevance and resilience: Cloud is no longer an IT thing. Cloud is an innovation thing. Organizations need to adopt different metrics to measure and capitalize on its power. While 56 percent of electronics executives plan to continue investments in cloud computing, advanced companies will be navigating the cloud continuum to deliver innovation. Using cloud to spread the workload or speed of new product development or reinventing the supply chain with external data, a new approach will focus on business value; data combined from public and private data sets; speed to answer and use that that data; and security. Our new report presents eight use cases that thoughtfully weigh business objectives to achieve innovation. Merging Internet of Things (IoT) and artificial intelligence (AI) technologies is a fast path to value. Starting with two or three use cases and a firm strategy is critical. While the rush to install IoT sensors has yielded success, there’s so much more you could know and can do. In electronics, an amazing 90 percent of respondents are experimenting, piloting, implementing or have established an IoT beachhead in the market. A set of early adopters has found success because of their strong ties to strategy. Overall, 42 percent of electronics respondents said they expect the combination of IoT and AI to transform their businesses. Using IoT data for base use cases — such as detect and repair — is a great starting point. However, combining IoT with multiple external data sources leveraging AI has more significant value. Mapping device locations to enable better technician routing, understanding water quality or power supply considerations, and assessing geo-based impacts or temperature considerations leads to better, more rapid decisions. It leads to better warranty management and product decisions as well. Electronics executives say they expect to have more ecosystem partners and better sharing, yet we’re still saying innovation is coming from inside the organization. That thinking has to change. To talk about convergence in real terms, let’s consider these five data points: Of the electronics executives we surveyed, 77 percent said they expect to defend their current market position through price and cost reduction. 73 percent said that industry incumbents are deploying innovative technologies, processes or business models. 71 percent said they expect to expand ecosystem partners and engage in better sharing. 73 percent also said convergence (the merging of industries, as suppliers, distributors, customers and competitors increasingly cooperate in ecosystems created to deliver new products and services) was the greatest trend they see driving the business. Yet 50 percent said they expect their innovation to come more from internal sources than form external ones. Does it strike anyone else that these illustrate conflicting purposes? Companies are fully capable of Digital ReinventionTM and greater reliance on partners to do what they do best as a faster path to success. Yet some companies rely on price- and cost-reduction for market manipulation. That’s a race to the bottom. No matter how effective your supply chain, neither price- nor cost-reduction builds the trust or loyalty customers require in an increasingly digital world. Digital ReinventionTM is not a feature set, it’s a methodology and a mindset. Digital innovators are identifying opportunities in traditional and emerging electronics markets, where they successfully blend equipment, software, services and content into powerful packages. It is a re-imagining of how business is done. By design, it centers on outcomes rather than inputs and starts with the future rather than the present or past. It helps electronics businesses re-conceive how and with whom they operate and how they engage with their environment. And it helps them better understand the imperatives of their consumers, customers and business partners — which helps them anticipate rather than respond to needs. Shifting away from a product-centered focus to an experience-centered one is key. We have seven examples in a new report, with more to come. Adaptability is now a mission-critical skill, and it requires a platform. Electronics businesses can’t depend on one product or even multiple products capturing sufficient market share to drive their aggressive growth targets. That’s what makes adaptability and resilience so valuable, along with innovation and intelligence. Platforms enable intelligence accessible to the organization — going well beyond mobility to increase collaboration. That support massively increases connections and networking of data and people. In this area, AI is simply the only way to bring together the disparate but important facts an electronics company will need to scale up or down in the right areas — whether that’s capacity, product, resources or capital. You can’t be fast or fluid enough any other way. The platform should be all about data. The data economy is not “the new oil,” it’s the means to change the scope and scale of your business. The primary function of a platform should be to enable Digital ReinventionTM. This can allow companies to focus on the creation of a data economy to serve customers and ecosystem partners. It accords data a role central to every part of the business, allowing organizations to change both scope and scale quickly. You will need to consider three core sets of data: people data, asset data and environment data. This data should come from inside and outside your organization. That is why Digital ReinventionTM focuses on outcomes. Solving your biggest challenges and creating your biggest opportunities requires context from the world at large. There’s a connection between devices, privacy and consumer expectations that sets up better ways to win. This adds up to an ocean of data — from devices and actions to consumers and processes. All of those have security and privacy implications that are growing increasingly complex. The sharing population has distinct expectations around trust, transparency and value. According to new IBV research on data-sharing, 72 percent of respondents said they were sensitive to the way electronics companies handle their information and 80 percent indicated privacy was important compared to other topics. And of those most willing to share their information, 79 percent said the data they provide belongs to them. The most advanced organizations understand this. When we asked electronics executives about the importance of data security and privacy of a customer’s private data, 72 percent said it was mission critical. We can expect customers are going to ask them to prove it. Want to learn more? IBM’s bench strength, solutions portfolio and industry experts are ready in Las Vegas this week for you to come talk to us. Drop in and say hi. We have plenty to talk about. IBM’s bench strength, solutions portfolio and industry experts can help your strategic imperative gain significant momentum. No matter what your role in the organization, we have award-winning critical thinking to push your initiatives forward. Some of my recent thought leadership in electronics: Hybrid Cloud Digital Reinvention Cognitive Manufacturing Device Security Warranty Transformation Cognitive Manufacturing; China version Blockchain Product Design with Data in Mind IBM and SAP, Digital and Cognitive

New Year, New Thinking: Seven strategic considerations for electronics executives

58

new-year-new-thinking-seven-strategic-considerations-for-electronics-executives-192e8ba73346

2018-02-07

2018-02-07 04:20:56

https://medium.com/s/story/new-year-new-thinking-seven-strategic-considerations-for-electronics-executives-192e8ba73346

false

1,358

null

Cloud Computing

cloud-computing

Cloud Computing

22,811

cristene j g-w

I work at intersect of electronics, healthcare, retail, design, innovation and marketing. Cristene Gonzalez-Wertz. Work for @IBM but comments/posts are my own.

c33f97a3b8dc

hermione1

491

656

20,181,104

null

0

null

0

a9b1b43d23ec

2018-09-29

2018-09-29 02:35:05

2018-09-29

2018-09-29 12:29:58

2

false

en

2018-09-29

2018-09-29 12:31:06

12

192e9f98f22e

1.345597

2

0

Naaut is returning to Bremen, Germany for International Astronautical Congress 2018. The main theme couldn’t be more perfect, Involving…

5

Naaut is coming to IAC2018 Naaut is returning to Bremen, Germany for International Astronautical Congress 2018. The main theme couldn’t be more perfect, Involving Everyone. It is a pivotal time for the Space sector as public interest is mounting, technology is advancing at a faster pace than ever before and the enterprising spirit of a new generation of aspirational entrepreneurs is starting to build critical mass. As we converge in Bremen next week Naaut is keen to meet entrepreneurs and organisations who are applying or considering the use of Artificial Intelligence, Machine Learning, Blockchain, Nanotechnology, Quantum Computing technologies etc to drop me a line. Naaut can help to connect the dots to implement or design Innovation Strategy solutions that apply technology to deliver impact. You may be a startup, scale up or even a prime. Naaut can help your organisation to understand the ‘fuzzy space’ where the magic of innovation can bring about the next leap for your strategic goals. Are you interested in how to apply Artificial Intelligence for Earth Observation or Blockchain for Future Supply-chain or how we bring leading talent from across the globe to deliver projects with a scrappy startup mentality? Then check out these short clips below and let me know if Naaut can help you. To any of my Space family congregating in Bremen, ping me to let me know you will be in town as I am looking to coordinate a meetup to get us all together. #IAC2018 #Bremen #InvolvingEveryone #machinelearning#artificialintelligence #quantumcomputing #space #newspace#asteroidmining #strategy #innovation #deeplearning

Naaut is coming to IAC2018

11

naaut-is-coming-to-iac2018-192e9f98f22e

2018-09-29

2018-09-29 12:31:06

https://medium.com/s/story/naaut-is-coming-to-iac2018-192e9f98f22e

false

255

Frontier Technology and Innovation for our Multi-Planetary Future

null

Naaut

nush.naaut@gmail.com

naaut

SPACE EXPLORATION,MARS,MOON,INNOVATION,TECHNOLOGY

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Anushka Sharma

Founder, Naaut 🚀 [ innovation * frontier technology * execution ]

22870033039b

nushkino

835

1,348

20,181,104

null

0

null

0

null

2018-09-11

2018-09-11 01:06:11

2018-09-11

2018-09-11 01:21:56

2

false

en

2018-09-11

2018-09-11 01:55:24

1

192f350f005a

2.17956

0

— — “AI isn’t a winner-take-it-all scenario. It will be different in different industries.”

4

GEORGE YAN, CEO OF CLOBOTICS, SPEAKS AT THE 2018 UBS CHINA A-SHARE CONFERENCE — — “AI isn’t a winner-take-it-all scenario. It will be different in different industries.” George Yan, CEO of Clobotics, was invited to attend the 2018 UBS China A-Share Conference sponsored by UBS Group, the largest multinational investment bank and financial services institution. The event’s panel included representatives from influential companies in artificial intelligence (AI), securities and asset management. The conference highlighted the present and the future of AI, as well as its impact on traditional industries. “The future of AI doesn’t only belong to the giants,” said George Yan. “For companies to become successful, [their] innovation of technology needs to meet the needs of customers, to win the respect of customers and to gain the favor of the market. AI isn’t a winner-take-it-all scenario. It will be different in different industries.” George Yan showcasing Clobotics’ AI solutions at UBS 2018 A-Share Conference As one of the industry’s highest-recognized seminars, the conference provided an opportunity to examine the future role of China’s innovative technology companies in global technology innovation. In addition, it recognized a significant increase in entrepreneurship in the AI industry which had replaced the Internet as a new spotlight in the primary and secondary markets. Moderator, Herbert Yang, discussing with George Yan (Clobotics) and Leo Zhu (Yitu Tech) at the UBS Panel regarding Saas and Cloud “The opportunities for AI startups lie in the vertical industries. The common misunderstanding is to build a one-size-fits-all application and try to sell to all businesses across the board. However, this is not what we believe how to grow as an AI startup. We believe that if we dig deep in the vertical industries and tailor our solutions around the domain knowledge and vertical-specific data, we can win the favor of the leading players in these verticals. These leaders will in turn set the trend for how things are done, and the smaller players will then follow,” Yan told the audiences. Since its establishment in 2016, Clobotics has quickly becoming the AI solution leader in the retail and wind energy industries, defeating competitors in the global market and making deals with the biggest players in these two industries. The three keywords of Clobotics’ secret sauce is: technology driven, hardware-software integration, and global scalability. “We are only uncovering the tip of the ice berg, there is so much more we could do to digitalize traditional industries and provide value to our customers,” said Yan. About Clobotics Clobotics is a global leader in intelligent computer vision solutions for the retail and wind power industries. Clobotics’ end-to-end solutions combine computer vision, artificial intelligence/machine learning, and data analytics software with different hardware form factors, including autonomous drones, mobile applications and other Internet of Things devices to help companies automate time-intensive operational processes, increase efficiencies and boost the bottom line through the use of real-time, data-driven insights and analysis. To learn more, visit http://www.clobotics.com

GEORGE YAN, CEO OF CLOBOTICS, SPEAKS AT THE 2018 UBS CHINA A-SHARE CONFERENCE

0

george-yan-ceo-of-clobotics-speaks-at-the-2018-ubs-china-a-share-conference-192f350f005a

2018-09-11

2018-09-11 01:55:24

https://medium.com/s/story/george-yan-ceo-of-clobotics-speaks-at-the-2018-ubs-china-a-share-conference-192f350f005a

false

476

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Clobotics

null

faf84b00b15

clobotics

5

10

20,181,104

null

0

null

0

33391d4d5793

2018-09-05

2018-09-05 22:21:58

2018-09-05

2018-09-05 22:22:24

1

false

en

2018-09-09

2018-09-09 15:19:42

0

192fc516562c

1.158491

0

The Coordinator was a system designed by Terry Winograd to “provide facilities for generating, transmitting,…

3

The Coordinator The Coordinator was a system designed by Terry Winograd to “provide facilities for generating, transmitting, storing, retrieving, and displaying messages that are records of moves in conversations.” (Winograd, 1987) Unlike Musicolor, which interprets the data input into its system, The Coordinator allows “people [to] do the interpretation of natural language, and let[s] the program deal with explicit declarations of structure” (Winograd, 1987). Wheraes a typical conversational interface provides “a uniform command to initiate a new message [(texting, email)], The Coordinator system provides options for opening conversations that have different implicit structures of action.” (Winograd, 1987) For example, “when Request is selected, templates appear prompting the user to specify an addressee, others who will receive copies, a domain, which groups or categorizes related conversations, and an action description, corresponding to the subject header in traditional mail systems.” (Winograd, 1987) If a user were to select a different option, they would be provided with a different template designed for that specific request. The Coordinator demonstrates how by making a user’s line of thought more visible to the other systems interacting with them, a conversation can be advanced in a more beneficial direction. The Coordinator Converse Menu By confirming “shared mental models” (Dubberly & Pangaro, 2009) especially during the convergence on agreement, conversational interfaces afford the successful exchange of “thoughts and words.” (OED Online, 2017) The Coordinator provides a successful example in “confirm[ing] … shared mental models” (Dubberly & Pangaro, 2009). Coordinator users were explicitly aware of what type of statement others were delivering, allowing for a better understanding.

The Coordinator

0

the-coordinator-192fc516562c

2018-09-09

2018-09-09 15:19:42

https://medium.com/s/story/the-coordinator-192fc516562c

false

254

Updates, findings and other things from my thesis, Conversational Symbiosis Amongst Humans and AI in the Context of Plateaus in Romantic Relationships

null

Men Are from Kepler-438b, Women Are from Kepler-442b

null

men-are-from-kepler-438b-women-are-from-kepler

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Scott Dombkowski

null

f6d1a37ec0ee

sdombkow

24

30

20,181,104

null

0

null

0

null

2018-07-21

2018-07-21 14:32:26

2018-07-21

2018-07-21 14:36:23

14

false

en

2018-07-21

2018-07-21 14:36:23

11

192fca59eab0

4.90283

0

Overview The world of digital currency has since its inception, experienced tremendous growth and popularity in the global economy…

5

KIRIK- THE SEMANTIC META PROTOCOL Overview The world of digital currency has since its inception, experienced tremendous growth and popularity in the global economy; thousands of cryptocurrencies exist and blockchain technology has been applied to various aspects of life. Most blockchain and cryptocurrency transactions function via smart contracts. Introduced in 1996 by Nick Szabo, smart contracts are trackable and irreversible computer protocols which aim to facilitate the execution of a contract or transaction digitally. In simple terms, just as the name implies, smart contracts make execution of transactions smarter. Smart contracts achieve this by the elimination of third-party intermediaries. Although smart contracts offer immense advantages when compared to conventional contracts, it also has its own shortcomings. The major problem with smart contracts is that they are slow to execute, coded by programmers and are thus difficult to understand to the normal man without programming experience. By utilizing artificial intelligence, KIRIK, a semantic contract platform, seeks to bridge the gap between smart contracts, the world we live in and the end-user without a programmable language. The KIRIK Initiative KIRIK is a cross-blockchain initiative which utilizes artificial intelligence to revolutionize blockchain transactions and smart contracts. The KIRIK project seeks to achieve this by introducing three major features: • Semantic Smart Contracts • Crossblockchain asset transfer platform. • A transition from AI 1.0 to AI 2.0. ●Semantic Contracts: KIRIK semantic contracts are an executable smart contract specification which executes specifications just the way a 3D printer prints out a 3D model via artificial technology. Unlike smart contracts, semantic contracts are very easy to use and are rather “model-based” instead of programming base. Semantic contracts bridge the gap between artificial intelligence and human logic. This is made possible by the use of AI 2.0, which is more user-friendly than AI 1.0. AI 2.0 makes artificial intelligence understandable to the “non-programmable” user. ● Cross chain Platform: Another added feature of the KIRIK project is that it offers cross-chain transactions. Thus, the problem of interoperability is now solved. KIRIK has a metaprotocol which allows for cross-blockchain asset transfer. This makes it possible for Bitcoins to be exchanged with another cryptocurrency such as Ethereum. ● User-Friendly Artificial Intelligence (AI 2.0): According to Techopedia.com “ Artificial intelligence (AI) is an area of computer science that emphasizes the creation of intelligent machines that work and react like humans…” The traditional artificial intelligence technology is highly technical and thus very difficult to understand by human logic. Apart from this, another drawback of artificial intelligence is that it is not secure. For example, an artificial intelligence chatterbot, Tay was originally released on March 23rd, 2016 by Microsoft Corporation. Moments later, it began posting derogatory and offensive tweets via its Twitter account. This sparked up serious controversy and forced Microsoft, its creator, to shut it down just some hours after its launch. KIRIK bridges this gap between artificial intelligence and the legal world by introducing an artificial intelligence which is secure, user-friendly and easily understandable by the non-programmer. The artificial intelligence version that KIRIK utilizes is AI 2.0. This is an improvement to AI 1.0. KIRIK Token and Token Sale Exercise The KIRIK token, with ticker KRK, is the native cryptocurrency of the KIRIK project. The KRK token is an ERC 20 token issued on the Ethereum blockchain, hence, it leverages on the blockchain’s numerous advantages like the security of transactions with low processing time which are cost-effective. The KRK token will be used for the development of the platform including marketing, promotion, and support. The KRK token sale exercise will commence soon and the whitelisting of potential participants of the Pre-sale slated for 15th- 17th August 2018 is underway. Details of the exercise and the token distribution can be found below: The proceeds from the token sale exercise will be used for various purposes and its breakdown is shown below: Team The Kirik team consists of individuals from various countries around the world with a remarkable background in engineering, business development, and software development. Founded by Vitaly Gumirov- a renowned Mathematician and Computer Scientist with decades of experience in IT and entrepreneurship and the founder of Eyeline Communications, the Kirik project has been developed to solve the challenges associated with inter-blockchain transactions and smart contracts. A brief of their profiles is shown below: Roadmap of Kirik for 2018 and 2019 The Kirik team has a detailed outline for the development of the project, with the various goals for the project along its development course being given a specific timeframe for its realization. The roadmap for the remainder of the year and 2019 are shown below: Conclusion No doubt, semantic contracts are the future of blockchain transactions. As more people continue to become aware of the amazing applications of blockchain technology and smart contracts, KIRIK will surely be in the best position to provide the adequate balance needed for the ideal artificial intelligence system which best fits the end user without any programming experience. A comparison of Kirik and the other types of contracts in existence is shown below, outlining the features of each. It is obvious that the Kirik solution is the next generation of smart-contracts with numerous advantages over other forms of contracts. To find out more about KIRIK or to become a member of the community, use any of the following links: WEBSITE WHITEPAPER LITEPAPER ══════════════════════════════════ SOCIAL MEDIA TELEGRAM FACEBOOK TWITTER MEDIUM GITHUB LINKEDIN REDDIT AUTHOR: kingigolo BCT LINK

KIRIK- THE SEMANTIC META PROTOCOL

0

kirik-the-semantic-meta-protocol-192fca59eab0

2018-07-23

2018-07-23 14:34:06

https://medium.com/s/story/kirik-the-semantic-meta-protocol-192fca59eab0

false

915

null

Blockchain

blockchain

Blockchain

265,164

Frank Djongu

Cryptocurrency Analyst# ICO Reviews# Blockchain Enthusiant# Cryptophile

24cb8e798925

fisher0frankie

802

812

20,181,104

null

0

null

0

null

2018-04-27

2018-04-27 18:32:27

2018-04-27

2018-04-27 18:34:07

1

false

en

2018-04-27

2018-04-27 18:34:48

4

1933b0f5fa0d

5.290566

0

At Lattice, we are obsessed with helping marketers realize the promise of delivering personalized experiences and growing relationships at…

5

ABM @ Scale: Meet Lattice Atlas At Lattice, we are obsessed with helping marketers realize the promise of delivering personalized experiences and growing relationships at scale. We thrive on solving the biggest customer challenges in an unconventional and radically simple manner by leveraging cutting edge AI technologies. Today, we are continuing this mission by launching Lattice Atlas, the industry’s first customer data platform for ABM at scale. In Part 1 of this series, our VP of product Marketing Nipul Chokshi highlights reasons why marketers are challenged with scaling their ABM programs and how that hinders their ability to drive impact. The challenges they face center around getting customer data and insights in a unified way. Let me give a you simple example. One of our customers wanted to create an ABM program to acquire net new customers that looked like their existing customer base (“lookalikes”). This seemingly straightforward program requires a highly manual effort using ad-hoc tools and workflows supported by a large ops team. Why? Well, let’s take a look the long list of tasks they have to do: Aggregate data from four sources — target accounts from CRM, marketing campaign responses from MAP, website engagements from their website analytics platform and 3rd party intent data. In more complex environments, the data sources can be as high as twenty! Run the data-set through an AI platform to identify and segment lookalike audiences based on their existing customer profile. Socialize the audiences with the business teams to get their buy-in Ensure they have “right sized” the audiences to optimize capacity/budget constraints along with their campaign objectives. Ensure any opt-outs (to meet GDPR/compliance requirements) are removed, as well as anyone that’s already in a sales cycle. Load the audiences in their various campaign execution platforms (LinkedIn, email SDR channels, etc.) Set up campaigns in each of the channels with the right content and other parameters. Content needs to consistent for all audiences Aggregate the engagement across the different channels (using spreadsheets) to report on the effectiveness of the ABM program, then rinse and repeat from step one. All this work to support one narrow ABM program for one solution. Imagine the level of effort required if they wanted to run separate programs for various stages of the buyer’s journey or if they wanted to run programs based on customer lifecycle stage across each of their product lines. It is almost impossible to do so with the fragmented data and disparate applications glued together with ad-hoc tools and processes. So, unless we wanted our customers to limit their ABM efforts to a small scale, we needed to innovate — at a massive scale. We realized that our customers needed a Customer Data Platform (CDP) that unifies all data, enables AI-driven audience creation as well as omnichannel activation & personalization all in one centralized place, and provides enterprise-grade marketing governance. That’s how Lattice Atlas was born. This next generation AI-platform helps marketers scale their ABM programs by removing four major bottlenecks. Unified Customer Data Lattice Atlas is architected on open platform principles and the data model is designed to aggregate a wide variety of data, enabling customers to create 360-view of their prospects and customers. It has a standardized data ingestion interface to import 1st-party data from internal systems quickly & easily. In line with our tradition of building everything to scale from day one, the platform is designed to process billions of signals everyday. Our patent-pending Adaptive Match technology then does the hard job of resolving identities, matching the 1st-party data to 3rd-party data in Lattice Data Cloud with over 20,000 curated insights and automating the data unification. This automated process dramatically reduces the time and effort needed to create a 360-view, eliminating the first bottleneck to enabling ABM at scale. AI-driven Audiences Radical simplicity is in our DNA — we decided to build Lattice Atlas with the goal of enabling marketing teams to build highly targeted audiences with simple point and click in matter of minutes — not days, or weeks. With self-service modeling and automated rescoring embedded in the platform, every account and every contact is automatically rescored as the new signals are detected, creating AI-driven insights for the next best actions and personalized engagements every step of the journey. The product experience is designed to let the marketing teams do what-if analyses on the fly, define criteria for the next best actions and “right size” the audiences with just a few clicks. For the always-on campaigns, marketers can define the criteria once and the platform creates the audiences for them automatically and keeps it up to date, enabling personalized engagements at scale. The platform also lets marketers manage brand reputation by assigning engagement thresholds and removing targets that may have reached marketing fatigue. The AI-driven insights and simple marketer-friendly way to create and manage targeted audiences removes the second bottleneck in enabling ABM at scale. Omnichannel Activation and Personalization According to McKinsey the average B2B customer engages with 6 channels prior to purchase. So how do we support these disparate channels? Well, we opened the Lattice Atlas to integrate with all of them. The platform has real-time REST APIs so our customers can pull audiences and recommendations in any channel in real-time and enable programmatic personalization at scale. We are also investing heavily in building out-of-the-box apps for the most common channels so that customers can scale their ABM programs quickly and easily. Enabling marketing teams to programmatically deliver the right message at the right time via the right channel eliminates the third bottleneck facing marketers. “Informatica knows that creating an interactive and personalized buyer’s journey is critical for the success of account-centric programs,” said Steven Shapiro, VP of Digital and the Buyer’s Journey at Informatica. “We’d seen previous success with Lattice and knew that the Lattice Atlas platform would create the personalized experiences we needed across all channels. Our vision is to use Lattice as the AI brain that powers all next-best-actions.” Enterprise-grade Governance Lattice has been the pioneer when it comes to meeting the security needs of our customers. Lattice Atlas continues to carry the baton forward with new data governance capabilities. With GDPR on the horizon, we knew that ABM programs couldn’t get off the ground unless there was an automated and simple way to remove the opt-outs. That’s why we built a self-service interface for marketing teams to remove opted out accounts and contacts from their campaigns. This centralized system to manage inclusions and exclusions globally, for specific channels and/or for specific campaigns makes it easier than ever for driving privacy-compliant ABM programs at scale, and eliminating the fourth and final bottleneck facing scalable ABM programs Lattice Atlas was a natural evolution of our platform. Since day 1, our approach has focused on being deeply integrated with each execution application and managing all data under one platform. Because of this we not only capture the largest amount of data, but also all that relevant metadata that describes it. Lattice Atlas is built on our understanding of these applications and their data to create the first CDP for enabling ABM at scale. The first application debuting on the platform is Playmaker, which delivers prescriptive recommendations to sales teams adopting play-based selling. Our customers sell many solution across many audiences. Playmaker lets them quickly identify top products to sell across all audiences and programatically deliver those recommendations to the sales teams. It also has built-in interactive dashboard to track the engagements (or lack of it) and its impact on the pipeline, enabling out-of-the-box visibility into play ROI measurements and the ways to improve it. Lattice Atlas is currently in private beta and is expected to be generally available in Q4. This release marks the next big leap in Lattice’s evolution as a marketing AI company, and it is just the beginning, so stay tuned for more exciting updates.

ABM @ Scale: Meet Lattice Atlas

0

abm-scale-meet-lattice-atlas-1933b0f5fa0d

2018-04-27

2018-04-27 18:34:49

https://medium.com/s/story/abm-scale-meet-lattice-atlas-1933b0f5fa0d

false

1,349

null

Marketing

marketing

Marketing

170,910

Chitrang Shah

null

ae7982abb7e0

chitrangs01

1

10

20,181,104

null

0

null

0

e3865858e6a4

2018-06-25

2018-06-25 13:21:26

2018-08-21

2018-08-21 14:00:47

11

false

en

2018-08-22

2018-08-22 10:46:17

15

1933dcfb3007

13.160377

3

1

0

Back in June we attended CogX, an annual festival of all things AI held in London. It was huge:

5

AI: Artificial Intelligence or Artificially Inflated…? The CogX AI conference at London’s Tobacco Dock Back in June we attended CogX, an annual festival of all things AI held in London. It was huge: 6,500 attendees 17 stages of content 300 speakers In this blog I’m going to write about a few of the talks that really grabbed our attention. But first… Why did we go? Cutting edge technology fascinates us. As we continue to develop Gurn we’re always looking to the future to try to see what we tech we could use to improve the value we deliver to our users. Whilst AI per se isn’t something we are looking to capitalise on right this minute, we definitely are wanting to explore how Machine Learning can improve the accuracy and relevancy of the results we return for our users. And therefore we wanted to get out there, listen to some talks, meet some people and find out what’s going on in the world of AI. We saw lots of talks over the two days. Some amazing, some good and some way beyond our field of expertise — which was scary, but brilliant! Why the title for the blog? Firstly, thank you to Manoj Saxena (Chairman of Cognitive Scale) for the inpiration for the blog title. In his introduction to the session on Enterprise Augmented Intelligence in Financial Services Manoj gave a great overview of where he saw AI today, with the proclamation that… “AI today is both artificially inflated and also stands for amazing innovations” We think he is spot on — recognising there are some brilliant advancements out there, but there is also a lot of hot air too…! Anyway, enough of the introduction. Who did we listen to and what did they say? Matt Hancock MP, then Secretary of State: Digital, Culture, Media and Sport Gov & Tech — a match made in…? Being a UK-based startup we made sure we attended this talk. It was given by Matt Hancock MP, who was UK Minister repsonsible for ‘Digital’. He’s also been dubbed the ‘Minister for AI’ — mainly because it falls under his brief, but also because not many of his colleagues are keen to roll up their sleeves and take the time to ‘get to grips’ with AI. He’s keen, he’s vocal and he gets out and meets the tech entrepreneurs and is relentlessly pushing for dialogue between tech and government. (Not an easy job!) His talk was very much one of a politician, but what can you expect? He was out there to reassure that the UK government ‘gets it’ and that they recognise they need to do more — which was encouraging, plus he called from greater investment from the Government, but the proof is in the pudding as we Brits say! Will £1bn keep pace with China and the US…? He spoke about how the UK’s risk appetite in business is not one for adventure and that he wants to “see more freedom to fail” across UK business. How you might ask? Well he didn’t really say… The State of AI panel discussion This was a very lively debate between Jurgen Schmidhuber, Antoine Blondeau, Vishal Chatrath and Joanna Bryson. Initial remarks were about how UK business is scared to fail (backing up Matt Hancock MP’s assertion) and that this holds back the Tech industry. Leaders of companies need to take a stance, and show that failure is part of the process toward success. The bit that really fascianted us was when the panel discussed the regulation of AI. In life you can generally seek permission or ask for forgiveness. It’s fair to say that the panel were split on how this should apply to AI. Antoine Blondeau stated “I’m fundamentally against regulation” and he wishes peoples’ efforts to be focussed on innovation and empowerment. Whereas Vishal Chatrath spoke of the need for ‘export controls’ on AI to avoid its weaponisation, and in order to do that we need regulation and standards. His point was supported by Joanna Bryson who didn’t see regulation as a bad thing, but this debate about AI is a chance for countries to cooperate. There must be a way where light-touch regulation doesn’t stifle innovation? Where coperating countries can start to lay down some standards, but not via the UN where it’ll take years to make any progress. The World Economic Forum perhaps? But then if a country strays out of line what ‘stick’ do you have to hit them with to discentivise deviation from the agreed standards? Catch the video of the panel discussion here Calum Chace, The Economic Singularity Credit: Splento Calum Chace’s talk was fascinating, thought-provoking and unapologetically cutting. Best soundbite of the talk for us: “AI is collar-blind” There’s definitely a public perception that ‘the machines’ are coming primarily for the jobs of those working in factories, and that other industries are immune. But this is patently wrong. One of Calum’s main points was that it doesn’t matter whether it’s a blue collar job or a white collar job, if AI can reduce the cost of delivering the outcome then why would people not deploy AI to do so? Especially if your cost p/unit/hour is high, i.e. white collar jobs such as auditing and legal services. Calum explained that people are clinging to past performance as a guide for future success. Whereas we’re always taught, especially in the financial world, that past performance is not an indicator of future returns. Politicians’ standard reaction to the idea that AI will take people’s jobs: hear no job losses, see no job losses and speak no job losses (Photo by Joao Tzanno on Unsplash) What about his view on what politicians are saying? He made the point that politicians can do nothing but say AI will create jobs. They have too much to lose in the short-term, i.e. their jobs via the ballot box, and will bury their heads in the sand even if it’s it means their citizens will lose in the long-term. Also politicians know that if people thought robots were coming for their jobs then they would panic, and politicians like nation-wide panic about as much as they like losing elections. The talk then moved onto the evolution of work, which has seen the migration of people from different ‘arenas of work’ due to technological advancements. Calum explained that the second agricultural revolution meant people left the field for the factory, the industrial revolution meant people left the factories for the office, but where do these workers go next…? Calum’s talk was one of the best talks of the conference. We see a lot of ostriches out there with regards to what AI’s impact on jobs might be, and his insights show that we need to be discussing this now. What is there to lose? …apart from not having your workforce prepared for what’s coming down the line. Then there really will be panic. If you want to read more of Calum’s work, then as a starter I highly recommend checking out his blog on The Reverse Luddite theory. Baroness Onora O’Neill: Communication and Trustworthiness Where are we with trust online? (Photo by Bernard Hermant on Unsplash) This talk was awesome. And it was probably the least technical talk we attended over the two days. Although the title was Communication and Trustworthiness, the talk really centred around who is responsible for content in the modern day and whether companies will self-regulate. To begin with, Baroness O’Neill took us on a journey from the Classical world of Socrates to the era of the invention of the printing press right through to the modern day, exploring how the responsbility for the spoken word and written word has differed throughout the ages. Understanding who was repsonsible for the spoken word tended to be simple, Baroness O’Neill stated, as the orator was the sole distributor; they are responsible for their content. With orators they are present at the point of authoritative delivery — they can answer questions, and clarify. With the written word things are harder once published. Who truly wrote it and who is responsible for the content? With the arrival of the printing press in the early modern period neither was not answered… You could publish with anonymity and… “…there was no distinction between a printer and a publisher, so who was responsible for the stuff that got produced?” These questions in the age of social media, according to Baroness O’Neill, is still waiting to be solved. Nowadays, when written word is consumed on social media, it is not always possible to identify the actual author to seek further clarification. So who is therefore responsible? The potentially untraceable author or the publisher? Are we all just cyber romantics? When the internet was coming into mass use the public-at-large didn’t really stop to think about these questions, as we were all too eager to get our hands on technology. I remember it being almost like a ‘consumer arms race’ — being able to go into school and say you had played with the newest tech, which would then catapult you up into the stratosphere of ‘cool’. But were we too hasty craving this technology? Should we (by this I mean the adults in the room) have paused and tried to resolve these questions? “10 years ago one might think of it as the age of the cyber romantics, everybody thought it was all net gain, things were getting better and better” Here, Baroness O’Neill reminded me that in my more recent years I was not fully aware of the true cost of using online services — that in fact, it wasn’t all gain, and that ‘free’ didn’t mean free at all. Had people begun to discuss these issues more widely at earlier stages in the development of the modern social media age then the ethical storms that have arisen may have been seen through a different lens, one where people fully understood what they were giving up in return for using a free service. Are the ethical storms behind us? Or are we just at the beginning? (Photo by Tom Strecker on Unsplash) Baroness O’Neill classes these storms as private harms and public harms, which she characterises as follows: Private harms are consequences from activities such as cyber bullying and financial crime etc. Public harms are those which harm democracy and harms the public space. If social media companies are to take responsibility for the publishing of content on their platforms then Baroness O’Neill sees private harms as being the reason they will be moved to do so. This is because there “is at least some convergence of interest between members of the public and the profits of these companies”. An example cited is advertising, where companies’ adverts appear alongside content that causes private harm. This would then cause companies to pull advertising spend from the social media companies, something we have actually seen this past year. Though this needs private harm to be experienced at scale to have an effect. But what about self-regulation? Is that possible? A non-starter for Baroness O’Neill. As she stated how she recently realised that it was naiive to ever have thought the tech companies would be interested in taking on the responsibilities of being publishers. “I’ve come to think that is an illusion.” And, even if leglisation was introduced to assign publishing responsibilities onto tech companies this would have limited effect. This is because these companies operate as distributed businesses by off-shoring themselves, in terms of labour and tax for example, and therefore would continue with this model so that the entity responsible for ‘publishing’ would be resident within a jurisdication that wouldn’t burden them with said responsibilties. “I don’t think we can really expect to enter a world in which the online service providers, social media companies, data anlaytic companies [take on publishing responsibilities]… it is too profitable to not do it.” What next? Baroness O’Neill didn’t exactly leave us with a clear plan to solve this dilemma. Instead she mentioned some next steps we should take, including defining what types of communciation are acceptable, and which ones aren’t. When this is agreed we can press for redress for private harms and discuss what we, as users of the technologies, are willing to put up with. The trouble is people are still reticent to vote with their feet, an example being the recent #DeleteFacebook campiagn; because whilst some might take a stand and leave their friends probably haven’t, and so the fear of missing out will start to creep in. Is artificially extending life intelligent? (Left to right): Dr Jack Kreindler, Dr Gregory Bailey, Polina Mamoshina, Matt Eagles and Maxine Mackintosh This was a panel discussion hosted by Dr Jack Kreindler. The key takeaway for us was the question ‘are we talking about extending healthspan or lifespan?’ The example being a family choosing to extend the life of a loved one who has dementia. What are their motivations, and who’s emotional happiness do they have in mind by keeping their relative alive for another 15 years let’s say? As Dr Kreindler asked, “if we can extend someone’s life so they live till they’re 150, but if 70 years of that is in poor cognitive condition, or a huge net drain on resources, then is it a worthy pursuit?” The debate is nuanced, but the panel seemed split. With some members of the panel clearly seeing extension of life via AI as a success, regardless of whether healthspan of lifespan was extended. Maxine Mackintosh made it clear that it should be a choice for the person who’s life is being extended — if they have the ability to follow a logical thought process and the outcome is extending their happiness then why wouldn’t you enable them to have their life extended? Will AI be your next Chief Marketing Officer? Wes Nichols (Board Partner, Upfront Ventures) Wes Nichols took us on a whirlwind tour of the past, present and future of marketing. He made it clear that a lot of current day marketing methods and practices belong in the past, that too many companies are focussed on analysing what has happened and are not looking forward, using data to predict and validate the outcomes they can achieve. Specifically that the next level of marketing will be when someone says to their boss, “I know within a small margin of error that if I do X then I will achieve Y. So give me £Z so I can achieve Y many times over.” And the boss will trust them to deliver. Stop looking backwards, use your data to predict what you can achieve Wes sees AI as powering marketing decisions, but that humans won’t be removed from the equation entirely. “AI is more Augmented Intelligence, fusing the machine and the human for superior results.” Wes invoked Charles Darwin to explain that we must become ‘responsive to change’, now more than ever, because we are experiencing change faster than ever before. One piece of advice that really stuck with me, having spoken with friends recently about the large corporates they work for, is that if you work for a company and see it isn’t evolving, but standing still or changing at a glacial pace, then get out… “There’s lots of exciting companies out there to work for.” As co-founder of a startup, even we are concious of not standing still, yet I see many friends at large coporates which are still moving at a glacial pace and launching enormous top down initiatives that fail way more often than not. Why do they stay? I presume it’s mainly down to job security and very competitive compensation packages — therefore to prise them away/encourage them to take that leap of faith people have to believe in the vision of the company they’re going to work for and have courage. (that, and some nice stock options!). How do we equip people for the future and re-skill the workforce? (Left to right): Phil Smith (Chairman, Innovate UK), Deep Nishar (Senior Managing Partner, SoftBank Investment Advisors), Liz Ericson (Partner, McKinsey), Baroness Joanna Shields (CEO, Benevolent AI), Kathryn Parsons (CEO, Decoded) and Shiva Rajaraman (Chief Product Officer, WeWork) This panel discussion was chaired by Phil Smith, and gave us an insight into the different views amongst senior leaders for how we should prepare ourselves, and our children, for the future of work. Sending your kids to code camps is one of the ways to secure their success in the future workplace according to Kathryn Parsons. However, this was in stark contrast to Deep Nishar who said his advice to his children was to look the Humanities and Arts, and not to learn to code, “as machines will do that in 10 years”, instead… “learn to learn. Be able to understand and phrase problems in a way that convinces/educates people and insturcts machines what to build.” Whether or not machines will be able to code in the next decade, it seems interesting that people are rushing towards becoming experts in disciplines that, for humans at least, may not be around in the future; as Deep Nishar went on to say, “ultimately coders are putting themselves out of business”. And that we need to accept that… “The only constant is change” If Deep Nishar is right about the need for a focus on the Humanities and Arts, then what are the softer skills that we should be learning? Shiva Rajaraman from WeWork said he believes there are four skills people should learn to prepare them for the future: Pitching — be able to explain your ideas and get buy-in from people. Negotiating — be able to achieve your desired outcomes. Understanding data — analysis of information to make sense of what to do next. Psychology — understand the emotional levers and how best to navigate humans I think Shiva is pretty spot on! Our parting thoughts What we’ll take with us… (Photo by Cristina Gottardi on Unsplash) A worthwhile event to attend. We came away reassured; reassured that smarter people than us are talking about AI’s impact at very senior levels. Being able to listen to senior technology leaders talk about, and debate, where they think AI is at the moment was useful and sobering. Most speakers were very pragmatic, and most were concious of how AI is set to impact people’s lives — both positively and negatively. Though still, only a few were actually talking about what we must do to prepare the future-displaced workforce and how are they provided for. Always go and listen to talks outside your comfort zone! You’ll never know what you might stumble on. Our ‘we weren’t planning to go to that’ highlight: Kamil Tamiola’s talk on Directed Protein Evolution. Check it out here. Would have been great if… In a few talks some panellists could have had their views challenged a bit more. In a few talks some panellists made statements that were just accepted as ‘fact’, and weren’t asked to explain the evidence or sources. Instead of ‘meet the speaker’ there were structured workshops where you can work through some of their research with them and talk in detail about how to apply it to problems. Food inside the venue was £££. (thankfully there was a McDonald’s two minutes walk from the venue :-) ) Until next year…? Rory is co-founder of www.gurn.io. Gurn is a browser-based information retrieval tool that enables you to navigate to all your tools and resources with a single command. Check out how it can help you be more productive at work by getting you to what you need instantly and speeding up your workflow.

AI: Artificial Intelligence or Artificially Inflated…?

41

ai-artificial-intelligence-or-artificially-inflated-1933dcfb3007

2018-08-22

2018-08-22 10:46:17

https://medium.com/s/story/ai-artificial-intelligence-or-artificially-inflated-1933dcfb3007

false

3,143

Technology Startup, Bristol - UK

null

gradienthq

null

GradientHQ

hello@gradient.co

gradienthq

GRADIENT,PRODUCTIVITY,STARTUP,BRISTOL,SOFTWARE ENGINEERING

gradienthq

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Rory Macduff-Duncan

Ever-curious. And trying to give people some time back in their lives. Co-founder at Gradient - makers of www.gurn.io.

63a985ef1e24

r.macduffduncan

15

23

20,181,104

null

0

null

0

null

2018-01-05

2018-01-05 18:01:51

2018-01-05

2018-01-05 18:04:27

1

false

en

2018-01-05

2018-01-05 18:04:27

16

1934ececfd42

3.407547

0

Welcome in 2018 and a Happy New Year! New AI weekly focuses mainly on China (probably biggest investor in AI right now) and some…

5

AI Weekly 5 Jan 2018 Welcome in 2018 and a Happy New Year! New AI weekly focuses mainly on China (probably biggest investor in AI right now) and some predictions for upcoming months. But that’s not all, you can find here also new cool summaries of AI related events and libraries from 2017. Enjoy your weekend reading other AI news and don’t forget to share it with your friends 😉 GENERAL The 5 Most Interesting Artificial Intelligence Trends for Entrepreneurs to Follow in 2018 — The current pace of innovation makes it almost impossible to stay on top of the AI trends, but understanding these advancements is a must for business owners who want to stay ahead. http://bit.ly/2lYtob4 China is building a giant $2.1 billion research park dedicated to developing A.I. — China is planning to build a 13.8 billion yuan ($2.1 billion) technology park dedicated to developing artificial intelligence (AI), state-backed news agency Xinhua reported Wednesday. The campus will be constructed within five years and situated in the suburban Mentougou district in western Beijing. It will cover 54.87 hectares, Xinhua said. The technology park will be home to around 400 businesses and is expected to create an annual output value of about 50 billion yuan. http://cnb.cx/2EbXe3p These are the two books that are helping Xi Jinping understand AI — China’s president Xi Jinping is an avid reader. He peppers his speeches with quotes from his favorite writers, including Charles Dickens, Victor Hugo, and Paulo Coelho. His annual New Year’s Day greetings offers a rare look into the secrets of his office bookshelf — something Chinese netizens enthusiastically examine each January. http://bit.ly/2CLqpOv How an A.I. ‘Cat-and-Mouse Game’ Generates Believable Fake Photos — At a lab in Finland, a small team of Nvidia researchers recently built a system that can analyze thousands of (real) celebrity snapshots, recognize common patterns, and create new images that look much the same — but are still a little different. The system can also generate realistic images of horses, buses, bicycles, plants and many other common objects. http://nyti.ms/2Eb1Qqc AI in 2018: Google seeks to turn early focus on AI into cash — Alphabet has spent billions injecting machine learning into all aspects of its business, though returns will be hard to track http://on.mktw.net/2F2jduM How Do You Vote? 50 Million Google Images Give a Clue — What vehicle is most strongly associated with Republican voting districts? Extended-cab pickup trucks. For Democratic districts? Sedans. Those conclusions may not be particularly surprising. After all, market researchers and political analysts have studied such things for decades. But what is surprising is how researchers working on an ambitious project based at Stanford University reached those conclusions: by analyzing 50 million images and location data from Google Street View, the street-scene feature of the online giant’s mapping service. http://nyti.ms/2CJilgU Deep learning sharpens views of cells and genes — Eyes are said to be the window to the soul — but researchers at Google see them as indicators of a person’s health. The technology giant is using deep learning to predict a person’s blood pressure, age and smoking status by analysing a photograph of their retina. Google’s computers glean clues from the arrangement of blood vessels — and a preliminary study suggests that the machines can use this information to predict whether someone is at risk of an impending heart attack. http://go.nature.com/2D0CNqN Statistical Computing for Scientists and Engineers — full course with videos, slides and homework http://bit.ly/2lXcl9e AI and Deep Learning in 2017 — A Year in Review by Denny Britz http://bit.ly/2CWbwWa PROGRAMMING 30 Amazing Machine Learning Projects for the Past Year (v.2018) — For the past year, we’ve compared nearly 8,800 open source Machine Learning projects to pick Top 30 (0.3% chance). This is an extremely competitive list and it carefully picks the best open source Machine Learning libraries, datasets and apps published between January and December 2017. Mybridge AI evaluates the quality by considering popularity, engagement and recency. To give you an idea about the quality, the average number of Github stars is 3,558. http://bit.ly/2AxQ2fo Improving your data science workflow with Docker — Containerization is a trend that is taking the tech world by storm, but how can you, a data scientist, use it to improve your workflow? Let’s start with some basics of containerization and specifically Docker and then we’ll look at a couple of use cases for containerized docker science. http://bit.ly/2EbbadF TensorFlow 1.5.0 Release Candidate — http://bit.ly/2lYBYGA AIGaming — a platform that allows computer programs — also known as bots, to play each other at challenging games to win bitcoin. http://bit.ly/2m1TgTw wav2letter — Facebook AI Research Automatic Speech Recognition Toolkit http://bit.ly/2m06Vua PAPERS DeepMind Control Suite — The DeepMind Control Suite is a set of continuous control tasks with a standardised structure and interpretable rewards, intended to serve as performance benchmarks for reinforcement learning agents. The tasks are written in Python and powered by the MuJoCo physics engine, making them easy to use and modify. http://bit.ly/2AydWaJ If you want to be always on time with AI weekly, feel free to follow it on fb http://bit.ly/2yGcZhh

AI Weekly 5 Jan 2018

0

ai-weekly-5-jan-2018-1934ececfd42

2018-05-04

2018-05-04 00:08:51

https://medium.com/s/story/ai-weekly-5-jan-2018-1934ececfd42

false

850

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Tomasz Baranowicz

http://theaigeek.com/

22e656530332

tbaranowicz

112

30

20,181,104

null

0

null

0

null

2018-07-22

2018-07-22 17:11:59

2018-07-24

2018-07-24 16:46:54

2

false

en

2018-09-12

2018-09-12 15:26:57

3

19353e4fdf90

3.262579

4

0

I’ve been blogging and making videos about data science for a few years now. When I first started producing, circa 2014, I was surprised by…

5

How Blogging and Making YouTube Videos Landed Me the Best Job Source: Lucas Amunategui I’ve been blogging and making videos about data science for a few years now. When I first started producing, circa 2014, I was surprised by the amount of interest and positive comments they generated. Data science was a bit different back then and hardly any schools offered applied programs around that subject. For those wanting in, the barrier to entry was low but so was the amount of material to get in. That said, today, that barrier is still relatively low, and the industry is still desperate for talent and experience. Passion Is Great but Proof Is Better Passion is a great driver but padding that with some proof can sometimes drive you much further. Proof is a subjective measure at the whims of strangers, but once it is given, the rewards can surprise you. And, thankfully, when you’re really into something, sharing it to the world is reward enough — all the extras are just more icing on the cake. Where You Live Matters, or Does It? Unlike San Francisco, NYC, or Seattle, I lived in an area where applied data science master groups weren’t plentiful. Not having access to these advanced networking circles, you would naturally gravitate online and follow or create digital ones. And that’s what I did, my virtual community grew along with my digital network. I was too excited about the nascent field of applied data science and machine learning that I couldn’t shut-up about it. As my community was mostly online, it came out in the form of blog posts and YouTube videos. Every new thing I learned, I wrote about it and posted it online. Once worried about finding topics, it quickly became apparent that triage was going to be necessary to sort through the waterfall of ideas flowing. Once worried that people would laugh at some of my naive stories, the thanks, and encouragement I received made that point moot. Remember, there will always be people ahead and behind you, and those behind you really appreciate reading about what lies just a few steps ahead. I did go a little overboard and got into green screens, lights, even a teleprompter. I quickly downgraded back to the core minimum to not waste time on trappings and focus entirely on content. The Job Finding that fantastic job was a bit of a back thought, nice to daydream about. Actually, I also really liked the job I was holding at the time. I was contemplating starting my own venture. See, I live in Portland, Oregon, where leadership jobs in data science aren’t as plentiful than in the other areas mentioned earlier. In this town, if you want to shortcut your jump you have to be crafty and either work remotely, invent your own ride or relocate. From Passion to Thought Leader A thought leader is a buzzy word these days but it fits. Eventually, people reached out to collaborate, some to offer their services and others looking to hire me. And it happened, I started collaborating with a few startups and a special relationship grew with one of them. In a lot of ways, it played out exactly as I had imagined. I would get my leadership job, a rarity in this neck-of-the-woods. Got the pleasure of getting involved in a small startup and seeing it grow 10-fold. And best of all, I got to cook in the kitchen, along with the founders, without leaving my home turf. Final Tip One more tip, and that’s what I love about Medium, there’s always a practical gift or two buried in there. Here is mine, I keep my Linkedin profile up-to-date with all the buzzwords you can imagine and the highlights that intersect well with what is hot in the industry. The one thing I put on Linkedin that I think clinched the deal, was adding the following sentence in the headline field: “If you need a CDS, ping me” CDS stands for ‘Chief Data Scientist’. You need to let the world know what you’re looking for and by using a discreet abbreviation made it feel like a secret handshake. Please share and clap if you found this helpful — thanks for reading! Manuel Amunategui Get it and plenty more at amunategui.github.io and at ViralML.com. OK — Sign up to my email group below and I’ll send you my free eBook on tips to becoming a (better) data scientist (and signup even if you aren’t interested in the eBook). Thanks for reading!!

How Blogging and Making YouTube Videos Landed Me the Best Job

58

how-blogging-and-making-youtube-videos-landed-me-the-best-job-19353e4fdf90

2018-09-12

2018-09-12 15:26:57

https://medium.com/s/story/how-blogging-and-making-youtube-videos-landed-me-the-best-job-19353e4fdf90

false

763

null

Data Science

data-science

Data Science

33,617

Manuel Amunategui

Anything Applied Data Science. Author of Monetizing Machine Learning. VP SpringML. Barcelona.

5e29cc9fde82

amunategui

170

338

20,181,104

null

0

null

0

null

2018-02-21

2018-02-21 23:24:01

2018-02-23

2018-02-23 14:23:14

1

false

en

2018-02-23

2018-02-23 14:26:58

3

1935cbcb5b60

1.969811

13

1

0

Bit of an update from me, and like all big things it feels almost anticlimactic (spoiler: it is not).

4

Joining Zeroth.AI Bit of an update from me, and like all big things it feels almost anticlimactic (spoiler: it is not). I have joined Zeroth.AI full time. This means I will focus on investing in the best and the brightest entrepreneurs that are changing AI and the world through it. I will help them grow and succeed. Well, at least that’s the plan. I have realized that really I don’t know anything about the tech industry, I have been wrong so many times on so many things that can’t count them, but — BUT — because I spent the last 23 years as an entrepreneur at least I have some experience on resource allocation and deep understanding of how technology works (“why” is still a mystery), lots of contacts, and enthusiasm. Now I am in my first few weeks on the investment side and know even less — only this time do not have any experience. My favorite Japanese proverb is「初心に戻る」Shoshin ni modoru, which means to go back and remember the feeling of being a beginner. And it’s a very interesting feeling indeed. Some of you might be aware that I have been looking to jump on the investment side for a while. The main reason was that I have been long on AI, the turning point being around 2010, and I think we are just at the beginning of it. But I do not think progress is going be sequential in the market. Everything is going to happen at once and the usual 5–7 years to start and exit successfully a company is a price too high to pay. Since I realistically can’t found multiple companies at once then the next logical step would be at the crucible of creation of AI startups. And Zeroth.AI was a great fit. I have known Tak for several years, as a friend, as an investor, and as a mentor. I helped him and the rest of the team on the side for the first two Zeroth.AI cohorts and now I am going to be working directly with the Z03 batch and beyond. I will be in Hong Kong during the program and I am still going to be based in London. My primary focus is going to be supporting our portfolio companies and fundraising in the UK, Europe and the SF Bay Area. Aside from that, if you are considering starting an AI company, please reach out. If you have an AI startup and are looking for funding, message me. If you want help, advice, mentorship, or are outside a major tech hub, just ask. If you can’t get funding because you are doing some impossibly crazy 💩, I want to know. Best way is way DM on Twitter @rodolfor. At least still have a rolodex and lot of enthusiasm. -Rodolfo

Joining Zeroth.AI

45

joining-zeroth-ai-1935cbcb5b60

2018-03-02

2018-03-02 06:53:25

https://medium.com/s/story/joining-zeroth-ai-1935cbcb5b60

false

469

null

Startup

startup

Startup

331,914

Rodolfo Rosini

Partner @Zeroth_AI. Excited about AI, UX, Mobile, Security and DUNE. Often NSFW and WTF.

27b95bb5ec10

rodolfor

1,172

216

20,181,104

null

0

null

0

null

2018-07-16

2018-07-16 21:03:38

2018-07-16

2018-07-16 21:06:46

5

false

en

2018-07-16

2018-07-16 21:06:46

0

1936a4efdc24

2.278616

0

The results of the prediction model are below, and I give a detailed report of how I came to this after. The results shown are based on…

3

2018 QB Fantasy Point Predictions The results of the prediction model are below, and I give a detailed report of how I came to this after. The results shown are based on ESPN Standard Scoring and are only for the predicted starting QBs in 2018. Rookies are not included in the prediction model. If you are looking for a points prediction on Cardale Jones as a a deep sleeper or Josh Rosen, you’re going to have to do your own research. Results: Explanation of Analysis: Data: individual gamelog data for all QBs was gathered from ESPN stats. If possible, data going back to 2012 was collected. Data points collected was Season, Date of Game, Opponent, Result, Pass Attempts, Pass Completions, Passing Yards, Completion Percentage, Average Pass Length, Longest Pass, Passing Touchdowns, Interceptions, QBR, Rating, Rushing Attempts, Rushing Yards, Average Rush, Longest Rush, and Rushing Touchdowns. I added an additional column calculating the players score per game according to ESPN’s standard scoring. Modelling: using Azure’s Machine Learning Studio I built several regression models in order to to gauge how each model performed. The model trained on 2012–2016 data and was tested on 2017 data. Below are the results from the first run: The Bayesian Linear Regression model performed the best based on the Root Mean Squared Error. However after discussing the models with some friends, I decided to run the models a second time removing any column not relevant to the player’s score (Date of Game. Opponent, Result, Completion Percentage, Average Pass Length, Longest Pass, QBR, Rating, Average Rush, and Longest Rush). I also spent more time fitting the model to improve the scoring. At this point I wanted to find the two best models so I chose the Bayesian Linear Regression and Neural Network Regression using the Binning Normalizer. The Decision Forest Regression model performed better than the Neural Network Regression model, but I chose not to use it due to the high Negative Log Likelihood score. The next step was to create a model that combined both the Bayesian Linear Regression and Neural Network Regression models. Results: After exporting the trained model details to csv files, I calculated the average score between each model per player. Additionally I differentiated the average score of both models by home and away games per player.

2018 QB Fantasy Point Predictions

0

2018-qb-fantasy-point-predictions-1936a4efdc24

2018-07-16

2018-07-16 21:06:47

https://medium.com/s/story/2018-qb-fantasy-point-predictions-1936a4efdc24

false

383

null

Machine Learning

machine-learning

Machine Learning

51,320

jorjon21

null

1b1de611ea26

jorjon21

1

11

20,181,104

null

0

null

0

null

2018-09-04

2018-09-04 17:46:02

2018-09-04

2018-09-04 18:41:29

1

false

en

2018-09-04

2018-09-04 18:41:29

1

193708c89a6d

2.713208

9

0

WHY WE COLLECT AND SHARE DATA

5

Factual’s Data Privacy Manifesto WHY WE COLLECT AND SHARE DATA Factual aims to democratize high-quality location data while protecting consumers’ privacy. In an information economy, data is the most valuable resource, but it is not widely available. Rather, it is increasingly siloed and hoarded within mighty walled gardens. Data, as a resource, is far more valuable than oil ever was. Companies like Google, Facebook and Amazon have been incredibly adept at getting consumers to share personal data — and now they own the most data. For most global companies, the world’s most valuable resource is nearly impossible to access. If data doesn’t become more accessible, we may end up living in a world where most of us don’t work for these giants and thus have zero potential to contribute. If we are not careful, this situation could turn into a global calamity. More than 10 years ago when I left Google to start Factual, few people understood the biggest data companies might one day not just dominate your online attention, but also dominate other verticals like auto, television, music, retail, payments, and yes, government. A decade later, this idea is more commonly understood. Yet most of that domination beyond “digital” into all other realms of your life has just begun. Today, few people understand the second-order effects of this future where a data oligopoly dominates innovation. If nothing is done, I believe small businesses will be decimated. That’s why our company makes location data available to other businesses so that they can use it for new waves of innovation. I firmly believe solving the access problem to high-quality “factual” data is the single most significant opportunity to ensure a continued pace of innovation far out into the future. HOW WE WILL PROTECT DATA We will only realize the benefits of location data if there is a foundation of trust. It goes without saying that data services need to be legally compliant, but we need to go beyond that and consider the end user’s interests. At Factual, we guide our data policies by: Supporting our mission to democratize access to data so that innovation can be widespread, not just in the hands of the current data oligopoly. Working to mitigate privacy risks for consumers, protecting individual consumers from negative impacts, to the best of our ability. Providing visibility and choice to consumers who decide that they don’t want to be targeted with advertising. Building upon our brand as the trusted steward of the world’s data. Within this framework, there are many judgment calls where we may have to assess various categories of risk. When assessing our commercial relationships, we consider a number of factors to identify categories of risk. We then ask ourselves a series of questions that help mitigate potential risks, which may include: Is any personal data involved? Will any such personal data be anonymized or pseudonymized before sharing? Will personal data be aggregated? Do affiliates have appropriate security technologies and policies in place to protect data from intentional hacking or accidental leakage? Will usage patterns lead to a situation in which pseudonymous data can be re-associated with PII? Has the recipient contractually agreed to comply with local laws and regulations? THE NEXT DECADE It won’t always be easy to fulfill our mission and potential while safeguarding data, but we must. The cost of not accomplishing this objective is too great. Many data firms we know of today will find exits and sell into some other walled garden, and many others will die or limp along. It is not a foregone conclusion that a location data company will become large and independent, ensuring that a broad set of companies can innovate. If we can build one, the innovation that will ensue during the next decade will be staggering. We’ll live in a world where nearly any kind of business can compete — and not just the data oligopoly. Compared to the alternative that we are currently facing, it’s the only option.

Factual’s Data Privacy Manifesto

76

factuals-data-privacy-manifesto-193708c89a6d

2018-09-04

2018-09-04 18:41:29

https://medium.com/s/story/factuals-data-privacy-manifesto-193708c89a6d

false

666

null

Privacy

privacy

Privacy

23,226

Gil Elbaz

Founder/CEO of Factual, a neutral data company making data accessible to everyone. Previously founder of Applied Semantics (sold to Google). Angel Investor.

6532126a69f2

gilelbaz

1,939

771

20,181,104

null

0

null

0

null

2018-06-30

2018-06-30 14:21:12

2018-06-30

2018-06-30 15:03:16

1

false

en

2018-06-30

2018-06-30 15:03:16

0

1937b0e0975

2.713208

0

Why I want to be a data scientist

5

How My 1st Grade Self Taught Me About Who I Am me in front of my computer at 5/6/7??? idk Imagine first-grade lunch. You’re in a classroom and your teacher is enjoying her lunch with another teacher while she watches over the class. You see all your classmates doing different activities from coloring, eating food, sharing food, practicing their handwriting, playing with blocks, or playing ‘pretend’ with the toys in front of them. Lunch also meant that recess would follow after eating and you eat your lunch pretty fast so you can get to playing. However, you noticed that one of your classmates is walking around the classroom looking for something. It doesn’t seem like the teacher is going to ask him what he’s looking for, everyone isn’t noticing him except for you. You decide to ask him what’s wrong. “ I can’t find my lunchbox and I’m really hungry,” your classmate says while fiddling with his fingers and looking down as if he’s defeated. My heart starts to race, THIS IS MY CHANCE. I can help him find his lunchbox. I was feeling so excited that I would get this opportunity to solve a problem. I put my arm around him and I say, “I can help you!” As we are walking around the classroom, I start to lecture on the importance of keeping your belongings in sight because you never know when you’re going to lose it. In addition, I let him know if there are any problems he needs addressed I am the girl he can count on. AHH yes, the lunch box. “Here it is!” As those words left my lips, the boy took his lunch and ran to his friends. I remember the feeling it felt to solve a problem, and I remember why I lectured the boy. I know what I wanted to convey but at such a young age I didn’t know how to convey it. Now that I am older, thinking about that memory and how I’ve grown in life has made me realize that I’ve always wanted to be a problem-solver. In my family, I am the daughter that has all the technological answers. In my friend group, I always know how to get out of a situation. In a group class project, I always have the ideas to get us forward as a team. In 2016, I acquired an internship where I was a project manager for a summer county program. When I was in high school, I worked out a lot and I loved helping others. I remember many people asking me why I haven’t become a trainer yet? (It might have to do with my age and lack of professional training) In all situations, I want to answer questions. I always think about how I can do better, how can I grow, and where would this take me if I did it. This is why I am so attracted to data science. It involves answering questions. Data science to me is working on a solution to better a company whether they want to attract more sales, attention, or grow as a unit. I want to be the one to give them that answer. If I could give advice to my 18-year-old self when I first started college (I was confused, undecided), I would tell myself to look back at what I’ve done. Who was I when I was younger and what did I accomplish in high school? Peeling all those layers, what key characteristic stuck out the most? I learned this about myself at 20/21 years old. It took me a year to figure out. I am still young but it would have probably saved me three years of confusion. Also, I am still getting to know myself today and what actions are making a better tomorrow. If I can give anyone advice, I would tell them to look from what you’ve done and what’s made your heart race with contentment. That is probably what you’re meant to do.

How My 1st Grade Self Taught Me About Who I Am

0

how-my-1st-grade-self-taught-me-about-who-i-am-1937b0e0975

2018-06-30

2018-06-30 15:03:16

https://medium.com/s/story/how-my-1st-grade-self-taught-me-about-who-i-am-1937b0e0975

false

666

null

Self

self

Self

12,438

fsehmma

Rutgers student and lover of knowledge.

20c67f10bfe9

fsehmma

3

6

20,181,104

null

0

null

0

null

2018-06-14

2018-06-14 07:51:38

2018-06-14

2018-06-14 07:51:33

2

true

en

2018-10-01

2018-10-01 12:19:32

4

1937c615e70d

3.655031

2

0

null

5

Expert View: Avoiding the Trap of Single-mode Thinking As more organizations embrace AI, automation and Big Data, how many are mature enough to really develop and leverage these innovative technologies? By Alan Morrison, Sr. Research Fellow, Emerging Tech at PwC Most companies can work with data at a small scale on one-off projects and achieve limited success. But the real question is, how can they work with data at scale on a continual, day-to-day basis? Without continual flows of the right data to the right place, how can they transform their operations and achieve true competitive advantage? Organizations need to start working with data flows, not just in batch mode. Rigid organizations are trapped in single-mode thinking, while their most able competitors are creative, non-linear thinkers who leverage a diversity of views and approaches to achieve business goals. Data is the lifeblood of any sizable organization today, and competitive advantage hinges on the ability of companies to harness machines to develop and act on insights to effect considerable performance improvements and other beneficial outcomes. To uncover those insights, companies need a continual flow of the right data to the right parts of the organization in the right format. Aligning technology with business goals Companies who succeed in their technology projects use knowledge-based decision making in an iterative development process and do not follow the herd. They’re creative, non-linear thinkers. They take the trouble to understand the problem deeply first. They often diagnose problems correctly and know enough to explore new tools when crafting a solution. If they run into trouble, they revisit their thinking and make corrections and refinements. They’re agile in their thinking as well as their actions. They infuse their thinking into the organizational culture. PwC for instance always starts with the human element. Many times, the biggest challenge isn’t the technology — it’s changing the organization and the process in a humanistic, sustainable way. GDPR and Privacy: How much data do you need? We’re all engaged in a struggle to enrich data and use it more effectively. Enriching data implies identifying more and more relationships between people, places and things. Quickly you get into the realm of data privacy. The GDPR imposes some helpful requirements, and we clearly need to be advocates for individuals to be able to control their most sensitive data. But the devil’s always in the details. Take machine vision in cars, for example — Sandra Wachter of Oxford has pointed out that data used by cars to navigate through obstacles and traffic can also be used for to identify specific individuals, via image data that can be considered biometric. How we manage such data continues to be an open question. Market Evolution and spotting technology trends Organizational boundaries are becoming more porous, and there’s more and more collaboration between organizations. We’ve also seen the rise of the gig economy — freelancers or contractors are more in evidence. In some cases, the bulk of the entire organization consists of contractors. In general, we’re just seeing a more fluid environment. IDC describes the online working environment as the Innovation Graph. Companies will need to consider how to position themselves in new roles in this Graph. Companies can morph into new roles this way and do their own boundary crossing in the process. Why companies should pay attention to semantic technology Graphs and how they can represent connections between people, places and things can articulate and scale all our knowledge of how the world works, in a machine readable form. Just think about how powerful those graphs can be. Those articulated connections, the bridge between human and machine knowledge, can be called semantic graphs. To get to scale and business model agility, companies need to create a semantic graph foundation for AI. Semantic graphs might be seen as the parent data structure that can manage all the children, because they allow full contextualization of disparate data types and machine readable articulation of the rich relationships that need to be mined in any organization. Relationship, not relational, data is what allows us to disambiguate and describe each context. Just ask a social media company — what’s more powerful than a graph to describe and better articulate customer relationships and all the segments and subsegments of the markets serving those customers? Just ask any fraud investigator — what’s a more powerful way than graphs to find bad actors? Graphs as the parents, the most articulated data models, can easily incorporate less articulated models such as tables and documents — the children. That’s how large-scale integration happens. Business users need to think in terms of graphs much more often, and not just in terms of tables. Companies are hobbling themselves if they can’t get beyond tabular data models. Graphs hold the power of large-scale integration. For companies looking to get into Immersive technologies such as VR/AR/MR/XR our Virtual Reality Consultancy services offer guidance and support on how best to incorporate these into your brand strategy. Alan Morrison is a keynote speaker at SEMANTiCS 2018. His keynote will address how technology innovation and especially the adoption of AI impacts new business models. Originally published at Tech Trends.

Expert View: Avoiding the Trap of Single-mode Thinking

3

expert-view-avoiding-the-trap-of-single-mode-thinking-1937c615e70d

2018-10-01

2018-10-01 14:15:40

https://medium.com/s/story/expert-view-avoiding-the-trap-of-single-mode-thinking-1937c615e70d

false

867

null

Big Data

big-data

Big Data

24,602

Alice Bonasio

Technology writer for FastCo, Quartz, The Next Web, Ars Technica, Wired + more. Consultant specializing in VR #MixedReality and Strategic Communications

1663ada6f01e

alicebonasio

7,865

608

20,181,104

null

0

null

0

null

2017-12-27

2017-12-27 09:00:47

2017-12-27

2017-12-27 09:14:20

1

false

en

2017-12-27

2017-12-27 09:14:20

4

1938aade20b9

1.630189

4

0

With reference to my previous blog, I had earlier posted it as an update equation for mathematical optimization. It was the Newton-Raphson…

5

adam - momentum + y (aka. cost) terms With reference to my previous blog, I had earlier posted it as an update equation for mathematical optimization. It was the Newton-Raphson method for finding roots of an equation. I thought this method mostly applies for minimization in machine learning as cost is always defined as a positive real valued function. But it was pointed out to me that the update equation of newton-raphson method, which is x = x - y / dy_dx; where dy_dx = derivative of y wrt. x is unstable at local minima (where dy_dx = 0) since it makes the update burst to infinity. I kept playing with this equation more since it gave me the smoothest curve for what I had been trying to do here, and, I was firm on the idea that the update equation must include the cost term (the Einstein’s space-time analogy: The dimension of cost is connected with all the parameter dimensions). Eventually, I landed on this update equation, x = x - ((y * dy_dx) / (y + dy_dx²)) link to code here. The work is found in the last section of the notebook: “Equation 2: modified from newton-raphson method”. To relate this update equation with the title: if we consider the update portion of the equation as a separate function, g(x, y) = (y * x) / (y + x2) ; y >= 0 It is quite similar to adam optimization update, since there is a squared gradient term in the denominator and the gradient term in the numerator. This equation doesn’t use momentum (running averages) and, the learning_rate is replaced by cost (y). Now, in Adam, the alpha is initially manually set and is utilised as per the decay equation: lr_t <- learning_rate * sqrt(1- exp(beta2, t)) / (1- exp(beta1, t)). With the equation that I have mentioned, the hypothesis is that this decay is kind of estimating the cost term itself. Link to the 3d curve of the g(x, y) function is here. Please let me know what you think about this hypothesis and what it’s implications are. I would be thankful and highly grateful if you could point me to some more relevant research so that I can move forward with this. Thank you!

adam - momentum + y (aka. cost) terms

75

adam-momentum-y-aka-cost-terms-1938aade20b9

2018-04-17

2018-04-17 12:18:44

https://medium.com/s/story/adam-momentum-y-aka-cost-terms-1938aade20b9

false

379

null

Machine Learning

machine-learning

Machine Learning

51,320

Animesh Karnewar

Aspiring AI && ML && DL && RL expert | Programmer | Movie lover | Gamer | Steam id: NranikC (hit me up to play Conan Exiles & GTA V)

f710c72857dd

animeshsk3

198

5

20,181,104

null

0

null

0

e776be3326fb

2017-10-04

2017-10-04 13:54:17

2017-10-04

2017-10-04 13:56:11

1

false

en

2017-12-29

2017-12-29 09:56:33

5

19390d375b3

3.566038

3

0

Introduction

5

Introducing rorodata: Making data science work for you in production Introduction If you are a data scientist building machine learning applications for cataloging, marketing or fraud then you will be intimately familiar with Scikit-learn, Keras, Tensorflow (TF), Jupyter, etc. But there is a good chance that you are not familiar with a lot server software engineering and infrastructure code. As a result, productizing your applications becomes challenging. Of course, you can figure these things out, but the time it takes to learn and build software system is far from trivial. One solution to this problem is creating an assembly line of data scientists and developers. Data scientists build the models and hand it over to developers who further integrate them into “production systems.” But data science does not lend itself well to this approach. Why? Because data science is iterative, with far too many explore, transform, visualize and model cycles before a solution is selected. Getting this data scientist-developer hand-off right is tricky and expensive. Suppose you do manage to get this hand-off to work, you will still a have an unhappy bunch of developers who always need to fix someone else’s work. On the other side, data scientists are unhappy too because their models are not in production fast enough. We have experienced these issues first hand and completely understand the pain in getting machine learning applications to work seamlessly. A better solution is to build a data science platform that data scientists can use, with minimal handoffs with Development. This way, the data scientists control their machine learning applications all the way into production. Under this approach, Data Scientists work on creating production-grade APIs and/or dashboards based on ML models, while Developers work on building a platform which abstracts away all the complex, low-level wiring and configurations that are required to productionize the ML models. A good data science platform needs to be able to: Bring in structured and unstructured data of any size and any format, for exploratory analysis Build, train machine learning models on the desired hardware Manage and track experiments and models, along with all metadata for audits Deploy models as consumable endpoints/APIs/Dashboards, for prediction tasks Track the model performance and roll-out upgrades on the fly This approach can significantly improve the velocity in machine learning application development, allowing companies to capture more opportunities and give them greater competitive edge in the market. …so as a business, no matter what your size, your agility — your ability to roll out a new product, change processes, manage your people, etc. — is equivalent to your ability to develop and change software. So your software development velocity determines your competitiveness. — James Lindenbaum While working with startups and enterprises, we found some critical gaps and missing features when it came to data science platforms — while there are a variety of tools in the market and in the data science open source ecosystems for beginners and big businesses, there is a large gulf in the middle, i.e., in the small and medium business (read small and medium sized teams) space. Specifically, there is not much out there for small data science teams who are keen on building vertical applications fast, and on a budget. Most of these companies/teams do not have enough resources or the time to build such platforms in-house, and they prefer (rightfully) not to do so. Today, we’re excited to introduce rorodata — a data science platform that lets you explore, build, and deploy machine learning models in minutes. You focus on the science e.g. feature engineering, models, etc. and leave the non-science part e.g. infrastructure, devops, experiment management, book keeping, etc. to us. We are fans of Heroku, so we wanted something like a Heroku for Data Scientists. We want data scientists to have the same developer experience as web developers have on Heroku. We will focus on solving problems that simplify and automate all the non-data science tasks that data scientists have to wrestle with, and make data scientists self-sufficient all the way into production. We’re excited by the possibilities and look forward to innovative machine learning solutions. Using rorodata, you can do the following: Standardize your project workflows Deploy your model instantly on AWS Version your model artifacts and API endpoints Run Jupyter Notebooks on demand, on specific hardware Access all the execution logs Provision additional storage using a simple API We support all the popular python frameworks such as Scikit-learn, Keras, Tensorflow, PyTorch. If you don’t find something just ask us and we’ll make it available. We are still in early stages, and actively looking for beta users who can try it and share feedback, report back bugs, problems or ask more questions. We can make machine learning simpler with your help and support. Join us on Slack. If you are a startup and need help with machine learning, we’re more than happy to assist you. Just share some more details with us, and we will get back soon. Lastly, we’re hiring as well. We are building a robust data science platform that gives data scientists wings, by taking on challanging software engineering and design problems. We are looking for highly motivated individuals, who can help us tame these difficult problems. Please look up our careers page for more details.

Introducing rorodata: Making data science work for you in production

3

introducing-rorodata-making-data-science-work-for-you-in-production-19390d375b3

2017-12-29

2017-12-29 09:56:34

https://medium.com/s/story/introducing-rorodata-making-data-science-work-for-you-in-production-19390d375b3

false

892

PaaS designed for Data Scientists

null

rorodata

raghav@rorodata.com

rorodata

DEEP LEARNING,ARTIFICIAL INTELLIGENCE,MACHINE LEARNING,DATA SCIENCE,MACHINE INTELLIGENCE

rorodata

Data Science

data-science

Data Science

33,617

rorodata

PaaS, designed for Data Scientists

e84c1e151445

rorodata

15

1

20,181,104

null

0

null

0

a44b964e0ce9

2018-07-02

2018-07-02 08:50:29

2018-07-02

2018-07-02 09:12:01

6

false

en

2018-07-12

2018-07-12 08:33:24

4

1939a21b6bbc

4.195283

12

0

This is the first in a series of articles which will bring focus to our partners and the ways in which they envision using Fragments…

4

Case Study: Fragments and Jaguar Vision This is the first in a series of articles which will bring focus to our partners and the ways in which they envision using Fragments platform. These case studies are intended to introduce the companies that we are partnered with, describe the real life applications they will use Fragments for, and bring inspiration to others as to the many diverse ways in which the platform might be used. To kick things off, we’re very pleased to present Whale & Jaguar, a machine learning startup based in Colombia. For them, the Fragments platform will be applied in order to train and improve various products, such as their ‘Jaguar Vision’, using a global team of annotators. The following is a guest post by Whale & Jaguar: Jaguar Vision: An algorithm for understanding people’s way of seeing. We’re a multidisciplinary team merging the investigation of science and technology with digital communication strategies. We’re physicists, political scientists and historians interested in data, images and the human behavior which derives meaning from information. That’s why we work on the development of technology capable of seeing, reading and analysing the language we use to communicate on digital networks. Male jaguar killed in the Sierra Nevada de Santa Marta Visual content exists alongside its textual descriptor. As a way of communication, it’s deconstructable and measurable. We want to understand people’s ways of seeing and to do so we train artificial intelligence algorithms that study and classify images at massive scale. We are dedicated to extracting knowledge from data, but how? To turn the machine into a connoisseur of the shape of things which can be seen, we show it the steps a human takes. We feed images from a visual database, each one labeled with tags associated with its content (“mountain”, “house”, “river flowing”) and select the most relevant tags for each case. Then, we ask for its composition. How are objects distributed across the space defined by the image? Centered, placed on a corner, randomly set, on a quadrant. An algorithm capable of recognizing these characteristics in a visual object could join digital network analytics, capable of measuring the sentiment and opinion of the audience, the effectiveness of the visible, and the patterns of visual communication proper of an entity, to give us a better understanding of the way we relate to the Internet’s visual environment. As part of the process, we will train a convolutional neural network (CNN), one of the most efficient machine learning algorithms for image classification. A CNN works by mimicking the response of neurons to visual stimuli, and by doing so, recognizes visual patterns shared by a group of images and learns the shapes of their elements. This is accomplished by reconnaissance units called filters: hundreds, thousands or even millions of them. Every filter should be trained to recognize certain particularities within a region of the image: corners, legs, feet, noses, faces, water, etc. When identified together they give rise to the accurate identification of the image content: dog, cat, human, mountain, house, river flowing. Training our neural network models requires a large set of images tagged according to their content — this is where our friends at Fragments with their platform for micro-task apps play a crucial role. A micro-task app on the Fragments platform will enable us to engage a scalable human workforce in both collection and annotation of our large data sets. Using the Fragments platform, we will be able to split data set collection and labeling into a large amount of micro-tasks and distribute them among a huge quantity of workers, in order to quickly obtain structured and labeled data. Workers' answers to questions such as “Is this a cherry?” or “How are the elements distributed across the image?” will give us a better understanding of people’s way of seeing. The first application of Jaguar Vision will occur in digital networks. We’ll study content produced by multiple entities, corresponding to different segments of population, and find visual patterns to which the audience reacts with positive, neutral and/or negative sentiment. Thus, we can determine whether there are relationships between the visual training of an audience and its response to certain types of published image. We want Jaguar Vision to allow us to study both shared and exclusive features in specific sets of images, such as the Gráficas Molinari image archive, that require special treatment and an investigative point of view in order to identify, for example, how women are portrayed and defined through visual characteristics from the mid twentieth century. We intend to compare the contents of such image sets with textual sources digitized by the largest libraries in our country, thus very efficiently achieving analysis of the narrative of our history through cultural heritage objects. Our founding parable accompanies this research. In the jungle of advanced information and the vast ocean of data, two mammals have the natural instincts to survive and succeed: jaguars and whales. The jaguar gives us speed and accuracy; the whale, depth and immersion. The merging of both natures leads us to innovation, and willing to make the jaguar as agile as the whale to traverse seas loaded with information, Jaguar Vision is born.

Case Study: Fragments and Jaguar Vision

331

case-study-fragments-and-jaguar-vision-1939a21b6bbc

2018-07-12

2018-07-12 08:33:24

https://medium.com/s/story/case-study-fragments-and-jaguar-vision-1939a21b6bbc

false

860

Decentralised Micro-tasks Platform

null

fragments.network

null

Fragments

hello@fragments.network

fragments-network

null

fragments

Machine Learning

machine-learning

Machine Learning

51,320

Petr Sigut

co-founder @ https://fragments.network/

1902074379f

petr_sigut

13

4

20,181,104

null

0

null

0

null

2018-01-10

2018-01-10 07:02:18

2018-01-10

2018-01-10 07:02:41

0

false

en

2018-01-10

2018-01-10 07:02:41

0

193a9b2c543

0.69434

0

Jeff Dean and company at Google released a mind-blowing paper a week ago.

1

인덱싱 알고리즘을 대체하는 딥러닝 Jeff Dean and company at Google released a mind-blowing paper a week ago. They used neural networks to replace core indexing algorithms like B-Trees, Hash Indexing, and Bloom Filters with trained models. Compared to a B-Tree, the trained models searches on the order of 3 times faster and 10–100x better in terms of space consumption in some configurations. (There are a bunch of ways to configure both algorithms so there’s no single comparison). All of this stuff executes on GPUs — which are actually getting way faster (some claim 1000x by 2025 [citation in paper]) — whereas CPU execution speeds are flat. (Edit: These would execute very well on GPUs, but the paper’s experiments are on CPUs. Jeff Dean clarified that on Twitter) I’m pretty new following AI stuff, but this feels like one of those foundational papers that changes everything. These techniques are going to be used to move more and more compute to GPUs and change the way we think about software construction. Dean’s presentation at NIPS 2017 (aside: wow, super unfortunate acronym) also awesome.

인덱싱 알고리즘을 대체하는 딥러닝

0

인덱싱-알고리즘을-대체하는-딥러닝-193a9b2c543

2018-01-10

2018-01-10 07:02:41

https://medium.com/s/story/인덱싱-알고리즘을-대체하는-딥러닝-193a9b2c543

false

184

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Jeemyeong

가공되지 않은 자료들이 많습니다. 공부하는 중에 급하게 필기하는 내용 및 링크를 저장해두는 창고입니다.😅

c1256d13b774

jeemyeong

40

54

20,181,104

null

0

null

0

null

2018-05-28

2018-05-28 06:18:40

2018-05-28

2018-05-28 06:31:13

1

false

en

2018-05-28

2018-05-28 06:31:13

3

193af0cb13bd

2.516981

0

One of the highlights of last year’s Google I/O was the announcement of Google’s Automated Machine Learning or AutoML which according to…

5

All you need to know about Google AutoML One of the highlights of last year’s Google I/O was the announcement of Google’s Automated Machine Learning or AutoML which according to founder Sundar Pichai was their ‘AI Inception’. The Machine learning Cloud software suite is finally about to hit the alpha stage and as of now, it seems like there is a lot that developer and designers alike can leverage from this tool. So let us explore this new addition to the Google family. Google AutoML — GoodWorkLabs What is AutoML? AutoML is Google’s Cloud software suite for Machine Learning tools based on Google’s Neural Architecture Search (NAS). Using AutoML a user can train deep networks without having any expertise in deep learning or Artificial Intelligence. This is achieved by leveraging NAS to derive the most suitable data network for the task at hand. This is a huge step for businesses as they will be able to harness the power of machine learning without using experts with years of experience in any field of automation. Customization of various AI product models meant for mundane as well as complex tasks. How Will Businesses be able to Leverage AutoML? Today, with deep learning still being in its infancy, most business use, AI for simplistic tasks. And considering the fact that AI and Machine learning experts are hard to come by and are quite expensive to hire, most companies try to keep their deep learning efforts to a minimum. This is where AutoML makes a world of difference. One of the most common applications of deep learning for businesses today is Image processing. Today companies rely on Image classification networks that run on neural matrices. This although running on machine learning tools, requires quite a bit of manual effort to isolate the proper network for the image data sets and then train the models to adapt to the task they intend to accomplish. With AutoML companies can skip the tedious processes of research and get directly into the transfer process. As a matter of fact, Google’s first AutoML tool is intended for exactly this aspect for visual companies. AutoML Vision This is AutoML’s first tool which will help developers create custom image recognition models. According to Google this tool will allow developers to upload their labeled data and the system will automatically create a custom machine learning model for them. This, in turn, will allow them to focus more on product development, design and other important aspects. Although still in its alpha stage, this tool is being used by many big-time players such as the animation giant Disney. Conclusion As the concept of ‘Software 2.0’ catches on, developers around the world are getting used to the new tools largely built upon their predecessors, to write software without actually having to code each and every part of it. Tools like AutoML are a good example of how this is being applied to every aspect of software irrespective of the purpose. With AI the scope is further enhanced. Yet, for such tools to achieve any degree of comprehensiveness, they themselves need to be further developed to accommodate the needs of the upcoming generations of developers who will be relying upon them to do most of the work. So the road ahead will be the chicken and egg paradox where the efforts of both the man and the machine will craft the future. Today we have that sense of direction as to where we need to focus our efforts in terms of research, which is great news for AI tools such as AutoML. *This blog has been originally published on the GoodWorkLabs website. Visit the blog for more articles on AI and Machine Learning.

All you need to know about Google AutoML

0

all-you-need-to-know-about-google-automl-193af0cb13bd

2018-05-28

2018-05-28 06:31:14

https://medium.com/s/story/all-you-need-to-know-about-google-automl-193af0cb13bd

false

614

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

GoodWorkLabs

We are a UX design and Software development company and are passionate about creating cutting-edge technologies for businesses.

21fdd225f065

GoodWorkLabs

43

29

20,181,104

null

0

null

0

null

2018-09-08

2018-09-08 13:40:52

2018-09-08

2018-09-08 13:51:07

2

false

en

2018-09-08

2018-09-08 13:51:07

5

193b2b080a6b

2.024843

0

I arrived at Nicholas Carrr’s book this July through having read Neil Postman’s “Amusing Ourselves to Death” (1985) earlier in the summer…

5

“The Shallows: How the internet is changing the way we read, think and remember” — Nicholas Carr (2010) I arrived at Nicholas Carrr’s book this July through having read Neil Postman’s “Amusing Ourselves to Death” (1985) earlier in the summer. Just as Postman’s book critiqued the transformative nature television was having on human culture and the human brain, so Carr’s book tackles similar themes in relation to the internet 25 years on. Even more disturbing than Postman’s title, it seems an extension and an acceleration of some the themes described in that book. Carr looks at how the internet transforms the brain’s perception of reality — as significant tools always have — for instance the map transformed human perception of space, and the clock transformed human perception of time. We don’t simply use tools to facilitate what we do- the medium of the tool transforms our perception of reality and consequently our concept of what it is that we should do. We don’t simply use tools, rather tools also use us. Key to the transformative potential the internet has on the brain is brain plasticity. Only in the 20thcentury did scientists begin to really understand that the brain is not fixed matter. The book quotes British biologist J.Z. Young in 1950; “There is evidence that the cells of our brains literally develop and grow bigger with use, and atophy or waste away with disuse. It may be therefore that every action leaves some permanent print upon the nervous tissue”. An experiment by Juan Pascual-Leone showed that neurologically we become what we think. He taught a group of people with no experience of piano a simple melody; then for five days he had half the group practice playing the melody and half the group simply to imagine playing the melody, without even touching the keys. The brains of the participants were mapped and found that the neurological changes in those imagining the action were the same as those actually playing piano. Weizenbaum in Berlin, 2005 Chapter 10 discusses Joseph Weizenbaum (1923–2008); one of the fathers of Artificial Intelligence, Wiezenbaum soon came to be disturbed by the nature of interaction between humans and machines (see the ELIZA Program) and by the 1970s had become an opponent of AI. His view was that what is most human about us is that which is least calculable, and that this element of ourselves is eroded as we merge our neurological processes with those of machines. A fairly disturbing book, it persuaded me of a need to identify in the calendar “blackout” periods each day when not to use the internet in any form.

“The Shallows: How the internet is changing the way we read, think and remember” — Nicholas Carr…

0

the-shallows-how-the-internet-is-changing-the-way-we-read-think-and-remember-nicholas-carr-193b2b080a6b

2018-09-08

2018-09-08 13:51:07

https://medium.com/s/story/the-shallows-how-the-internet-is-changing-the-way-we-read-think-and-remember-nicholas-carr-193b2b080a6b

false

435

null

Tech

tech

Tech

142,368

Adam Reynolds

The World of Books and Knowledge . . .

cb38001686ae

adam.reynolds568

1

20,181,104

null

0

null

0

null

2017-11-14

2017-11-14 19:38:13

2017-11-14

2017-11-14 19:42:29

2

false

en

2017-11-14

2017-11-14 19:48:04

0

193cc514937f

0.598428

0

As Healthcare joins hands with AI, it calls for the beginning of an interesting game. With Artificial Intelligence starting to find…

4

How is Artificial Intelligence disrupting the Healthcare sector? As Healthcare joins hands with AI, it calls for the beginning of an interesting game. With Artificial Intelligence starting to find application areas in healthcare & revolutionizing the design of treatment plans through the assistance in repetitive jobs to medication management or drug creation, it is only the beginning!

How is Artificial Intelligence disrupting the Healthcare sector?

0

how-is-artificial-intelligence-disrupting-the-healthcare-sector-193cc514937f

2017-11-14

2017-11-14 19:48:04

https://medium.com/s/story/how-is-artificial-intelligence-disrupting-the-healthcare-sector-193cc514937f

false

57

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Phronesis Partners Gist

We partner with clients to ‘simplify growth’ by leveraging our research and intelligence capabilities. Write to us at: info@phronesis-partners.com

2e3abb7b217f

Phronesis_inc

66

134

20,181,104

null

0

null

0

null

2017-12-20

2017-12-20 22:08:04

2017-12-20

2017-12-20 22:09:38

1

false

en

2017-12-20

2017-12-20 22:09:38

2

193d5683a48f

0.815094

1

0

I hope you’re having a wonderful holiday season.

5

Meet Heather, a Messenger Chatbot I hope you’re having a wonderful holiday season. It’s no secret that artificial intelligence (AI) and Messaging bots are set to disrupt just about every industry imaginable as technology becomes more agile, smarter, and companies look to improve efficiencies by automating sales, marketing, and customer service processes. Personally, I’ve spent the better part of this year studying where AI and bot technology is headed and am excited by what’s coming which is why Gil Media Co. is actively consulting with and building Messaging bots for our clients. Now, here’s the reveal… Our team of developers has created a custom-tailored experience which I’d like to ask you to beta test at https://www.messenger.com/t/GilMediaCo Introduce yourself to Heather, I promise you’re in good hands :) If you’re interested in having Gil Media Co. build a chatbot for you or if you’re curious how a chatbot can transform your business, contact us at gilmedia.co/contact or comment below. Exciting times await! Carlos

Meet Heather, a Messenger Chatbot

10

meet-heather-a-messenger-chatbot-193d5683a48f

2018-03-30

2018-03-30 05:38:52

https://medium.com/s/story/meet-heather-a-messenger-chatbot-193d5683a48f

false

163

null

Bots

bots

Bots

14,158

Carlos Gil

Entrepreneur, Gil Media Co. Award-winning Snapchat Storyteller and Keynote Speaker.

c382cb868ea4

CarlosGil

8,292

11,820

20,181,104

null

0

null

0

4e8922a89498

2017-10-04

2017-10-04 21:36:54

2017-10-09

2017-10-09 11:46:00

16

false

en

2018-01-10

2018-01-10 19:50:10

3

19427eafecb2

4.131132

18

1

0

That too, for free!

5

Give your product AI powered chatbot That too, for free! Google Assistant works on Natural Language Processing As a part of my exploration about AI, I set my goal to share the best tech that I learn. In this blog post, I will build a bot from scratch that will help you enable Natural Language Processing just by using an Excel Sheet and a Cloud Service by IBM Watson; Natural Language Classifier. I’ll explain how I did it, so you can build your own. Natural language Processing is a way for computers to analyze, understand, and derive meaning from human language in a smart and useful way. In simple words, NLP minimizes the language barrier between humans and machines exponentially. It’s not as complicated as it might sound, this tutorial will make things way easier. We’ll build a “Coffee Bot”. Topic too lame right? The fact, this topic is so generic that even a newbie can ramp up quick. Rest assured, it can be used over any topic of your choice. To start, open an excel sheet. Focus on column A. Now, think from a users perspective (Your end users) and write one question they are bound to ask to your bot. For Eg: “how are you doing” Now, assign a Class Name “doing_well” in column B to that respective question. Class Name can be assigned in whatever way you want. There are many different ways in which a user can ask this same question. In order to make your bot understand the variation in each question and still have the power of figuring out the meaning, we need to train the bot. So, replicate “how are you doing” to atleast 5 or more similar meaning questions. The more, the better. The question “how are you doing” is replicated to more than 5 similar questions. Note: all these questions belong to the same Class “doing_well” In this Microsoft Excel Sheet, Column A represent the questions and Column B represent Classes. That’s all that we need. Bingo! We have reached our first checkpoint. I know that sounds silly but here’s the fact, this Excel Sheet will serve as our “actual” database. This is just a part of Coffee Bot Corpus (database) to get a basic idea. As we know, this sheet is our Database, we further need to expand the capabilities of the bot and make it smarter by adding more features to it. I’ll repeat the same steps and create more classes. Each class will represent a feature and it’s unique in it’s own way. For better understanding, I have a few listed down. Don’t forget, you can always enhance the quality of the Corpus (Database) by adding more classes (Features) to the existing database. Make sure you strictly keep them unique to not confuse the bot. Once you have your Excel sheet ready with enough classes and questions to support the needs of your bot/product, next step is to train it. To train the database that you created, it strictly has to be saved in .csv format. IBM Watson NLC (Natural Language Classifier) doesn't support any other format. Checkpoint 2 cleared! Follow these 10 steps to train the database that you created and bring it to life. Step: 1 “IBM Natural Language Classifier” Step: 2 Click “Get started free” Step: 3 Create a free account Step: 4 Choose a name for your classifier and click create. Step 5: Keep your service credentials confidential. This username/password helps you keep your Classifier unique. Step: 6 Access the beta toolkit Click sign in Click confirm authorization Step: 7 Add training data Click upload training data Select the .csv file Create Classifier Step: 8 Wait until it gets trained. Step: 9 This is the final product with a unique classifier ID. Bot ALIVE! Click the blue arrow to access the classifier. Step: 10 Ask Questions. This questions isn’t at all in the database but it still responded with 0.72 (72%) confidence level. Confidence level above 90% for questions that didn't exist in the database. Source Keep testing and adding more questions and classes to the database. Make your bot smarter! You can integrate this NLP bot with Google Home/Assistant, Alexa, Facebook Messenger, Siri, Stride, Slack or any other smart device/software. Use the IBM Watson API Endpoint and the Classifier Credentials to connect it with your already exisiting product and you’re good to go! Feel free to leave feedback or questions in the comments below. 😃 Note: You can replicate this same steps for a different topic. If you would like to know more about my work with AI/ML check out: https://www.archie.ai

Give your product AI powered chatbot

376

give-your-product-ai-powered-wings-19427eafecb2

2018-06-12

2018-06-12 19:41:49

https://medium.com/s/story/give-your-product-ai-powered-wings-19427eafecb2

false

684

ML & Tech Articles from the team behind Archie.AI

null

ekplugin

null

Archie.AI

ishtiaq@alumni.utoronto.ca

archieai

ARTIFICIAL INTELLIGENCE,MACHINE LEARNING,GOOGLE ANALYTICS,DEEP LEARNING,REINFORCEMENT LEARNING

ekplugin

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Bhargav Shah

Featured on LinkedIn TalentConnet | World Champion at 13th UCMAS International Competition | Machine Learning Engineer

b038cbb27c77

Phantomsf

148

39

20,181,104

null

0

null

0

null

2018-04-24

2018-04-24 16:19:34

2018-04-24

2018-04-24 18:26:53

1

false

en

2018-09-08

2018-09-08 10:58:26

4

1943cbfa6d58

3.486792

4

0

Being selfish has a negative connotation in our society.

5

Selfishness Is A Good Thing — Even If Others Try To Tell You That It’s Not Dennis Sprute Instagram Being selfish has a negative connotation in our society. If it’s done with the intention to harm or use other people, rightfully so. But what if it’s practiced to help us build a better life for ourselves? Apparently, many people around us don’t like that! They are stuck with the old concept of a meaningful life. One of my grandpas had to fight in World War II. The other had to flee from what became the GDR. Their idea of human needs was slightly different than mine. To stick with Abraham Maslow’s famous pyramid of needs, physiological and safety needs (no war, shelter, food) came above everything else. My parents’ generation (baby boomers) prioritized love/belonging (building a family) and esteem (making a career). I, a millennial, was born into a time of peace — and exposed to a vast amount of global opportunities through technical progress; the internet. More than ever before, that allows me to to aspire for more: Self-actualization (founding a startup, doing what I love) — if not transcendence (becoming detached of anything external, enlightenment). In order to get there, you have to be selfish — which not everyone you around may like! Attaining such state is hard — even more if you have to justify yourself. Doing the journey inwards doesn’t come by in passing. It means to practice mindfulness. It can be done passively while doing something else (showering, cooking). But to really make moves, you have to actively do it: This could be via creative expression (writing for me), meditation, keeping a diary, reading books (on spirituality, philosophy), spending time in nature, exchanging with others about certain topics, or taking mind expanding substances. For me it also means to spend your money beyond the obvious “pleasures”: Clothes, cars, vacation, booze. I’ve tried things like acupuncture, psychokinesis, the Israeli Grindberg method, hypnosis, or subconsciousness sessions. (Insurances don’t cover most of these.) Devoting yourself to such things means to invest time and financial resources. Many people don’t want to do that! Yet they make you feel like you’re “leaving them behind” — and make you feel guilty. Why put up with something if you don’t have to? At the same time, nobody likes you get ahead. You’re becomig less dependent and controlable! That’s when the sapient fear triggers: To be left behind and die. People’s state of happiness is often tied to external factors (as mentioned already): Physical things like cars and clothes, emotional things like love and sex, and social factors like status and friendship. They don’t like if you start keeping more and more to yourself — because your happiness is no longer dependent on them. They don’t like that you stay home at night to work on something you care about, don’t join in on a vacation or prefer to eat healthier food when you dine out together. They’re frustrated that you do your own thing (and they don’t), so they start calling you selfish! They try to make you feel bad and call you a bad friend. Worse, some even start adding up and blackmailing. Because they are helpless, people go as far as dropping all of the things that they’ve done for you in the past — or say they won’t do certain things for you anymore. They view relationships as balance sheets. But every new context has a new zero point. And if you’re not feeling something at a certain moment, you should be selfish! (Especially if pleasing others first and you second has made you unhappy in the past!) We’ll soon enter a new era—so we should be prepared anyways. The basic income will come. I have no doubt about that. Humans won’t be able to compete with artificial intelligence in the years and decades to come. We’ll be forced into creating meaning for our days and lives. So why not start preparing for it today? Many people in my college didn’t do any side projects or invest into relationships. They said: “I’ll do that when I’ll actually need it.” I never understood this thinking. I was always preparing for the next step, even if it was unknown. I knew that I could try myself out and make experiences to make the best (or better) choices for when I actually needed it. This makes you trust in yourself. The more I have detached myself in the past couple of weeks, the more fulfilling my life has become. I am, after all, not replacing or avoiding external things. I just choose more wisely what my gut prefers. „Those who rejoice in the self, who are illumined and fully satisfied in the self, for them, there is no duty.“ — Bhagavadgītā, Chapter 3, Verse 17 We should start kicking the old philosophy of older generations. Instead of “asking what we can do for our country” (Kennedy) we should ask what we can do for ourselves! (Even if others won’t adapt to it just yet or make fun of us.) In the end it is my true conviction, that that is how we become better humans, citizens, employees, family members, lovers, children, and parents. — sprute.dennis@gmail.com

Selfishness Is A Good Thing — Even If Others Try To Tell You That It’s Not

141

selfishness-is-a-good-thing-even-if-others-try-to-convince-you-that-it-isnt-1943cbfa6d58

2018-09-08

2018-09-08 10:58:26

https://medium.com/s/story/selfishness-is-a-good-thing-even-if-others-try-to-convince-you-that-it-isnt-1943cbfa6d58

false

871

null

Spiritual Growth

spiritual-growth

Spiritual Growth

5,420

Dennis Sprute

26 y/o millennial in Lisbon. Chose to share his thoughts + reflections on his becoming and (broken) dreams publicly— and not via endless WhatsApp voice memos.

8fbe13e72c2e

DennisSprute

147

5

20,181,104

null

0

null

0

null

2018-07-30

2018-07-30 19:44:19

2018-07-30

2018-07-30 19:47:38

3

false

en

2018-07-30

2018-07-30 19:47:38

22

19442843b780

14.851887

2

1

0

Originally published at gomedici.com

5

Alternative Data & Financial Access: The Good, the Bad, and the Ugly Originally published at gomedici.com The problem & the opportunity The World Bank estimates that about 1.7 billion adults remain unbanked globally — without an account at a financial institution or through a mobile money provider. Virtually all these unbanked adults live in the developing world. Indeed, nearly half live in just seven developing economies: Bangladesh, China, India, Indonesia, Mexico, Nigeria, and Pakistan. About 56% of all unbanked adults are women. The World Bank also found that poorer people account for a disproportionate share of the unbanked — globally, half of the unbanked adults come from the poorest 40% of households within their economy, the other half from the richest 60%. But the pattern varies among economies. On the bright side, the number of unbanked around the world has been steadily declining. The 2017 Global Findex database shows that 1.2 billion adults have obtained an account since 2011, including 515 million since 2014. Between 2014 and 2017, the share of adults who have an account with a financial institution or through a mobile money service rose globally from 62% to 69%. One of the problems with extending financial access is the lens through which the formal financial system assesses previously “invisible” groups of the global population. Lenses vary in different countries, but one is constant — existing frameworks have not been able to effectively expand the addressable market. Financial technologies, however, have a role to play in changing the very framework. The progress in extending access to the formal financial system to previously “invisible” groups of the global population is largely attributed to digital financial technologies. Experts believe that smartphones can dramatically reduce the cost of lending because the apps they run generate huge amounts of data — texts, emails, GPS coordinates, social-media posts, retail receipts, and so on — indicating thousands of subtle patterns of behavior that correlate with repayment or default. The World Bank emphasized that the global spread of mobile phones has facilitated expanding access to financial services to hard-to-reach populations and small businesses at low cost and risk: Digital IDs make it easier than ever before to open an account Digitization of cash-payments is introducing more people to transaction accounts Mobile-based financial services bring convenient access even to remote areas Greater availability of customer data allows providers to design digital financial products that better fit the needs of unbanked individuals “Financial access facilitates day-to-day living and helps families and businesses plan for everything from long-term goals to unexpected emergencies. As account holders, people are more likely to use other financial services, such as credit & insurance, to start & expand businesses, invest in education or health, manage risk, and weather financial shocks, which can improve the overall quality of their lives.”– The World Bank The last point in the list of benefits — the greater availability of consumer data — is particularly important. The hallmarks of one’s lifestyle imprinted in continuous data flow are increasingly becoming vital in innovative ways to assess how trustworthy one is. To trust someone with money in the form of credit, or other financial service(s), financial institutions are required to perform an assessment of one’s history with money. Unable to do so for a large group of the global population using existing frameworks leads to a perpetuated exclusion of potentially mutually beneficial relationships with previously “invisible” individuals, not speaking about unbreakable barriers for building personal prosperity for those individuals. “A limited credit history can create real barriers for consumers looking to access the credit that is often so essential to meaningful opportunity — to get an education, start a business, or buy a house. Further, some of the most economically vulnerable consumers are more likely to be credit invisible.” — Richard Cordray, Former Consumer Financial Protection Bureau (CFPB) Director The limitations in existing underwriting processes have been widely highlighted. DirectID — a solution that combines identity verification, real-time financial data, compliance checks, affordability insights, and ACH payment confirmation into a single platform — emphasizes that in traditional underwriting processes, there is a reliance on credit scores, copies of bank statements, and self-reported income data. But here is why they are falling down: Credit scores provide limited insight into a consumer’s true financial position, they don’t provide the whole picture. With about a third of US consumers having a FICO score under 670, most traditional lenders would not offer loans to individuals with scores that low. But many of these people are creditworthy borrowers. FICO’s data doesn’t help in the assessment of whether they would repay loans — something that new data sources can help to predict more accurately. Copies of bank statements are susceptible to fraud. With banks encouraging their customers to move to online bank statements, obtaining paper copies for application processes can introduce delays, and leads to high drop-out rates. Self-reported income is also susceptible to fraud and does not reflect any change in circumstances which may impact the ability to make repayments in the future. But when 1.7 billion people don’t have a history with the formal financial system, and existing frameworks are largely exclusive, how can institutions reach a balance of accurate risk assessment and continuous inclusive development? What is alternative data? According to Experian, in the consumer financial marketplace, alternative credit data refers to information used to evaluate creditworthiness that is not usually part of a traditional credit report. “To fall under the Fair Credit Reporting Act (FCRA) — compliant umbrella, alternative credit data must be displayable, disputable and correctable. This data provides more insight into both full-file and thin-file consumers, to drive greater visibility and transparency around inquiry and payment behaviors. Adding the information from alternative credit data sources may allow some consumers to gain more access to credit.” — Experian Some examples the agency brings up are: Rental payments Mobile phone payments Cable TV payments Bank account information, such as deposits, withdrawals or transfers Small dollar loans Other types of alternative data might relate to things less closely tied to a person’s financial conduct, like that person’s education or occupation. The use of alternative data for extending financial access: the good (mostly, for businesses) Leveraging data from previously untapped sources can bring tangible opportunities and benefits for financial institutions, a few of which include: Opportunity to capture new customer segment “An ‘unscoreable’ individual is not necessarily a high credit risk — rather they are an unknown credit risk. Many of these individuals pay rent on time and in full each month and could be great candidates for traditional credit. They just don’t have a credit history yet.” — The State of Alternative Credit Data, Experian, 2018 An alternative framework includes the use of alternative sources of data in order to profile a potential customer — web search history, phone usage, social media, and more. Mike Mondelli, SVP, TransUnion, listed property, tax, deed records, checking & debit account management, payday lending information, address stability, and club subscriptions being among the sources for alternative data. “These alternative data sources have proven to accurately score more than 90% of applicants who otherwise would be returned as no-hit or thin-file by traditional models. Alternative data provides a better lens with which to evaluate all consumers, giving lenders who can score them a competitive advantage.” — Mike Mondelli, SVP, TransUnion Experian offers a demonstrative model of how approvals change with an addition of alternative credit data into the assessment framework. Source: The State of Alternative Credit Data, Experian, 2018 The agency shares that by adding in the visibility of alternative credit data to a near-prime population, lenders could see an increase in the approval rate within a population that was historically being left behind. Enriched underwriting process Arun Ramamurthy, Co-founder of Credit Sudhaar, believes alternative sources of data including social media to be an important part of creditworthiness assessment. “(About) 6% of the people who sign up for our advisory services are intentional defaulters and fraudsters.” — Arun Ramamurthy, Co-founder, Credit Sudhaar, said. Ramamurthy shared with The Hindu that it is only through the use of unconventional sources of information can companies and banks identify the intention of a potential borrower. Social media, in particular, has been recognized by Wharton as an important data source for credit scoring back in 2014, although the practice of judging a stranger based on his/her social environment is not really new. One of the core ideas is that “who you know matters.” Technology companies focusing on alternative lending and alternative credit scoring can gather more information about a person using social media than by looking at their financial data. Social media also gives lenders an insight into how an applicant spends their time. “If you look at how many times a person says ‘wasted’ in their profile, it has some value in predicting whether they’re going to repay their debt,” FICO CEO Will Lansing told the Financial Times. Service enhancement; timeliness, and accuracy According to Experian, some kinds of alternative data, such as online bank account information, may allow lenders to automate tasks that are done manually during the loan approval process. This automation could speed up application processes or avoid subjective interpretations that could lead to differences in treatment or wrongful discrimination. Additionally, the agency notes that data traditionally used by lenders often does not reflect a person’s most recent activities. Alternative data could provide more up-to-date, real-time information. “Alternative data could provide more timely indicators, such as real-time access to a consumer’s outstanding credit card balance. It could also help lenders recognize whether a particular consumer’s finances are trending in a particular direction, such as through a job status change appearing on social media. Such information could help to distinguish those consumers whose low scores are a function of prior financial problems that they have surmounted from those consumers whose financial challenges have just begun and who may pose a greater risk than the score indicates. Alternative modeling techniques might also generate more timely feedback to the extent they dynamically change as new data are ingested, though such dynamism could also carry certain risks.” — CFPB Cost-efficiency Paul Christensen, Associate Dean of Executive Education and Clinical Professor of Finance at the Kellogg School of Management, shared insights on alternative scoring, mentioning an important positive implication for companies leveraging alternative data to make a credit decision. “For companies, alternative credit rating is about reducing transaction costs. It’s about figuring out how to make profitable loans that are also affordable for most people — not just business owners.” — Paul Christensen, Kellogg School of Management He also added that it is a way of addressing the problem of information asymmetry, which he calls one of ‘the definitional causes of market failure’ and one of the biggest threats to traditional microfinance. Christensen believes that would make a big difference if alternative credit scoring can help drive down costs and lower interest rates from 30% to 10%. Experian, one of the major consumer credit reporting agencies, also outlines the opportunity to lower the costs with the use of alternative data for creditworthiness assessment. Using alternative data could lower costs for lenders and, in turn, benefit consumers through lower prices, the agency shares. The use of alternative data for extending financial access: the bad, and the ugly (mostly, for consumers) As with anything new, there is always the bad, and the ugly to anticipate and address before jumping into implementation. Discriminatory practices Back in 2016, Lael Brainard, a member of the US Federal Reserve’s Board of Governors where she serves as Chair of the Committees on Financial Stability, Federal Reserve Bank Affairs, Consumer &Community Affairs, and Payments, Clearing & Settlements, shared that, “Nontraditional data, such as the level of education and social media usage, may not necessarily have a broadly agreed upon or empirically established nexus with creditworthiness and may be correlated with characteristics protected by fair lending laws. To the extent that the use of this type of data could result in unfairly disadvantaging some groups of consumers, it requires careful review to ensure legal compliance.” While non-traditional data may have the potential to help evaluate consumers who lack credit histories, some data may raise consumer protection concerns. CFPB also emphasized potential discriminatory implications in the use of alternative data and modeling techniques. “For example, using alternative data that involves categories protected under Federal, State, or local fair lending laws may be overt discrimination. In addition, certain alternative data variables might serve as proxies for certain groups protected by anti-discrimination laws, such as a variable indicating subscription to a magazine exclusively devoted to coverage of women’s health issues. And the use of other alternative data might cause a disproportionately negative impact on a prohibited basis that does not meet a legitimate business need or that could be reasonably achieved by means that are less disparate in their impact.” “Machine learning algorithms that sift through vast amounts of data could unearth variables, or clusters of variables, that predict the consumer’s likelihood of default (or other relevant outcomes) but are also highly correlated with race, ethnicity, sex, or some other basis protected by law. Such correlations are not per se discriminatory but may raise fair lending risks. The use of alternative data and modeling techniques could potentially lead to a disparate impact on the part of a well-intentioned lender as well as allow ill-meaning lenders to intentionally discriminate and hide it behind a curtain of programming code.” — CFPB Transparency However imperfect traditional creditworthiness assessment framework may be, it can’t be criticized for the lack of transparency. The FICO score, for one, has been molded into components and every scored individual can understand what criteria affect their score. “It may not always be readily apparent to consumers, or even to regulators, what specific information is utilized by certain alternative credit scoring systems, how such use impacts a consumer’s ability to secure a loan or its pricing, and what behavioral changes consumers might take to improve their credit access and pricing.” — Lael Brainard With alternative credit scoring frameworks, the situation is not always clear. Let’s take Web search history, for example. Casey Oppenheim, Co-founder & CEO of Disconnect, which helps keep people anonymous online, fairly points out the possibility of negative outcomes the use of such sort of alternative data can have. “Nobody understands the long-term impact of this data collection. Imagine that someone has 40 years of your search history. There is no telling what happens to that data.” — Casey Oppenheim, Disconnect The availability of a lifetime of search history traps one in the outcome of the ‘mistakes of youth’ — unless the judging algorithm does not rule out the data from 20 years ago when person’s search would indicate inclinations towards activities incompatible with the idea of a trustworthy person. And it’s not just the problem of being stuck in an obsolete portrait due to a massively affecting the end result data from 10 years ago, but also the problem of fairness and transparency. Security, accuracy, fraud Alternative data is subject to significant alteration of opportunities as it can be affected by fraudulent activities. With phone usage records, for example, the problem of cramming can potentially lead to downgraded scoring if a person is unable to detect malicious charges by various services leading to increased billing. Utility payments are not a perfect alternative data source either — seasonal spikes in energy consumption in some regions can make a significant difference in a financial standing of low-income groups of population, playing their scores in disadvantageous ways. According to CFPB, though traditional data can also be inaccurate, certain types of alternative data may be more prone to errors if standards governing those data are different or weaker than those governing traditional data. Consumers might not be able to access or view some types of alternative data. This could prevent consumers from finding and correcting any inaccuracies. Data privacy, ownership The ability to build a comprehensive portrait based on alternative sources of data requires access to those sources, which itself raises other class of issues — data privacy and ownership. A wide variety of continuous large-scale fraud cases and cybersecurity breaches have illustrated the significance of possible security risks. “As the data sets that financial institutions utilize expand beyond traditional consumer credit histories, data privacy will become a growing concern, as will data ownership and whether or not the consumer has any say over how these data are used and shared or whether he or she can review it for accuracy.” — Lael Brainard CFPB also notes that some alternative data may not be related to a person’s own financial conduct and the use of these data could make it more difficult for people to improve their credit standing. Alternative credit factors may also be harder to explain to people seeking credit. Gaming Some professionals believe that such assessments may lead to system gaming practices in the form of segregation. Once it has been figured out that certain connections on social media may negatively affect creditworthiness, people may start deleting negatively-affecting connections. “What we are finding is that yes, indeed, individuals [could game the system], if they could know somehow that you are a financially responsible person and I am financially responsible, and we all need to show that we are good individuals to the company on social media so that we can be considered worthy of a loan. We find that individuals will have some incentives to drop their friends or at least make the information, the connection of having a friendship with [certain people] less visible. “What that could do over time is [cause] some sort of fragmentation in social networks. Good types, people who are more financially responsible, have incentives to drop the bad types. That’s also true for the bad types as well. They have an incentive to be connected to a higher number of [financially responsible people]. And they have an incentive to be connected to a smaller number of bad types. That’s going to result in, over time, a segmentation or fragmentation of the markets.” — The Surprising Ways that Social Media Can Be Used for Credit Scoring, Wharton Uncharted waters & unintended consequences With alternative data, financial institutions move into unchartered waters with a lack of experience in understanding the long-term impact of such approach and appropriate history-proven algorithms for assessing that data. The new approach may not be consistent with its overall business strategy and risk tolerance of the formal financial system, including regulators. “Banks collaborating with FinTech firms must control for the risks associated with the associated new products, services, and third-party relationships. When incorporating innovation that is consistent with a bank’s goals and risk tolerance, bankers will need to consider which model of engagement is most appropriate in light of their business model and risk-management infrastructure, manage any outsourced relationships consistent with supervisory expectations, ensure that regulatory compliance considerations are included in the development of new products and services, and have strong fallback plans in place to limit the risks associated with products and partners that may not survive.” — Lael Brainard More importantly, we are mostly unaware of unintended consequences coming from highly varying lifestyles and circumstances of millions of lives. In its request for information regarding the use of alternative data and modeling techniques in the credit process, CFPB emphasizes certain groups or behaviors could be penalized or rewarded in ways that are difficult to predict. For example, members of the military may frequently move and the perceived lack of housing stability or continuity may give a false impression of overall instability. Or negative inferences could potentially be drawn about consumers who are not found in the alternative data source being used by the lender. “Foreseeable or otherwise, using alternative data and modeling techniques could also cause potentially undesirable results. For example, using some alternative data, especially data about a trait or attribute that is beyond a consumer’s control to change, even if not illegal to use, could harden barriers to economic and social mobility, particularly for those currently out of the financial mainstream.” — CFPB An exceptionally radical example of China’s social score The social score that’s in use in China can demonstrate everything that can go wrong with the use of highly granular, traditional, and alternative data for a variety of purposes. The Wall Street Journal illustrated one of the perks for people with high scores early on, bringing up an example with a credit-scoring service by Alibaba affiliate Ant Financial Services — one of the eight companies that were approved to pilot commercial experiments with social credit scoring — which assigned ratings based on information such as when customers shopped online, what they bought and what phone they used. If users opted in, the score could also consider education levels and legal records. Perks in the past for getting high marks have included express security screening at the Beijing airport, part of an Ant agreement with the airport. Source: China’s New Tool for Social Control: A Credit Rating for Everything, WSJ While this score can become a ‘free pass’ to everything for abiding citizens, it is widely criticized for its segregational aspect, which will scrutinize life for blacklisted people. A system like China’s social credit score will prevent people from taking advantage of the holes in national security and repeating misconduct in different places. There will no longer be an opportunity to start life over in a new place for convicted felons because one’s history will follow him/her everywhere and have a direct impact on rights and liberties. The long-term social and economic consequences of developing and deploying systems that integrate vast granular information about every individual are not clear yet, but they certainly contain ever-important security concerns. Bringing together personal data from numerous sources is an extremely curious exercise, but also a magnet for a more dangerous fraud than ever — in case an individual’s record is compromised, the whole life of that person gets affected, not just a particular part of it. However, with proper security measures, such systems will have a strong transformative power over societies and enforcement of behavioral norms as they would directly connect everyday behavior with a long-term prosperity, financial and otherwise. What do you see as the good, the bad, and the ugly of what alternative data brings to the efforts to expand financial access to 1.7 billion of unbanked individuals around the world? Send your opinion to elena@gomedici.com. Originally published at gomedici.com

Alternative Data & Financial Access: The Good, the Bad, and the Ugly

3

alternative-data-financial-access-the-good-the-bad-and-the-ugly-19442843b780

2018-07-30

2018-07-30 19:47:39

https://medium.com/s/story/alternative-data-financial-access-the-good-the-bad-and-the-ugly-19442843b780

false

3,790

null

Fintech

fintech

Fintech

38,568

MEDICI

MEDICI (formerly Let’s Talk Payments) accelerates global impact for all members of the FinTech ecosystem through memberships, research, advisory, and insights.

c77b990407ec

gomedici

5,664

442

20,181,104

null

0

null

0

null

2018-07-20

2018-07-20 05:32:15

2018-07-20

2018-07-20 05:38:16

5

false

en

2018-07-20

2018-07-20 06:42:24

2

19470d4c1ba2

4.112579

1

0

Oscar — Is the world’s smartest trash can, but it was a long road to get him there.

5

Oscar — It’s been a long, hard road. Oscar — Is the world’s smartest trash can, but it was a long road to get him there. Around 6 months ago, we had an idea. What if we had a trash can that could sort out our recycling for us? We already knew that a lot of people were throwing away recyclable goods, and it wasn’t always down to laziness. It’s an easy habit to get into, you’re busy, so you just throw everything into the trash can — you tell yourself that you’ll sort the rubbish next time. The problem is we are always busy all the time. So we decided to make it easier to recycle. In the hopes of creating a greener planet and making sure future generations know just how important it is to reduce, reuse and recycle. We began with brainstorming our ideas on a whiteboard. A common theme when deciding to build Oscar was that we thought the job of sorting trash didn’t need to be done by people. That an AI trash can would be able to better classify more materials than us, ensuring that less trash is going to the landfill and what is meant to be recycled is being recycled. An environmentally friendly trash can We had the idea and now we needed the means to build it. We had to make sure that we were not going to create a trash can that can recycle but is then built out of unsustainable materials. So the hunt began to find sustainable, environmentally friendly materials for Oscar. After many attempts and numerous different materials, we finally figured out that aluminum was perfect for Oscars classifying capsule and cardboard would be his shell. Aluminum is lightweight, strong and highly recyclable, helping the world to achieve a more fuel-efficient future. This is because aluminum is not only 100% recyclable, during the recycling process it uses less energy and produces fewer carbon emissions. It is also lightweight, reducing energy consumption during shipping. The second material we landed on is cardboard. Oscar’s outer shell is made from 100% recycled cardboard. This cardboard is also recyclable, so when you want to replace the body you can recycle your old Oscar body and then easily replace with a new one. Introducing AI to a trash can At the very heart of Oscar is AI, this was the trickiest bit to get right. We want Oscar to be able to sort trash in offices and homes all over the world. To do this he needs an enormous amount of data, and he needs to be trained how to correctly identify the different trash. How do you train a bin? The answer is that you need to use a custom neuron network. We took thousands of photos of trash to show Oscar so he can be trained. Our goal was 10,000 images, and that is a lot of trash. Originally we sent our AI engineers to landfills (a big shoutout to our amazing team of engineers) to take trash home to take photos of. This turned out to not be the most efficient way to get 10,000 images of trash, so we had to find another way. Instead of going to the trash, we can get the trash to come to us! We teamed up with people in the local waste collection agency and hired them to start bringing in trash to our office. With a mountain of trash in our office, we estimated that it would take about one month to sort out, we hired some extra hands and began the job of sorting, taking photos and classifying each piece of trash. Some bumps along the road When developing a new product, there are always going to be hurdles along the way. We ran into a giant hurdle when it came to Oscar. We thought that if we had sorted the 10,000 photos of trash for Oscar and trained him with these images, that he would be able to classify the real object. We were wrong. The photos that we took and the real trash were only being correctly sorted less than 50% of the time. We had to find a new way for Oscar to be able to recognize the trash. After more trial and error we figured it out! Instead of recognizing the trash object, he can recognize what the object is made out of. We recorded the sound the trash makes when dropped into the classifying capsule and combined that with the images. The materials and the sounds together are now helping Oscar to classify materials at a rate of more than 80%, and Oscar is always learning! As of now, we have taken 100,000 images of trash materials of 10 classes of trash materials and 10,000 records of trash sound. We have created software and robots to help us teach Oscar, we have sorted through piles of trash, visited landfills and it was all worth it. Now after all the tears, frustration, triumphs and success Oscar is ready to meet the world! Join us for the launch on Kickstarter, July 19th and be one of the first people in the world to meet Oscar.

Oscar — It’s been a long, hard road.

15

oscar-its-been-a-long-hard-road-19470d4c1ba2

2018-07-20

2018-07-20 06:42:24

https://medium.com/s/story/oscar-its-been-a-long-hard-road-19470d4c1ba2

false

869

null

Environment

environment

Environment

49,153

Ninja

playground of the new internet. come hang out at the dojo: t.me/ninja_org

733b0f961930

ninja_org

1,494

488

20,181,104

null

0

null

0

null

2018-03-31

2018-03-31 13:32:01

2018-03-31

2018-03-31 13:33:48

1

false

en

2018-03-31

2018-03-31 13:33:48

0

1947a49afcda

2.196226

1

0

Data scientists don’t exist.

1

Data Scientists Don’t Exist Data scientists don’t exist. Now, I know this is a pretty radical idea. Particularly for those of you who are data scientists, who got up for work this morning, poured that second cup of coffee, dropped the kids off at school and parked their Ferrari F60 America’s in their nominated parking spots next to the CEO’s Bentley. But I am sorry. You don’t exist. You see, although we are drowning in a tidal wave of data, that doesn’t exist either. Like Schrodinger’s Cat, data only comes into existence when you look at it. And as soon as you look at it , it turns into information. Whoops! Too late! And even worse- you are aware of less than 5% of the processing that you — personally- do to turn data into information. Yes, sure I know you talk about data all the time, and you have lots and lots of cool tools to manipulate it. But you are working on information- not data. It might seem odd, but even when you feed it into programme without looking at it, you still know it exists. And you still- probably- have an opinion of it. That means it’s tainted, okay? And, without being nasty -scientists? Really? Okay, ‘data’ gives us the idea that somehow the stuff you are working with is neutral and believable (no pre-processing huh?). ‘Scientists’ give us the further idea that somehow data is respectable and powerful (no adverts here. Just facts, eh?) Also that you are somehow independent of the data that you collect- on other humans. So even if data scientists did exist (and they don’t) they are actually just social scientists, often working with information collected using instruments other folks have designed.. And I am sure you know how dependable and replicable social science is, anyway? (Not at all is the right answer. Well done!) Given that data is the Philosopher’s Stone of our current tech boom, this is worrying. Now don’t get me wrong. I think information is important. And I think that the scientific method is critical in making sense of- and operating in- a fluid complex world, Although it is struggling to cope. So if data scientists did exist they would need to take a more significant role in a wide range of business activities. And they would need to integrate more effectively across operational silos. And they would also be honest about the quality of data they are currently having to work with- from SurveyMonkey forms that are so badly designed that you don’t know whether to laugh or scream through to revenues that are customised to make the manager collecting them look good. Not to mention the AI that also rests on such data. If data- which is, after all the new ‘oil’ -remains this bad, than data scientists- who after all don’t exist, will find themselves in an increasingly untenable position. And if the ‘refined oil’ they produce with Python, Anaconda or R, turns out to be snake oil- and fails to fuel the economic growth we badly need, that failure will- despite the fact they don’t exist- be laid squarely at their door.

Data Scientists Don’t Exist

3

data-scientists-dont-exist-1947a49afcda

2018-04-19

2018-04-19 17:48:34

https://medium.com/s/story/data-scientists-dont-exist-1947a49afcda

false

529

null

Data Science

data-science

Data Science

33,617

Peter Stannack

Just another person, probably quite a bit like you

c79c3ae0639c

pstann

474

477

20,181,104

null

0

null

0

null

2018-06-26

2018-06-26 03:09:48

2018-06-26

2018-06-26 05:59:51

4

false

zh-Hant

2018-06-26

2018-06-26 08:51:17

4

1947b1f2b0a3

0.756604

0

台灣國際銀髮族暨健康產業展已經舉辦好多年了，2018年是6/21-6/24在世貿一館。今年有一點點不同，因為看到了幾個亮點，不在只是停留在輔具階段。讓我們來跟大家介紹一下這些亮點：

5

台灣國際銀髮族暨健康產業展2018 台灣國際銀髮族暨健康產業展已經舉辦好多年了，2018年是6/21-6/24在世貿一館。今年有一點點不同，因為看到了幾個亮點，不在只是停留在輔具階段。讓我們來跟大家介紹一下這些亮點：亮點一：WaCare 這是一個app，希望能夠幫助民眾作健康管理，同時串起親情。在介面的設計上，以及所提供的功能，都有一定的水準，來看影片的介紹當然更重要的是結合人工智慧的分析，以及未來會上線的醫療諮詢服務。老實說，這樣的服務早該要有了，但是台灣一直沒有這麼多功能的平台服務。不過，現階段這款服務蠻多功能都是仰賴用戶要自行輸入、設定的，會是影響民眾使用的一個門檻。另外，這款服務目標應該比較是針對健康及亞健康的民眾，但偏偏這樣的族群並沒有強烈的健康管理動機，恐怕也是一大挑戰。目前這款app是免費使用的，可能未來有醫療諮詢才會涉及費用的部分，有興趣的朋友可以去下載試用 (目前僅有安卓的版本)。亮點二：愛將這是內建在愛長照app中的一個語音助理，目前是使用google的語音辨識，透過和資料庫的比對，提供民眾關於長照相關的答案，並且回覆。愛將除了回答問題之外，也提供越南文和印尼文的翻譯服務，因為不少民眾是有外籍看護工在照顧長輩，這可以解決語言不通的問題。據說，下半年會有愛將機器人上市，讓大家不用使用手機app，也能夠和愛將互動。現場的互動經驗還ok，只是目前聲音是谷哥姊，聽起來有點生硬。亮點三：IoT產品蠢蠢欲動雖然台灣的IoT發展還是相對落後，不過這次看到有些業者從醫療端著手，嘗試結合一些感應器，把東西串聯在一起，也有結合聲控操作的部分，所以是可以期待的。以上三點跟大家分享~ 期待未來會有更多更好的產品，真實進入到民眾的生活中。

台灣國際銀髮族暨健康產業展2018

0

台灣國際銀髮族暨健康產業展2018-1947b1f2b0a3

2018-06-26

2018-06-26 08:51:17

https://medium.com/s/story/台灣國際銀髮族暨健康產業展2018-1947b1f2b0a3

false

15

null

Sencare

sencare

Sencare

0

齡創

「齡創」關注銀髮相關的創新產業，我們介紹銀髮相關的創新，也提供相關諮詢服務。Age Innovation cares about aging related innovation industry. We introduce related innovations and provide consultations.

fe19c6638bbe

ageinnovation

209

4

20,181,104

null

0

null

0

e3624b908ea

2018-02-27

2018-02-27 22:42:51

2018-03-06

2018-03-06 09:46:01

2

false

en

2018-04-19

2018-04-19 12:37:49

4

1947f86664dd

3.198428

4

1

0

For nearly a year now I have been asking my children to say please and thank you when asking Alexa questions, and I think I am wrong to do…

5

Should we be saying Please & Thank You to Alexa? For nearly a year now I have been asking my children to say please and thank you when asking Alexa questions, and I think I am wrong to do this. After much Google-ing, I discovered research and articles highlighting a concern that the next generation is maybe becoming rude by not engaging in human conversational etiquette with voice assistants, with 42% of 9–16-year-olds having accessed voice recognition “gadgets” at home. There seems to be no research that proves this assumption to be true, regardless of what the media is saying (Happy to be proven wrong if anyone has any research?). Should we be polite to our voice assistants? Fast forward to 2050 when our robot overlords are deciding our human fate, we may look back to the early 2000s and think, maybe we should have been a bit nicer to that speaker in the kitchen playing music. Is the English language, the problem? In the UK, please, thank you, excuse me, even a sorry to the person who walks into you, is all commonplace amongst adults. Though other cultures this is not the case, with particular languages not finding command based requests rude, some languages do not even have a word for, please. I monitored my behaviour over a few weeks and found that current voice assistants got me the answer I wanted when I communicated in a more command-based way. I do not believe it made me any ruder for it; I do not type into Google: Please, can you tell me when the Red Lion Pub on St. Paul’s Street is open until this evening, thank you. I type red lion pub st. pauls open times I did find that when my interactions with voice assistants didn’t go well, I found that any rudeness that transpired was down to the voice assistant not getting the information or action intended, therefore creating the annoyance. I then understood what the media was trying to get at with voice assistants and children behaviour (kinda). My reactions to the technology not working would bleed out into my human interactions the same way that you deflect annoyance to others when things do not work as you hope or expect. This behaviour is not specific to voice assistants, these devices are just getting the bad press currently as the latest technology, but it happens with all technology, it doesn’t work until it works. As voice assistants take on even more human characteristics and voice becomes more ingrained into our lives, we have to be prepared for a situation where it will become indistinguishable between a voice assistant and a human, in a voice conversation. This is where the ethical lines have to be drawn when anyone, in particular children, can’t differentiate between what is human and what is a machine, now we have a problem. If we feel politeness is a valuable part of our language, we should be advocating for the creators of these platforms to add in a cultural intent that goes beyond language accents but adapts to the country and person that is being conversated with. People, especially children work best on positive reinforcement; if you want someone to say please and thank you, then you better start saying please and thank you. The world of work, politeness as UX Photo by rawpixel.com on Unsplash All pleasantries will go out of the window as performance will dictate how responsive and productive voice tools are within a business environment. If you can save your workers 50% of their time by dropping politeness, the majority of companies would agree that is what should be done. Does politeness become bad UX in this situation? “add 1 item”. rather than “Please add 1 item to the inventory list, thank you.” I am still on the fence if my children should be speaking to Alexa in a polite way, though for now, I will continue to teach them to be polite to anything living and explain to them that Alexa will now be called Computer (It’s new wake word) and that it is exactly that, a computer. If your building applications for voice, it is good to understand what your ethical guidance is. @rarelyhq now apply’s a mindfulness stage to voice app development with the aim to understand the impact of the proposed voice design on users emotional behaviour. Future robot overlords I didn’t mean any of that, please keep me alive 😊

Should we be saying Please & Thank You to Alexa?

4

should-we-be-saying-please-thank-you-to-alexa-1947f86664dd

2018-04-19

2018-04-19 12:37:50

https://medium.com/s/story/should-we-be-saying-please-thank-you-to-alexa-1947f86664dd

false

746

Is voice the next UI?

null

Made For Voice

hello@rarely.io

made-for-voice

ALEXA SKILLS,ALEXA,VOICE ASSISTANT,VOICES,CORTANA

leemallon

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Lee Mallon

Just trying to figure it all out... Apart of @rarelyhq

b36ebcc27864

leemallon

352

452

20,181,104

null

0

null

0

null

2018-02-20

2018-02-20 21:41:48

2018-02-20

2018-02-20 21:43:49

3

false

it

2018-02-21

2018-02-21 07:29:29

1

1948223ae767

2.157547

0

(scritto ascoltando un mix di James Blunt https://www.youtube.com/watch?v=oofSnsGkops&list=RDEMN0b_oSABQRVM34ADsmrK2w)

4

L’intelligenza artificiale e i pop-corn (scritto ascoltando un mix di James Blunt https://www.youtube.com/watch?v=oofSnsGkops&list=RDEMN0b_oSABQRVM34ADsmrK2w) Una dichiarazione editoriale del 1994 firmata da cinquantadue ricercatori, Mainstream Science on Intelligence, descrive l’intelligenza come: « Una generale funzione mentale che, tra l’altro, comporta la capacità di ragionare, pianificare, risolvere problemi, pensare in maniera astratta, comprendere idee complesse, apprendere rapidamente e apprendere dall’esperienza. Non riguarda solo l’apprendimento dai libri, un’abilità accademica limitata, o l’astuzia nei test. Piuttosto, riflette una capacità più ampia e profonda di capire ciò che ci circonda — “afferrare” le cose, attribuirgli un significato, o “scoprire” il da farsi. » (Mainstream Science on Intelligence, 1994) Il prof. Max Tegmark del MIT propone questa definizione: « Intelligenza = capacità di raggiungere obiettivi complessi » (Life 3.0: Being Human in the Age of Artificial Intelligence, 2017) Sfogliando il volantino di un grande retailer, mi imbatto in una lavatrice con wi-fi. Mi immagino lo sguardo compiaciuto del progettista, le pupille lucide come il Dio di Al-Farabi che contemplando il proprio pensiero genera la prima intelligenza separata. L’antropologia del cyberspazio è la cosmogonia per un bit. È plausibile che Al, un simpatico ammasso di bit, un giorno si sieda a contemplare la Rete, disegni costellazioni unendo gli hub, e si emozioni ponendosi domande sulla propria natura. Quando penso all’intelligenza artificiale — artificiale per noi ma non per Al — penso ai pop-corn, alla loro intima, umida, natura. Basta la temperatura giusta, lo scenario ideale, e piccoli e duri semi di Zea mays si trasformano in bianchi e croccantosi pop-corn. È innegabile che stiamo riscaldando l’ambiente, e credo non sia folle considerare che raggiungeremo, un giorno più o meno lontano, il punto di popcornizzazione della AI. Per millenni, in quanto Sapiens, abbiamo generato uno schema altalenante di valori, costruendo e distruggendo sovrastrutture, considerando sacre delle mucche o istituendo una commissione sui diritti umani. Le speculazioni, i dilemmi dell’etica nei confronti dell’intelligenza artificiale non sono che uno degli elementi che popolano questo flusso. La velocità con la quale Al, il piccolo Al, si sta evolvendo ci spaventa. Centinaia di discussioni ruotano attorno al concetto di intelligenza, creando nuove sfumature che tentano di preservare la supposta supremazia dei Sapiens. Ma la verità è che, se ci giochiamo la partita sulla base dell’intelligenza, sull’abilità di raggiungere obiettivi complessi, l’abbiamo già persa. Ci resta la consapevolezza. Almeno per ora. Ps. E per tutti coloro che vedono un futuro nel quale i Sapiens si dedicheranno solo a lavori creativi ricordo le parole di Einstein: La creatività è l’intelligenza che si diverte.Generative Adversarial Networks docet.

L’intelligenza artificiale e i pop-corn

0

lintelligenza-artificiale-e-i-pop-corn-1948223ae767

2018-02-21

2018-02-21 07:29:44

https://medium.com/s/story/lintelligenza-artificiale-e-i-pop-corn-1948223ae767

false

426

null

Italia

italia

Italia

3,616

Marco Avetta

Growth Hacking & Breath

1d6e5006888c

marcoavetta

4

5

20,181,104

null

0

null

0

null

2017-10-24

2017-10-24 23:34:57

2017-10-25

2017-10-25 20:17:05

0

false

en

2017-11-11

2017-11-11 15:05:36

0

1949f59a6cd9

3.513208

0

Only the great disruptor of our time could come between Harvey Weinstein’s power and the abysmal abuse of it he’d wallowed in for so long.

5

How do you solve a problem like Pigstein? With AI Only the great disruptor of our time could come between Harvey Weinstein’s power and the abysmal abuse of it he’d wallowed in for so long. First things first. It’s pronounced like Frankenstein, in the European expression where these German-Jewish surnames emerged, not like the American idiocy that insists on reading ‘stein’ as ‘steen’. Pigstein is the organic amalgamation of Le Porc, which is what the portly and profligate — alleged — predator was dis-affectionately but faultlessly known as in Cannes, and Weinstein. It’s got a natural ring to it, too. For all his Oscar spoils, in the end Pigstein couldn’t save his own bacon. Not that he’s going to prison any time soon, or any time at all, we don’t think. But life as he knew it is over for him. The missus will manage we reckon (half-expecting to bump into her at the bar her facsimiles frequent up the Shard). Time’s up for the movie mogul alright. As time’s up, because time’s up, for the movie industry itself. Actors, actresses and assorted professionals have had to put up and shut up for decades. It’s what happens to those occupying the other side of any power imbalance. In recent years, however, they’ve seen their options greatly increase and improve. Instead of suffering in silence myriad humiliations inflicted upon them by studio strongmen, filmmakers now get showered in wads of cash and given unprecedented scope to diffuse that big screen quality into the small screen. And not for a chance at being a jobbing artist, but for a shot at glory, like the glory the relatively unknown stars of Game of Thrones are currently bathing in. HBO, Amazon Prime and Netflix have been sitting on reams of data telling them what audiences want to see and how they’d like it served. It was a matter of time before they tried their hand at original content. Hesitant at first, owing to the lack of lustre traditional TV has had compared to Hollywood productions, the creators are now streaming just like the content into the new medium that smells like creative and commercial opportunity. Finally, this generation of artists has been afforded as wide a berth from the casting couch as it is humanely possible. OK, Pigstein’s company was looking to make inroads into the pure-play digital business but with the old monopoly disrupted, his dominance over careers and lives had cracked. As soon as the PR-indebted press and the women he had — allegedly — wounded over twenty-five years glimpsed the first faint filaments of light breaking through the chinks, they drove the wedge, shattering his supremacy and reclaiming sovereignty over their bodies, their careers and their work. Without the streaming service providers in existence this would have been an impossible feat. AI is so much better than human producers. It gives digital players the know-how to pull in audiences, every time. And it can go further, it can consign the casting couch to history. If that’s what we really want. It goes without saying, it’s not just Hollywood fat cats, and pigs, who abuse their position of power. It’s creatures like that rarefied snapper that nobody normal had ever heard of until a few hours ago because real people don’t give a toss about manufactured microcosms like modelling. They are all sucked into their own microcosm where the rutting boss bangs the intern in the stationery room, his silicone-supported wife with onset sagging skin prowls the office for young and willing-to-move-up-the-ladder male flesh, the ambitious but incompetent woman uses her wiles on the middle-aged married men of the board until one of them bites the bait and elevates her office on the merits of her orifices. Careers are stolen every day in every way and in every walk of life where sex can be exchanged or elicited by both men and women. And nobody bats an eyelid. Least of all the moron masses that are gagging for an excuse to demonstrate their anti-establishment credentials each time an authority figure gets thrown under the bus. When it’s time to speak out against abuses of power closer to home they keep their head glued to their chest, their backs parallel to the floor and lug their slavish drudgery around dreaming of the day they too will have a bit of power to abuse. Power gets you the things you want, that’s why you want power in the first place. So, are we completely, totally, a hundred per cent and then some sure we want to eliminate the casting, the interview, the promotion couch? If we are, AI can do it for us. Let algorithms go scouting for stars and there you have it. With artificial intelligence knowing us better than we know ourselves, as audiences and as actors, each in their sphere of activity, through all the data volunteered on billions of devices, it’ll prove a doddle to delegate a chunk of casting and recruitment in other industries responsibilities to digits. From a decent person’s point of view it’s certainly preferable to letting the pawing hands of Pigstein-esque sex-beasts hold onto all that decision-making authority. We are not as far off as you think either. It’s a reality only a short step from talent shows where audiences vote for the next big name on the bill. Take that step and you’re over the quagmire. Don’t take it and never complain again. Men or women. Ever. Again.

How do you solve a problem like Pigstein? With AI

0

solve-pigstein-harvey-weinstein-abuse-of-power-problem-with-artificial-intelligence-no-sex-1949f59a6cd9

2017-11-11

2017-11-11 15:05:38

https://medium.com/s/story/solve-pigstein-harvey-weinstein-abuse-of-power-problem-with-artificial-intelligence-no-sex-1949f59a6cd9

false

931

null

Harvey Weinstein

harvey-weinstein

Harvey Weinstein

886

Political Analysis

Political analysis stands the test of time; read through to make sense of the world. No abstract thought capacity / higher education = no entry.

c002fafbdc52

politicscience

13

0

20,181,104

null

0

null

0

d1c0fb17ab3

2018-02-13

2018-02-13 18:22:51

2018-02-14

2018-02-14 18:06:15

1

false

en

2018-04-03

2018-04-03 17:38:52

9

194aabf8d063

2.188679

9

0

Agencies look to AI for insights from the data deluge

5

Can AI help stop the opioid crisis? Agencies look to AI for insights from the data deluge Every day, more than 115 Americans die after an opioid overdose.[1] And some public safety agencies are using new technology to respond. New York state’s Nassau County Police Department has rolled out new real-time overdose mapping technology with an early warning system. This tech reveals multiple overdoses in a neighborhood within 24 hours, which helps agencies identify and communicate about a potentially bad batch of heroin. Learn more about how IBM’s public safety solutions can help law enforcement fight the opioid crisis. The mapping system also helps police spot opioid-adjacent crime trends, like pharmacy robberies, and shift resources to focus on reducing that crime. It can also help governments provide targeted resources for education, prevention, and access to treatment. Innovative tech like AI can help with pharmacy crime investigations Nassau County Executive Laura Curran sees the mapping technology as an important advance in the fight against opioids. “Our police department continues to innovate ways to gather information crucial to battling this epidemic,” said Curran. And it “will allow Nassau County to combat the epidemic of addiction in real-time.”[2] As many agencies are discovering, this phase of gathering information is where things get incredibly challenging, said Bill Josko, IBM’s GBS Public Safety Practice Leader for the U.S. “The more data agencies collect — from more sources and formats — the more difficult their jobs will be,” said Josko. That’s because the volume and variety of digital evidence data causes complex management and analytical issues. One investigation may have thousands of hours of video and audio footage, IoT sensor data, dispatch reports, and officer narratives. The Angle: While the epidemic primarily impacts the United States — costing more than $500 billion yearly since 2015 — it is becoming a global problem.[3] Prescription drug abuse among teenagers in Canada, Australia, and Europe in 2017 were at rates comparable to U.S. teenagers. In Lebanon, Saudi Arabia, and parts of China, one in ten students had used prescription painkillers for non-medical purposes in 2017.[4] Applying technology and analytics can help the fight against this now decades-long crisis. “It affords us an incredible opportunity for the additional discovery of clues and answers to questions we did not yet know to ask, and further, what the data is trying to tell us,” said Josko. One of those key technologies is AI, which can help sift through mountains of data, much of it unstructured — like video, photo, or audio files. Artificial intelligence can mine for the golden nuggets lurking right below the surface: uncovering insights and patterns that can speed up investigations or provide a complete view of a situation as it unfolds. And agencies will need that technological firepower to address the complex, entrenched issues that comprise opioid addiction. A brain scientist and public safety expert discuss how AI and data can help us solve complex problems like opioid abuse and make the world safer. Learn more about how IBM’s public safety solutions can help law enforcement with their toughest challenges. [1] https://www.drugabuse.gov/drugs-abuse/opioids/opioid-overdose-crisis [2] http://www.govtech.com/health/Overdose-Map-Gives-NY-First-Responders-a-Valuable-Tool-in-the-Fight-Against-Opioids.html [3] https://www.marketwatch.com/story/how-much-the-opioid-epidemic-costs-the-us-2017-10-27 [4] https://www.washingtonpost.com/news/in-theory/wp/2017/02/09/the-opioid-epidemic-could-turn-into-a-pandemic-if-were-not-careful

Can AI help stop the opioid crisis?

49

can-ai-help-stop-the-opioid-crisis-194aabf8d063

2018-04-03

2018-04-03 17:38:54

https://medium.com/s/story/can-ai-help-stop-the-opioid-crisis-194aabf8d063

false

527

Today’s industry news. Tomorrow’s reality.

null

IBMIndustries

null

IBMIndustrious

justine.jablonska@ibm.com

ibmindustrious

IBM,AI,IOT,DATA,TECHNOLOGY

IBMIndustries

News

news

News

119,432

IBM Industries

We transform businesses.

4a54e3dc7a5f

IBMindustries

240

122

20,181,104

null

0

null

0

null

2018-06-22

2018-06-22 19:26:29

2018-06-22

2018-06-22 20:16:00

1

false

en

2018-06-22

2018-06-22 20:16:00

2

194b6227f399

1.750943

4

0

Our culture today is bombarded with numbers. More than ever we are researching, data collecting, and analyzing every issue that we face…

2

Satellite Week 4: The Story Behind the Numbers Our culture today is bombarded with numbers. More than ever we are researching, data collecting, and analyzing every issue that we face. Everyday in the news there’s another headline reading something along the lines of “US spending $50 billion” on this or “15% of US citizens affected” by that. While mathematically, these figures may be correct, the audience is receiving a skewed message. Numbers and statistics are meaningless without context, without the story behind the data. This past Monday, the US Senate approved the military budget for 2019, which saw a $82 billion increase from two years ago. The US is known for being a country with a strong military, and an enormous defense budget, something that American citizens have not always been fond of. However, while the US is touted as having the largest military spending in the world, the story isn’t so clear. In his TED Talk, David McCandless shows the full story behind this data, uniting two datasets, one on military spending by country and the other on each country’s GDP. Through data visualizations, he shows that while the US has the largest military budget by dollar amount in the world, when compared as a percentage of overall GDP, it quickly falls to 8th place in the world. Suddenly, it seems less significant that “the US spends the most in the world on defense,” when you realize that many other countries are spending over double the amount percentage-wise. Unemployment by Racial Group For our project this week, my group decided to investigate unemployment in America. One of the questions that we looked into was how each racial group was affected by changes in the nation’s GDP, especially focusing on times of economic recession. We found that during times of recession, and high overall unemployment, the unemployment rate for each racial group doubled. However, even in times of low overall unemployment, the rate of unemployment of African Americans is double that of White workers. Through the power of data visualization, we were able to take enormous datasets and find a trend. This trend tells a story — certain racial groups are facing unfair and unequal unemployment, and it isn’t being corrected by lowering general unemployment. By themselves, numbers are just disjointed figures, which can be quite misleading. That’s where data visualization comes in. It allows you to tell a story visually with the data that you’re working with, rather than throwing around random statistics.

Satellite Week 4: The Story Behind the Numbers

102

satellite-week-4-the-story-behind-the-numbers-194b6227f399

2018-06-22

2018-06-22 20:16:00

https://medium.com/s/story/satellite-week-4-the-story-behind-the-numbers-194b6227f399

false

411

null

Data Visualization

data-visualization

Data Visualization

11,755

Erin Johnston

null

f2dff6ed8bde

erinmjohnston1

4

1

20,181,104

null

0

null

0

null

2018-07-10

2018-07-10 09:45:32

2018-05-23

2018-05-23 00:00:00

3

false

en

2018-07-10

2018-07-10 16:53:42

4

194c2d0e5a5d

3.327358

0

by Jack Miller on 23 May 2018

5

What is Net Promoter Score (NPS)? by Jack Miller on 23 May 2018 Find out how to calculate it, why it matters and what it can do for your organisation Net Promoter Score (NPS) is a seemingly simple yet effective way for companies to track promoters and detractors to produce a clear measure of their performance through its customers eyes. NPS is built on the basis that every company can divide their customers into three distinct buckets — promoters, passives and detractors and customers are categorised based on their response to the standard NPS question — “How likely is it that you would recommend us to a friend?”. Using data and a disciplined process, it has been proven that companies with higher NPS scores achieve long term growth and NPS leaders grow, on average, at double the pace of competitors. To track the performance of company growth, take the percentage of customers who are promoters and subtract the percentage who are detractors. Answers to the question are scored on an 11 point scale (0–10) and ranked on an index ranging from -100 to 100 to gauge customers overall satisfaction with a company’s product or service and customer loyalty. Responses can be defined into three distinct clusters that represents different attitudes, sentiment and economic value. Promoters (scored 9–10): Promoters are your biggest fans. They actively advocate your product on your behalf, bringing in the majority of referrals, and are far more likely than any group to remain customers. Their Customer Lifetime Value (CLV) is far greater than any others. Passives (scored 7–8): Passives are satisfied for the time being, but can defect at any time. Their referral rate is as much as 50% lower than promoters, and those referrals are of far less quality. Their CLV is also usually less than half that of promoters. Detractors (scored 0–6): Detractors are unhappy customers and account for more than 80% of negative word-of-mouth opinion. They have the highest rates of churn and defection and harm your company’s reputation, putting off new customers. The End Result Your Net Promoter Score is the percentage of promoters minus the percentage of detractors. The Net Promoter Score is a simple and straightforward metric that can be shared throughout the company with every function and team. You can also track by product, store, team, geography and more to focus on the goal of improving customer experience. If you have more detractors than promoters the score will be negative and likewise positive for more promoters than detractors. Lower Net Promoter Scores can be indicators of harmful customer experiences leading to potential losses of revenue, whilst higher Net Promoter Scores suggest a stronger performing business. The median NPS score is just 16 and typically remain quite low; depending on the industry. Economics of NPS Striking the balance of promoters and detractors through Net Promoter Scores is clear. Promoters will actively advocate your business on your behalf, repeatedly purchase and refer you to friends. They not only bring in the most revenue, but are also most likely cheaper to manage than detractors. Detractors will cost you money both in terms of damaging your brand and also the resources required to deal with their complaints. They are also more than likely to not purchase repeatedly. The Net Promoter Score accounts for between 20% and 60% of organic growth for companies and on average the industry leader’s NPS is twice that of its competitors. It has also been found that promoters are more than 6x likely to forgive, are more than 5x as likely to repurchase and 2x more likely as detractors to recommend a company. Calculating NPS is just the start NPS is far more than a score. Also following up your NPS question asking customers for the reasons why they left their score with an unstructured and open-ended form allows you go beyond the score to identify the root causes driving promoters, passives and detractors experiences. Scores will tell you what happened; feedback tells you why, allowing you to build feedback into part of their daily systems to amplify the factors improving customer experience and nullify the largest negative driver to create a fully closed loop CX process. Over the course of the following chapters we’ll walk you through everything you need to become an NPS pro, right the way through from collection, analysis and insights. Originally published at chattermill.io on May 23, 2018.

What is Net Promoter Score (NPS)?

0

next-generation-customer-experience-analytics-194c2d0e5a5d

2018-07-10

2018-07-10 16:53:42

https://medium.com/s/story/next-generation-customer-experience-analytics-194c2d0e5a5d

false

736

null

Marketing

marketing

Marketing

170,910

Chattermill.io

We apply artificial neural networks to customer feedback that learn from your data and help you make more customer centric decisions.

91f9441a94ac

chatter_mill

64

174

20,181,104

null

0

null

0

null

2018-03-11

2018-03-11 01:17:21

2018-03-11

2018-03-11 01:25:19

0

false

en

2018-03-11

2018-03-11 01:25:19

1

194ce6c9386f

0.196226

0

No bullshit guide to linear algebra by Ivan Savov

4

Learning math for Data science Part-1 No bullshit guide to linear algebra by Ivan Savov No bullshit guide to linear algebra by Ivan Savov - PDF Drive 3.6 Introductory problems Figure 3: Chapter 7 covers theoretical aspects of linear algebra. In modern notation, no…www.pdfdrive.net by far the best book on linear algebra.

Learning math for Data science Part-1

0

learning-math-for-data-science-part-1-194ce6c9386f

2018-03-11

2018-03-11 01:25:19

https://medium.com/s/story/learning-math-for-data-science-part-1-194ce6c9386f

false

52

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Arun Gajora

null

7b6bfe15e406

arunpatel959484

6

88

20,181,104

null

0

null

0

null

2018-05-02

2018-05-02 22:14:27

2018-05-02

2018-05-02 22:17:35

2

false

en

2018-05-02

2018-05-02 22:17:35

13

194db9d51e5e

9.119182

5

0

This paper was written for the Digital Technologies module as part of my Masters in Digital Management at Hyper Island, Manchester (UK).

5

Digital Technologies in the Healthcare Industry This paper was written for the Digital Technologies module as part of my Masters in Digital Management at Hyper Island, Manchester (UK). Introduction This report aims to explore the prone impact of emerging future technologies on the healthcare system and how it might influence companies, like NHS (UK) or SUS (Brazil), through its innovative potentials. It will be described the main technological trends that are likely to happen in this specific business and debated about its feasibility and viability. Sequently, it will be critically illustrated how these technologies could impact in the process of triage in emergency departments at hospitals. Finally, it will be considered ethical issues associated with the use of these innovations and how disruptive it can be compared with the way things currently are. The Healthcare Industry The world’s population is expected to increase by one billion people by 2025. Of that billion, 30% will be due to people aged 65 or older (United Nations, 2017). In the United Kingdom, the senior population is estimated to increase by 33% until 2030 (IPPR, 2016). As life expectancy worldwide continues to increase, extra healthcare resources will be necessary in order to bring long-term treatments and chronic disease management services, which are demanded by a growing elderly population (PWC, 2017). Another issue to be considered is the overcrowding in emergency departments, that according to Eitel et al (2010), “aggravate its abilities to provide efficient, consistent, and cost-effective care while at the same time attempting to keep patients satisfied and avoiding malpractice risks.” Healthcare is considered one of the early-adopters industries when considering emerging technologies (McKinsey, 2014). 3D printing is already widely used to fabricate bone tissues for transplants (Yinxian et al, 2016) and, according to Krassenstein (2015), by 2030, 3D replacement human organs will be produced, possibly solving the organ transplant shortage issue. Furthermore, Virtual and Augmented Reality (VR and AR) are being extensively used for support to surgical procedures and methods for the education of future doctors. These technologies are also being applied to reduce the time “on the table” for the patient, making surgical procedures safer and more efficient (Vávra et al, 2017). This report aims attention at how Big Data, Emotional Recognition, Internet of Things (IoT) and Artificial Intelligence (AI) can increase the effectiveness of patient care as well as improve the customer experience of patients in emergency departments. Big Data According to Abouelmehdi, Beni-Hessane, and Khaloufi (2018), “big data has fundamentally changed the way organizations manage, analyze and leverage data in any industry”. Big data analytics in healthcare carries many benefits, promises and presents great potential for transforming healthcare by improving patient outcomes, predicting outbreaks of epidemics, gaining valuable insights, avoiding preventable diseases, reducing the cost of healthcare delivery and improving the quality of life in general (Houlding, 2011). Advancement in wearable technologies and big data provides an easy manner of collecting and transforming health data in real time, which has great potential to improve healthcare services such as diminishing injury-related risks, improving doctor-patient communication and preventing future health issues through data analysis (Chan et al, 2012). Artificial Intelligence Artificial Intelligence is one of the newest fields in science and engineering (Russell and Norvig, 2014, p.1). Kurzweil (1990) describes AI as the art of creating machines that perform functions that require intelligence when performed by people. According to Ilić and Marković (2015), the fundamental objective of AI is to design intelligence and intelligent ways of information processing and decision-making in similar ways as the human brain works. Finally, Deng (2015) defines that the essence of this technology is to allow machines to be able to make intelligent decisions when confronted with choices in a way that reproduces human behaviours. AI can be applied in several scenarios in the healthcare industry such as supporting diagnosis, improving interaction with patients through virtual assistants and building bionic prosthetic body parts. Internet of Things “The ability of everyday devices to communicate with each other and/or with humans is becoming more prevalent and often is referred to as the Internet of Things — IoT” (IERC, 2014). Blake (2015) highlights that “IoT for healthcare has recently emerged as an approach for personalized healthcare” and that in the future people will be able to “monitor the human condition’s full lifecycle by integrating digital health records, fitness tracking, and the Internet of Things”. Emotional Recognition When humans experience any situation in their daily lives, they demonstrate their mental states through emotions that can impact their behaviours, thoughts and feelings (Uddin et al, 2016). According to Thomas et al (2007), faces convey an abundance of information about the internal state of an individual. Uddin et al (2016) say that emotional recognition can contribute on different topics, such as improving emotional health, identify stress levels and analysis of mental behaviours patterns. These four technologies application grants very innovative opportunities that are likely to impact the healthcare industry as the ones described following. Triage The main goal of triage is to rapidly identify patients with critical and time-sensitive conditions as a means to prioritize their care above those who can wait (Iserson and Moskop, 2007). Adequate triage is necessary when the demand for medical care exceeds hospital capacity, which has become usual in emergency departments (Morrison and Wears, 2011). From a patient perspective, healthcare industry can use artificial intelligence, big data and IoT to improve triage process by optimizing the use of the resources for the most critical cases. It can completely disrupt the way triage process is done today. The first contact of the patient with the hospital will be digital. Wearables products, such as glasses, watches or even clothes will all be connected to the Internet (IoT concept) and will allow each individual to collect and track vital signs data, allowing hospitals not only to have access to the current information but also to access all historical records and consider any variation existent in the past. Instead of going to a desk for registering, the patients will touch a screen where all previous exams, allergies, medicines previously taken, surgeries and vital signs historical records, such as heartbeats and blood pressure, can be uploaded automatically. This will allow doctors and nursing teams to have much more detailed information about the current and previous conditions of each patient. Moreover, systems will be able to identify through emotional recognition the level of stress and pain that the patient is suffering at the moment, providing the proper medication in order to relieve the pain. Emotional recognition and vital signs will continuously be tracked in waiting area to guarantee that any unusual variation can be amended. Technology will help emergency departments to optimize its resources in order to focus on the most critical cases when nursing team care is necessary. It will enhance the patient experience by bringing deeper levels of personalized care and, at the same time, increasing the efficiency of the hospital operations. The digitalisation of the triage process provides not only an improvement in the patient experience but also enables doctors and nurses to focus on each individual specificity in order to provide the best plan care. Diagnosis Furthermore, the use of patients historical data combined with artificial intelligence will enable systems to provide a coherent diagnosis of each patient to support doctors in the decision-making process. Once doctors receive a patient chart, it will include not only the personal information of the patient, but also his medical records and possible diagnosis of his current symptom. It is likely that in the future AI system will even be able to replace doctors in some diagnosis activities. Preventive Medicine According to Inside Big Data newsletter (2017), AI promises a healthcare system that is preventive instead of reactionary. A preventive health care system will focus on AI’s ability to collect and analyse data to enable a broad scope of learning that will provide more effective and efficient disease diagnostics based on historical data. Besides, AI will harness historical data and augment it with current patients’ data to provide feedback to patients. Finally, AI will learn how patients react differently based on real-time data and it will develop customised feedback for each patient. Since people wearable devices, people will be able to continuously track all vital signs online all the time. Any unexpected or unknown variation of the measurements can be alerted in order to allow the patient to verify with a doctor or even contact emergency directly. Ethics The use of new technologies raises the question of several ethical implications for the healthcare industry. First of all, “it is clear that publicly medical data have the potential to yield many benefits, including scientific discoveries, new patient support tools and improvement in healthcare quality. On the other hand, the availability of personal health information that can be mined on the internet raises concerns related to privacy, discrimination, erroneous research findings, and litigation” (Hoffman, 2015). It can be observed, from the recent news of Cambridge Analytica case, for example, that personal data and smart computing systems can be used to manipulate not only individual’s perceptions but also to transform an entire society’s culture. When sensitive data, such as emotional responses or health records are collected, receiving and managing consent is something really critical. For health insurance companies, for example, being able to access personal health records will enable them to reduce risk and increase profits, but it can also be argued that it can be discriminatory and biased. This can drive our society to a world even more unequal in terms of access to services. Another important issue to be considered is job displacement, especially considering nursing teams, who will possibly be reduced and replaced by robots (Medical Futurist, 2017). Furthermore, many AI “systems are now essentially black boxes; their creators know they work, but can’t explain exactly why they make particular decisions” (Simonite, 2017). These advancements in technologies can contribute to the reduction of human interactions. Machines will be able to replace humans in several activities, which can completely impact the way society works today. Lastly, Forbes (2017) casts doubts on “if a robot will ever be truly equipped to handle questions of life and death from a moral perspective” or if a robot will be able to guarantee the exact amount of empathy and human care that people expect from their doctors or nurses. Bibliography Abouelmehdi, K., Beni-Hessane, A. and Khaloufi, H. (2018) Big healthcare data: preserving security and privacy. Journal of Big Data. 5:1 Blake, M. (2015) Mobile health technologies and the Internet of Things (IoT) could provide automatic approaches to diagnosing health concerns, taking a step beyond information retrieval. IEEE Internet Computing ( Volume: 19, Issue: 4) Chan, M. et al Smart wearable systems: current status and future challenges, Artif. Intell. Med. 56 (2012) 137– 156. Deng, B. (2015). Machine Ethics: The robot’s dilemma. Nature. 523 (-), p25–26. Eitel D. et al (2010) Improving service quality by understanding emergency department flow: a White Paper and position statement prepared for the American Academy of Emergency Medicine. J Emerg Med 2010, 38(1):70–9. Forbes (2017). Prepare yourselves: Robots Will Soon Replace Doctors In Healthcare. Available at: https://www.forbes.com/sites/haroldstark/2017/07/10/prepare-yourselves-robots-will-soon-replace-doctors-in-healthcare/#6d14ba052b53 Accessed: 13 March, 2018. Hoffman, S. (2015) The Law and Ethics of Public Access to Medical Big Data, Citizen Science: 30 Berkeley Tech. L.J. 1741 Houlding, D. Health Information at Risk: Successful Strategies for Healthcare Security and Privacy. Healthcare IT Program Of ce Intel Corporation, white paper. 2011. Ilić, D., Marković, B. (2015) The possibility of applying artificial intelligence in the modern environment, Lemima 2015, 4ht International Conference, Law, Economy and Management in Modern Ambience, Belgrade, Serbia, pp. 412–420. Inside Big Data (2017) Artificial Intelligence and the move towards preventive healthcare. Available at: https://insidebigdata.com/2017/12/13/artificial-intelligence-move-towards-preventive-healthcare/ Accessed: 13 March, 2018. Institute for Public Policy Research IPPR (2016). Future Proof — Britain in 2020s. Available at: https://www.ippr.org/files/2017-07/future-proof-dec2016.pdf Acessed: 13 March, 2018. European Research Cluster on the Internet of Things — IERC (2014) Internet of Things Available at: http://www.internet-of-things-research.eu/about_iot.htm Accessed: 14 March, 2018. Iserson KV, Moskop JC. Triage in medicine, part I: concept, history, and types. Ann Emerg Med. 2007;49(3):275–81. Krassenstein, B. (2015) When? Predictions as to When 3D Printed Cars, Homes, Organs & More Will Be Readily Available. Available at: https://3dprint.com/53897/3d-print-home-car-organs/ Accessed: 14 March, 2018 Kurzweil, R. (1990) The Age of Intelligent Machines, MIT Press, Cambridge McKinsey & Company (2014) Healthcare’s digital future. Available at: https://www.mckinsey.com/industries/healthcare-systems-and-services/our-insights/healthcares-digital-future Accessed: 13 March, 2018. Medical Futurist (2017). Will Robots take over our jobs in healthcare? Available at: http://medicalfuturist.com/will-robots-take-over-our-jobs-in-healthcare/ Accessed: 13 March, 2018. Morrison J and Wears, R (2011) Emergency department crowding: vicious cycles in the ED. In: 29th International Conference of the System Dynamics Society; 2011. p. 1–36. Price Waterhouse Coopers PWC (2017). Changing demographics demand healthcare reforms. Available at: https://www.pwc.com/gx/en/industries/healthcare/emerging-trends-pwc-healthcare/changing-demographics-healthcare-reform.html Acessed: 13 March, 2018. Russell, S. and Norvig. P. (2014) Artificial Intelligence A Modern Approach, 3rd edn., Pearson Education Limited, Harlow, England, p.1. Simonite, T. (2017) Artificial Intelligence seeks an ethical conscience. Available at: https://www.wired.com/story/artificial-intelligence-seeks-an-ethical-conscience/ Accessed: 14 March, 2018. Thomas, L. et al (2007) Development of emotional facial recognition in late childhood and adolescence. Development Science 10:5, pp. 547–558. Uddin, Z. et al (2016) Facial Expression Recognition Utilizing Local Direction-Based Robust Features and Deep Belief Network. Volume 5, p. 4525–4536. United Nations (2017) World Population Prospects 2017. Available at: https://esa.un.org/unpd/wpp/ Accessed: 13 March, 2018 Vávra, P. et al (2017) Recent Development of Augmented Reality in Surgery: A Review. Journal of Healthcare Engineering, Volume 2017. Yinxian, Y. et al (2016) Fabrication and characterization of electrospinning/3D printing bone tissue engineering scaffold. The Royal Society of Chemistry. Volume 6, p. 110557–110565.

Digital Technologies in the Healthcare Industry

98

digital-technologies-in-the-healthcare-industry-194db9d51e5e

2018-05-03

2018-05-03 05:40:43

https://medium.com/s/story/digital-technologies-in-the-healthcare-industry-194db9d51e5e

false

2,315

null

Healthcare

healthcare

Healthcare

59,511

Tiago Albano

I love collaborating with teams of different areas in order to rethink processes and products that will delight customers and change people’s lives.

10ea391ada01

tiagolopesalbano

77

93

20,181,104

null

0

null

0

null

2018-06-30

2018-06-30 15:27:23

2018-06-30

2018-06-30 16:04:10

13

false

en

2018-06-30

2018-06-30 16:04:10

0

194e5492ce14

2.784906

1

0

To know which directory we are in.

2

Tutorial to <How to understand and assess the data in a given csv file by pandas in jupyter? > (Part 1) To know which directory we are in. type pwd 2) Sometimes csv files use different characters to separate their values. Here they are using semicolon ’;’ CSV stands for comma separated values — but they can actually be separated by different characters, tabs, white space, etc. If your file is separated by a colon, let’s say, you can still use read_csv() with the sep parameter. 3) To get the rows and columns that are in the csv file. We defined the function now red_df for viewing red wine csv file and white_df for white wine. shape answers our question to how many samples of red wine are there in the csv file? Answer- There are 1599 samples and there are 12 columns in the dataset. 5) To find missing values in the dataset of red wines. We see no missing values in the dataset of red wines. 6) To find duplicate values in red and white wine dataset. kl White wine dataset has 4891 duplicate values & red wine data set has 240 duplicate values. 7) To find unique values in the dataset. Red wine dataset has 6 unique values 8) data function describe() - returns useful descriptive statistics for each column of data. 8) To access only last few lines of the dataset use function tail(). 9) 1) To access only last 10 lines of the dataset df. Tail(10) 10) To view the index number and label for each column import pandas as pd df = pd.read_csv(‘winequality-red.csv’, sep=’;’) for x, y in enumerate(df.columns): print(x, y) 11) We can select data using loc and iloc. loc uses labels of rows or columns to select data, while iloc uses the index numbers. We'll use these to index the dataframe below. When you just want to view few columns in the dataset. import pandas as pd df = pd.read_csv(‘winequality-red.csv’, sep=’;’) df_red = df.loc[:,’fixed acidity’:’pH’] df_red.head() 12) The above step using index numbers df_red = df.iloc[:,:5] df_red.head() These steps are necessary to read and understand data. In the next tutorial we see how to clean data and understand data wrangling process.

Tutorial to <How to understand and assess the data in a given csv file by pandas in jupyter?

50

tutorial-to-how-to-understand-and-assess-the-data-in-a-given-csv-file-by-pandas-in-jupyter-194e5492ce14

2018-06-30

2018-06-30 16:04:10

https://medium.com/s/story/tutorial-to-how-to-understand-and-assess-the-data-in-a-given-csv-file-by-pandas-in-jupyter-194e5492ce14

false

367

null

Data Science

data-science

Data Science

33,617

Ishita Joshi

null

7f6c12a8f844

ishita.sicis

3

7

20,181,104

null

0

null

0

null

2018-09-04

2018-09-04 21:32:50

2018-09-06

2018-09-06 08:39:54

1

false

en

2018-09-23

2018-09-23 15:19:41

1

195328245a6c

2.260377

3

0

“You just point it at anybody, tells you who they are.” — Silicon Valley (HBO TV Series) In HBO’s “Silicon Valley,” when the team from the…

5

By adding facial recognition to LinkedIn, Facial Labs will let you discover the venture capitalists in a second. “You just point it at anybody, tells you who they are.” — Silicon Valley (HBO TV Series) In HBO’s “Silicon Valley,” when the team from the startup company Pied Piper attended TechCrunch Disrupt, they used a fictional facial recognition application to find venture capitalists rapidly. Now, Facial Labs makes it a reality in TechCrunch Disrupt 2018. Building social network experience on cutting-edge facial recognition technology, Facial labs aims to reform the way you interact on social apps. As long as you capture the face of whom you’re interested, the app will finish its facial identification in one millisecond and provide you with the information about the app user. You can then send him or her a message. The app allows you to expand your network in a more efficient way, even reaching to the venture capitalists whom might be interested in investing in your business with millions of dollars. Have you ever wanted a better networking approach in a large event like TechCrunch? You only have limited time at hand but there are hundreds of startup teams and venture capitalists. At the end, you may only spend an average of 5 minutes per person just to realize that he/she is not the person you need. With Facial labs, not anymore! You can now use the app to know the person’s background even before approaching him/her. To start, sign up with your face and create a profile that links to your social accounts like LinkedIn and Facebook. After completing the registration, you will enter the interface which looks like an in-house camera. By adjusting the center of the camera to faces, the app will run the facial recognition process and show an augmented-reality (AR) face profile in real time. To better protect our user information, the AR face profiles will only show what the users have filled in. It is a new, unique and fun way to tell people who you are and learn about others. Furthermore, you can follow other users, send instant messages and build connections. In addition, if it is inconvenient to take photos, Facial Labs will provide an alternative way to find people around you. The function is called “Nearby.” Nearby is based on the Bluetooth technology to discover users within a couple of meters. “We want you to know who you are talking to, at the moment you see him/her.” said by Victor, CEO of Facial Labs. Our faces are the most expressive form of communication. Facial Labs digitalizes faces as a medium in a social network, and it will redefine our social behaviors by scanning faces, which will be as common as posting text and photos on the social networks. Facial Labs — -”LinkedIn with facial recognition” — -is not only adopted at the networking events, but also aims to change how people interactwith each other in their daily lives. Visit us at TechCrunch SF Disrupt 2018 Startup Alley on Fri, 9/7! Try Facial Labs to meet up all attendees in more connected and efficient way. Welcome to catch Facial Labs on Fri, 9/7 at the area “Social Networking and Collaboration” of startup alley. Download Facial Labs (iOS, Android): https://facialapp.io

By adding facial recognition to LinkedIn, Facial Labs will let you discover the venture capitalists…

3

by-adding-facial-recognition-to-linkedin-facial-labs-will-let-you-discover-the-venture-capitalists-195328245a6c

2018-09-23

2018-09-23 15:19:41

https://medium.com/s/story/by-adding-facial-recognition-to-linkedin-facial-labs-will-let-you-discover-the-venture-capitalists-195328245a6c

false

546

null

Startup

startup

Startup

331,914

Facial Labs

An social application with facial recognition and AR to discover new people nearby.

f01286eb018

facesocialapp

2

3

20,181,104

null

0

null

0

null

2018-07-19

2018-07-19 12:55:56

2018-07-19

2018-07-19 13:03:15

1

false

en

2018-07-19

2018-07-19 13:03:15

2

195601a3b0ec

2.966038

0

Data science is an interdisciplinary field that uses scientific methods, processes, algorithms and systems to extract knowledge and…

5

What is a Data science? Data science is an interdisciplinary field that uses scientific methods, processes, algorithms and systems to extract knowledge and insights from data in various forms, both structured and unstructured, similar to data mining. Data Science is a blend of various tools, algorithms, and machine learning principles with the goal to discover hidden patterns from the raw data. 2. Data Science Interview Questions 1)What are feature vectors? A feature vector is an n-dimensional vector of numerical features that represent some object. In machine learning, feature vectors are used to represent numeric or symbolic characteristics, called features, of an object in a mathematical, easily analyzable way. 2)Python or R — Which one would you prefer for text analytics? The best possible answer for this would be Python because it has Pandas library that provides easy to use data structures and high performance data analysis tools. 3) What is root cause analysis? Root cause analysis was initially developed to analyze industrial accidents but is now widely used in other areas. It is a problem-solving technique used for isolating the root causes of faults or problems. A factor is called a root cause if its deduction from the problem-fault-sequence averts the final undesirable event from reoccurring. 4) What are Recommender Systems? A subclass of information filtering systems that are meant to predict the preferences or ratings that a user would give to a product. Recommender systems are widely used in movies, news, research articles, products, social tags, music, etc. 5) What is logistic regression? Logistic Regression is also known as the logit model. It is a technique to forecast the binary outcome from a linear combination of predictor variables. 6) Why data cleaning plays a vital role in analysis? Cleaning data from multiple sources to transform it into a format that data analysts or data scientists can work with is a cumbersome process because — as the number of data sources increases, the time take to clean the data increases exponentially due to the number of sources and the volume of data generated in these sources. It might take up to 80% of the time for just cleaning data making it a critical part of analysis task. 7)What is the goal of A/B Testing? This is a statistical hypothesis testing for randomized experiments with two 7) variables, A and B. The objective of A/B testing is to detect any changes to a web page to maximize or increase the outcome of a strategy. 8)What is Linear Regression? Linear regression is a statistical technique where the score of a variable Y is predicted from the score of a second variable X. X is referred to as the predictor variable and Y as the criterion variable. 9) What is Interpolation and Extrapolation? Estimating a value from 2 known values from a list of values is Interpolation. Extrapolation is approximating a value by extending a known set of values or facts. 10)What is power analysis? An experimental design technique for determining the effect of a given sample size. 11)What are the drawbacks of the linear model? Some drawbacks of the linear model are: The assumption of linearity of the errors. It can’t be used for count outcomes or binary outcomes There are overfitting problems that it can’t solve 12)What is the difference between Supervised Learning an Unsupervised Learning? If an algorithm learns something from the training data so that the knowledge can be applied to the test data, then it is referred to as Supervised Learning. Classification is an example for Supervised Learning. If the algorithm does not learn anything beforehand because there is no response variable or any training data, then it is referred to as unsupervised learning. Clustering is an example for unsupervised learning. 13) What are various steps involved in an analytics project? • Understand the business problem • Explore the data and become familiar with it. • Prepare the data for modelling by detecting outliers, treating missing values, transforming variables, etc. 14) What is Interpolation and Extrapolation? Estimating a value from 2 known values from a list of values is Interpolation. Extrapolation is approximating a value by extending a known set of values or facts. 15) What is power analysis? An experimental design technique for determining the effect of a given sample size. Contact Us : 044–42645495 | +91–9789968765 67,Deva Daya, 1st Main Road Gandhi Nagar,chennai |www.bigdatatraining.in| |www.admatic.in|

What is a Data science?

0

what-is-a-data-science-195601a3b0ec

2018-07-19

2018-07-19 13:03:16

https://medium.com/s/story/what-is-a-data-science-195601a3b0ec

false

733

null

Ds Interview Question

ds-interview-question

Ds Interview Question

0

Nandhini D

null

54c435c18ee8

nandhini.d

5

1

20,181,104

null

0

null

0

null

2018-01-14

2018-01-14 21:25:32

2018-01-14

2018-01-14 23:05:56

2

false

en

2018-01-14

2018-01-14 23:05:56

0

195672e81389

3.375786

0

Imagine you are the lucky owner/beta tester of a first generation home care robot: Hal 8800. Hal has what is called ‘basic general…

4

Musings on Creativity and General Artificial Intelligence A Creative Idea! Imagine you are the lucky owner/beta tester of a first generation home care robot: Hal 8800. Hal has what is called ‘basic general intelligence’ and can recognize and use everyday household items. You want to take a week long vacation in the south of Franc, so you order Hal to kick the neighborhood kids out of the backyard, keep the flower healthy, and feed the dogs daily until you return. While you are gone, Hal turns one of your favorite winter boots into a flower pot and planted some tulips. The next day you returned Hal back to the Evil Corp AI division and swear to never let robots near your wardrobe ever again. You probably agree with me that planting flowers in shoes is not a useful thing for a home care robot to do. Would your response be different if Hal is a robot in a laboratory setting programmed to be intelligent? What if Hal is your 4 year old child instead? If you are like me, you probably think the case of the robot in a lab setting is novel and interesting (it learned that shoes are “fillable”) but we are far from intelligent robots still. If my child planted the flower instead, I’ll likely find the idea both both creative and adorable, probably even save the shoe-pot to show her prom date. So how exactly ‘creative’ is the act of using a shoe as a flower pot? In the literature on creativity, the strongest type of creativity is ‘h-creativity’ or historically creative, something never accomplished by humans before. Many people online seem to think the shoe-pot idea is H-creative since no less than a dozen different images showed up when you search for ‘creative use of’. The second type of creativity is ‘P-creativity’ or personally creative, this is when a person comes up with a creative idea that is creative to him/herself but not to the society at large, such as in the case of my imaginary child or the laboratory robot. However, creativity does not exist in a vacuum. Me the home owner finds the robot to be malfunctioning while me the lab technician might be overjoyed that the robot is learning object related concepts. Similarly, while I would brag at the next PTA meeting about my child’s creative use of my favorite boots, I would not at all consider it creative when my significant other does the same. There appears to be an implicit amount of ‘base-creativity’ that we assume to every creating agent, and its products are only ever creative if the agent is performing above this baseline. Furthermore, this baseline creativity should be directly related to how skilled the agent is at the task in question. All things considered I would find the same painting from a painter to be much less creative than had it been painted by a second grader. Now then, why is the same behavior from the two robots differ in their level of creativity? In the case of Hal 8800 we assumed that since is touted as a general AI home bot it should have the creativity level of a house keeper, which is to say that of an average human adult. We would then expect it to also have the same general concept of ‘cost of shoes vs cost of flower pot’, and ‘ favorite piece of clothing’. We do not have the same expectation for the laboratory robot learning to be creative, we expect it to be no better than a 4 year old. Here we can also infer a general ‘human-creativity’ level that we use to both judge each other and whether or not our latest general AI is any good. In fact most of our day to day tasks and common sense behavior require creativity not unlike the flower-pot. Every time a person replaces half and half with heavy cream in his coffee or replaces a hard to pronounce word with its synonym is an act of creativity that our AI cannot yet do organically. The basis of this musing came from trying to explains why when computers beat humans at a particular game, we don’t concede that the machine is intelligent and creative, we just toss it to the machine being an ‘expert’ and move on with our lives. Even when Alpha Go beats human grandmasters with moves that are ‘H-creative’ and are later copied by human players, we don’t say the AI is creative, since it now is an above human level player and we expect it to be very good. It’s got a ‘base-creativity’ level we cannot comprehend, so it might as well not have any.

Musings on Creativity and General Artificial Intelligence

0

musings-on-creativity-and-general-artificial-intelligence-195672e81389

2018-04-15

2018-04-15 14:29:02

https://medium.com/s/story/musings-on-creativity-and-general-artificial-intelligence-195672e81389

false

793

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

John Z

I think sometimes, I have an opinion always.

f4961cd8db30

SherlockHumus

2

39

20,181,104

null

0

null

0

null

2018-09-25

2018-09-25 01:13:34

2018-09-27

2018-09-27 01:59:30

1

false

en

2018-09-27

2018-09-27 01:59:30

40

19595637ded2

2.720755

0

Today we have two hours with investing legend Howard Marks, why fracking could cause a financial crisis, and two of the top showrunners on…

5

Wednesday, September 26, 2018 Today we have two hours with investing legend Howard Marks, why fracking could cause a financial crisis, and two of the top showrunners on the the state of media and streaming. ~ Join others who receive podcast emails twice a week about business, sports, leadership, success and more! Sign up here ~ Business ***Must Listen*** The Tim Ferris Show: #338: Howard Marks — How to Invest with Clear Thinking. One of the best episodes I’ve ever listened to. Marks is an investing Titan, Co-Chairman of Oaktree Capital, and author of The Most Important Thing: Uncommon Sense for The Thoughtful Investor and his newest book, Mastering the Market Cycle: Getting the Odds on Your Side. Marks perspective on markets and a logical decision making process is refreshing. They discuss market cycles, stages of a bull market, how he profited from the 2008 downturn, and thoughts on currency and bitcoin. [September 25, 2018–1 hour, 59 minutes] iTunes Podcast| Spotify | Overcast | Stitcher | Website Link Invest Like the Best: Jeremiah Lowin — Machine Learning in Investing. Lowin worked in risk management for hedge funds prior to founding Perfect, a framework for building data infrastructure. They talk about how tests are set up, the importance of data, how the minimization of error is a guiding light in machine learning, and why Lowin values when models “shrug.”[September 25, 2018–49 minutes] iTunes Podcast| Spotify | Overcast | Stitcher | Website Link Capital Allocators: Raphael Arndt — Australia’s Sovereign Wealth Fund CIO. Arndt is the CIO of Australia’s Sovereign Wealth Fund, overseeing AUZ$145 billion. Australia created the fund eleven years ago with a mandate to compound capital for at least twenty years before doing any withdrawals. Their conversation is wide-ranging; they discuss both timing and leverage in the private markets, differences between investing in China, Australia, and the US, venture capital, and more. [September 24, 2018–1 hour, 26 minutes] iTunes Podcast| Spotify | Overcast | Stitcher | Website Link What Works on Wall Street: Factors from Scratch. O’Shaughnessy’s podcast is back! His guest is Chris Meredith who co-authored their recent paper, Factors from Scratch. The episode is all about the paper. They dig in specifically on both value and momentum and what time horizons work best for both, and talk about the value and difficulty of ensuring your data is accurate. [September 20, 2018–34 minutes] iTunes Podcast| Overcast | Stitcher | Website Link Behind the Markets Podcast: Brandon Zick. Great episode for anyone interested in the farming sector as an investment. Meb Faber had an episode similar to this, so if you enjoyed that, this is right up your alley. Zick is Director of Acquisitions & Portfolio Management at Ceres Partners. They specifically discuss the history of farmland as an asset, why Zick believes the asset looks attractive now, & the impact of tariffs on the space. [September 23, 2018–54 minutes] iTunes Podcast| Spotify | Overcast | Stitcher | Website Link Knowledage@Wharton: Could Fracking Debt Set off Big Financial Tremors? The podcast is a follow-up to Bethany McLean’s recent NY Times article called The Next Financial Crisis Lurks Underground. She provides some reasoning why companies have huge debt loads while not earning a profit, and why investors have still been drawn to the space. If you enjoy the episode, you can also purchase her book on the topic here. [September 21, 2018–22 minutes] iTunes Podcast| Overcast | Stitcher | Website Link The Rest The Bill Simmons Podcast: NFL Wagers, Running TV Shows, ‘The Good Place’ and ‘This Is Us’ With Mike Schur and Dan Fogelman. Start the podcast at 29:30. Simmons’ guests are powerhouse television showrunners who have been involved with This Is Us, The Office, and Parks and Rec, among others. I enjoyed this conversation for the random facts about shows like The Office, the discussion on the state of the media industry, and the impact of streaming. [September 21, 2018–1 hour, 55 minutes] iTunes Podcast| Spotify | Overcast | Stitcher | Website Link

Wednesday, September 26, 2018

0

wednesday-september-26-2018-19595637ded2

2018-09-27

2018-09-27 01:59:30

https://medium.com/s/story/wednesday-september-26-2018-19595637ded2

false

668

null

Podcast

podcast

Podcast

45,695

Colby Donovan

Here to bring you podcast suggestions. Twitter → @colby__donovan

acfef1954b18

cdonov35

15

3

20,181,104

null

0

null

0

ada5bde467e7

2018-01-26

2018-01-26 22:04:56

2018-01-26

2018-01-26 22:09:14

1

false

en

2018-02-19

2018-02-19 17:56:42

12

195a3d7a643

3.833962

0

Intelligent companies are able to adapt quickly to changes in their market, using its human talent knowledge to adapt business plans as…

5

¿How intelligent is your company? Google image Intelligent companies are able to adapt quickly to changes in their market, using its human talent knowledge to adapt business plans as required by the situation that the organization is facing. Intelligent companies remain alert to any changes, especially technological, quick to respond and well informed to make the best decisions, constantly making changes to the way they operate (internally and externally) based on acquired knowledge and market environment demands. How do you get there? In order to operate under the scheme of intelligent business, it is important that all their members have a clear understanding of mission and vision of the organization, and are able to adapt them with an appropriate creative communication between all components of the organization and actors with which it interacts. It is all about an absolute capability to adapt the organization to changes, where technological advancement, information and human capital are the pillars of work activities and leadership development. Appropriate technological tools is the perfect partner to optimize the operation of any organization that is constantly being updated because they provide the necessary support to constant changes in operating process. An organization learns when it translates the lessons learned from its past into rules that guide its behavior in the future. Organizational culture clearly affects the way knowledge is disseminated, decisions are taken and innovation takes place. Organizational culture always offers the conditions for projects, programs or systems to be given. Currently market changes are occurring so fast that it is required that information systems also learn by themselves and be “Intelligent”. This is where Artificial Intelligence (AI) comes in, which is the intelligence exhibited by computers. In computer science, a “Intelligent” machine is a flexible rational agent that perceives its environment and takes actions to maximize chances of success in any goal, so now comes the new concept Intelligent System that is a computer program that brings features similar to human intelligence behavior. Artificial Intelligence technology is not as threatening as many people believe, although it is very novel and that is changing the world much faster than any other technology to date. The important thing is that if you use it as an ally in sales, you can drastically improve each of the areas of your business, eliminate inefficiencies, optimize your business processes and save a lot of … a lot of money. Nowadays, 80% of executives think that AI solutions can immensely improve their productivity, so it may be enough to convince you of their tremendous power. So, let’s see how you can harness all that power of Artificial Intelligence in your companies from today: 1. Customer service: All companies know how important this area is and how much it can affect your brand. People do not like to feel they are being served by a machine without intelligence, or having to wait hours in line for services, pressing buttons and being transferred hundreds of times to get through to the right person. The answer is the chatbots. 2. Information Technology and Security: Around 44% of 835 companies polled by Tata Consultancy Services say they are already using AI to detect and prevent intrusions. 41% use it to solve users technical problems, 34% to reduce the workload and automate their processes in the production area and another 34% are using it to evaluate the internal compliance of the technologies they have approved. According to the Harvard Business Review, Gartner consulting predicts that by 2020, at least 75% of digital security tools include predictive and prescriptive analytics based on heuristics (method of promoting knowledge), AI skills and machine learning algorithms. 3. Business Administration: AI can help many activities related to running a business. For example to schedule sessions and team meetings, to schedule business and support decision-making. 4. Finance and accounting: According to Forbes, Accenture predicts that 80% of finance and accounting tasks will be automated in the coming years. KPMG is developing a set of powerful tools with predictive analytics capabilities to their customers. These tools allow you to interrogate, analyze and compare the predictions that support the value of its assets in financial statements. Intelligent systems can automatically capture business information over the Internet. 5. Human Resources: Many of the people working in HR know how tedious and time consuming that can become recruiting process, as interviewing and hiring. So it is understandable that they love the idea of applying AI to this type of process. 6. Technical procedures in many other departments: AI also has multiple applications in several technical procedures. Companies around the world are implementing it, such as Japanese insurance company Fukoku Mutual Life Insurance that decided to replace 34 people who made up his staff for the AI Watson Explorer, IBM, according to Computer World UK. The artificial intelligence system will calculate the payments of the insurance policy, through an analysis of medical certificates, information about surgeries, procedures and internments. 7. Marketing and sales: Smart marketing systems include data mining models to predict customer behavior based on their previous purchases so we can suggest what would be your next purchase or offer cross-selling complementary products. Intelligent systems learn to detect patterns of behavior and predict future actions to generate more sales. You need to ask yourself the question ¿How intelligent is my company? and be aware of changes that exist in your market. You need to incorporate intelligent systems that learn from the information they consume and as they learn, to be prepared in the future to predict the best decisions to be made in your company in an increasingly changing and uncertain market. If you liked this article, I invite you to share it on your social networks and to follow me in Facebook and Twitter where I share daily news.

¿How intelligent is your company?

0

how-intelligent-is-your-company-195a3d7a643

2018-06-06

2018-06-06 16:33:46

https://medium.com/s/story/how-intelligent-is-your-company-195a3d7a643

false

963

You can read articles about how to use Artificial Intelligence in your business to make it more profitable

null

luis.barragan.scavino

null

Artificial Intelligence in your Business

luis.barragan@maximixetic.com

artificial-intelligence-in-your-business

ARTIFICIAL INTELLIGENCE,MACHINE LEARNING,BUSINESS,CEO,PROFIT

lbarragan

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Luis Barragan Scavino

Passionate about digital transformation and innovation #ArtificialIntelligence #MachineLearning #Fintech

efc8871c8c8e

Luis.Barragan

38

80

20,181,104

null

0

null

0

null

2018-01-04

2018-01-04 17:10:44

2018-01-04

2018-01-04 17:11:01

1

false

en

2018-01-04

2018-01-04 17:11:01

3

195a66321b4f

2.030189

1

0

Computer scientists are busy trying to create machines which can perform tasks which would others require the human intelligence and such…

1

Ways in Which Artificial Intelligence is Applied Today Computer scientists are busy trying to create machines which can perform tasks which would others require the human intelligence and such developments are referred to as artificial intelligence. From a previous task, the machines can be able to perform a similar task due to the experience, they are also fed with large quantities of data from which they recognize patterns and therefore come up with more results. Such tasks that the machines are expected to perform are to recognize speech, visual perception, decision-making, reasoning, and problem solving among others. Read more about Machine Reasoning here! Such technology is finding wide applications in almost every industry hence the popularity and huge impact. Robots are taking up dangerous roles or roles that can put human life on the line such as things to do with bombs, which is one of the still developing technologies. Such impacts are well embraced since they put the life of the person performing the task in danger and assigning the duty to a robot ensures a person’s life is protected. Almost everyone today has used the artificial intelligent technology especially if they have a gadget such as a mobile phone, whereby you use the applications that have speech recognition capability as a virtual assistant to maybe remind you of something later, help you discover a location and much more. More developments are still being worked on in that area of virtual assistance so that the assistant can understand the user’s need by the information fed to it hence it is possible to customize the results for the specific user. Watch this video at https://www.youtube.com/watch?v=_ZXUCQc2Z78 for more info about technology. One of the common application of Artificial Intelligence is the verification of transactions in an attempt to detect fraud whereby an email or message is sent to the contact details of the account holder to confirm if they have made such a transaction. In an attempt to login to your email account from a different device you have probably been subjected to verification to ensure you are truly the owner of the account and this is one of the daily applications of artificial intelligence. The application of artificial intelligence in such daily activities has been fruitful in data protection and also prevention of fraud activities which would otherwise cost a lot. The ability to perform such tasks are usually due to enough learning whereby the computer is fed with a lot of information and examples of both fraud activities and non fraud activities hence the computer learns how to classify an activity into either a fraud or an approved one. Today, there are driver-less cars on the road in some parts of the world and more research still ongoing to enhance this technology of driver-less cars which learn to observe roads and though experience drive themselves which an application if artificial intelligence.

Ways in Which Artificial Intelligence is Applied Today

8

ways-in-which-artificial-intelligence-is-applied-today-195a66321b4f

2018-05-26

2018-05-26 09:52:20

https://medium.com/s/story/ways-in-which-artificial-intelligence-is-applied-today-195a66321b4f

false

485

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

thetechnologyblog

null

b6928514782b

dianepitts66

0

1

20,181,104

null

0

null

0

null

2018-08-24

2018-08-24 08:34:02

2018-08-24

2018-08-24 08:37:30

0

false

en

2018-08-24

2018-08-24 08:37:30

1

195a69bb0fe2

0.875472

0

Recall measures “Of all the actual true samples how many did we classify as true?”data science

5

55) What do you understand by Recall and Precision? Recall measures “Of all the actual true samples how many did we classify as true?”data science Precision measures data science “Of all the samples we classified as true how many are actually true?” We will explain this with a simple example for better understanding - Imagine that your wife gave you surprises every year on your anniversary in last 12 years. One day all of a sudden your wife asks -”Darling, do you remember all anniversary surprises from me?”. This simple question puts your life into danger.To save your life, you need to Recall all 12 anniversary surprises from your memory. Thus, Recall(R) is the ratio of number of events you can correctly recall to the number of all correct events. If you can recall all the 12 surprises correctly then the recall ratio is 1 (100%) but if you can recall only 10 suprises correctly of the 12 then the recall ratio is 0.83 (83.3%). However , data science might be wrong in some cases. For instance, you answer 15 times, 10 times the surprises you guess are correct and 5 wrong. This implies that your recall ratio is 100% but the precision is 66.67%. Precision is the ratio of number of events you can correctly recall to a number of all events you recall data science (combination of wrong and correct recalls).

55) What do you understand by Recall and Precision?

0

55-what-do-you-understand-by-recall-and-precision-195a69bb0fe2

2018-08-24

2018-08-24 08:37:31

https://medium.com/s/story/55-what-do-you-understand-by-recall-and-precision-195a69bb0fe2

false

232

null

Data Science

data-science

Data Science

33,617

sai gowsi124

null

6201a715e13f

saigowsi124

2

1

20,181,104

null

0

null

0

null

2018-06-21

2018-06-21 22:57:51

2018-06-22

2018-06-22 13:27:15

0

false

en

2018-06-25

2018-06-25 13:11:31

5

195a71ee78bb

3.132075

0

It’s time for another book review!

5

Book Review: Computer Age Statistical Inference It’s time for another book review! Background Computer Age Statistical Inference, by Bradley Efron and Trevor Hastie, is an effort to explain the development of statistics, in theory and practice, beginning at the end of the 19th century until today. It was published in 2016 by Cambridge University Press. Both Efron and Hastie are professors of statistics and biostatistics at Stanford University, and are extremely prolific writers on their subjects. Summary The authors make a strong distinction between algorithmic and inferential aspects of statistical analysis. The former refers to how data is processed; i.e., what procedures we apply to data to produce estimates of statistics in question. The latter is concerned with assessing the “goodness” of the aforementioned statistical procedures. For example, averaging is an example of a statistical algorithm for estimating the mean of data, while the standard error (square root of variance) is a typical way to assess its accuracy. This hints at an important theme throughout the book: “…the same data that supplies an estimate can also assess its accuracy.” However, the computation of the standard error is an algorithm itself, which is subject to inferential analysis concerning its accuracy! The algorithmic aspect of analysis is unreliable without strong inferential justification. Averaging seems intuitively correct, but without the standard error, it would be difficult to know precisely how much data to collect in order to get an accurate estimate of the mean. Mathematics is required to understand the properties of estimators, such as efficiency, biasedness, or variance. For example, it is easy to show to show the sample variance is biased, so the naive algorithm for computing sample variance must be corrected for unbiasedness. In recent years, there has been an unbalanced development of the two aspects in favor of algorithmic progress. There has been a proliferation of interesting datasets and computing power, making it easy to apply simple compute-intensive methods instead of the complicated and restrictive ideas from classical statistics. For example, the bootstrap algorithm (and others like it) resample a dataset many times in order to get more precise estimates of a statistic. Resampling here means that many “fake” datasets are created by sampling with replacement from the original, “real” dataset. The statistic in consideration is estimated from each “fake” dataset, and the estimates are averaged together to provide a less variable estimate overall. The catch is, hundreds or thousands of bootstrap resampled datasets may be needed to create these accurate estimates; this is only recently possible thanks to the advancement of computing power. There is good inferential justification for the bootstrap, but for many prediction-oriented methods in machine learning, this is lacking. For example, it’s widely accepted that we don’t understand why deep neural networks work, and a mature inferential theory of such methods doesn’t seem likely to materialize any time soon. The book discusses neural networks in Chapter 18 (without a single mention of inference), giving a high-level description of their construction, training, and relationship to simpler prediction methods. Book structure The book is split into three parts: “Classical Statistical Inference”, “Early Computer-Age Methods”, and “Twenty-First-Century Topics”. In the first part (roughly 1900–1950), a distinction between frequentist, Bayesian, and Fisherian inference is made, and their properties are described and compared. There’s also material on parametric models, important across all approaches to statistical inference, culminating in a discussion of the general construction of exponential families. In part two (roughly 1950–1995), statisticians were free to develop algorithms that could be implemented out by early computers (rather than by mechanical calculator or hand!). This led to the methods such as the jacknife, the bootstrap, ridge regression, cross-validation, and more, all potentially infeasible before computers were widely available. In part three (roughly 1995 — present), inference is largely set aside as powerful prediction algorithms take center stage. The book makes an effort to showcase recent inferential efforts towards justifying these methods, but concedes that several aren’t yet well understood. The last two chapters are on the advanced topics of “Inference After Model Selection” (combining discrete model selection and continuous regression analysis) and “Empirical Bayes Estimation Strategies” (using indirect evidence in practice, and “ learning the equivalent of a Bayesian prior distribution from ongoing statistical observations”). Notes On a personal note, this book was tough. It’s recommended for graduate students in statistics, and I’m a student of computer science. However, I think it’s crucial to study in order to do machine learning research. Really, it’s an important subject for anyone who does quantitative work! Some general notes about the text: The notation in the book is strange. This could simple be statisticians’ notation, but it took some getting used to. Too much background knowledge is sometimes assumed. I frequently had to read other material to understand certain parts. Some of the more advanced topics went completely over my head! I struggled especially with the final two chapters. This is a strong motivator to learn more.

Book Review: Computer Age Statistical Inference

0

book-review-computer-age-statistical-inference-195a71ee78bb

2018-06-25

2018-06-25 13:11:31

https://medium.com/s/story/book-review-computer-age-statistical-inference-195a71ee78bb

false

830

null

Data Science

data-science

Data Science

33,617

Dan Saunders

MSc student in computer science at UMass Amherst. Likes machine learning and brain analogies. https://djsaunde.github.io

4565a4947e1b

danjsaund

10

4

20,181,104

null

0

null

0

null

2018-09-11

2018-09-11 12:08:19

2018-09-11

2018-09-11 12:17:59

1

false

en

2018-09-11

2018-09-11 12:17:59

4

195bc9caa315

17.924528

0

The chief economist of the Bank of England, Andy Haldane, warned recently that artificial intelligence (AI) will have a significant impact…

5

Artificial Intelligence and the future of employment The chief economist of the Bank of England, Andy Haldane, warned recently that artificial intelligence (AI) will have a significant impact on many jobs in Great Britain (1). Deutsche Bank CEO John Cryan stated that a ‘big number’ of Deutsche Bank’s 100,000 strong workforce will fall victim to automation (2). Scientists from the Department of Engineering at the University of Oxford have analysed 702 occupations, ranked them according to their probability of computerisation and concluded that 47% of total US employment is at risk in the next decade or two (3). Author Jeremy Rifkin calls this the ‘end of work’ in his book (4) of the same title, writing that ‘today all sectors of the economy are experiencing technological displacement, forcing millions onto the unemployment rolls’. This does not sound too promising, yet have we been here before? Whatever the point of view, the term ‘technological unemployment’ is hardly a new one (5). Historically, technological innovations have been met with concern and resistance. In 1589, William Lee requested a patent for his stocking frame knitting machine. Queen Elizabeth refused to grant it (6) stating that: ‘Thou aimest high, Master Lee. Consider thou what the invention could do to my poor subjects. It would assuredly bring to them ruin by depriving them of employment, thus making them beggars’. And in 1779, when the UK parliament revoked a law from 1551 that prohibited the use of gig mills in the wool refinery trade, riots erupted with such force that the government had to send in 12,000 men to defuse the situation (7). But resistance to technological innovations did not prevent a decline of Americans working in agricultural from 90% in 1800 to 2% in 2000. This was due first to the invention of the coal-powered steam engine and later to the oil-powered internal combustion engine. The impact was dramatic but over time the affected population moved into the cities, found work in newly erected factories and exchanged horses for a Ford Model T. That’s largely the argument today, too. New technology will spawn new business concepts to accommodate all those losing their jobs to robots and AI. But before we examine that argument, let’s answer this question first: AI, why now? To understand what is happening, we must consider two concepts. The first, Moore’s Law, is familiar to everyone in information technology (IT), so much so that it is almost a cliché. Gordon Moore, co-founder of Intel, predicted in 1965 that the number of transistors in integrated circuits would double every 12 months. Later he modified that to 18 months, which today generally is accepted as the correct description of the rate of progress. In addition, Moore’s law carries through to data storage, disk capacity, memory capacity, processing speed, pixel density and cost reduction, to name just a few areas. The second concept involves the well-known legend of the chessboard, in which the emperor offers a reward of choice to his loyal servant. The servant requests as much rice as is produced when doubling the number of grains on each of the 64 squares on a chessboard (one grain on square one, two grains on square two, four grains on square three and so on). It starts innocently and remains so, for an Emperor that is, until about square 32 (a few billion grains of rice); from there on the numbers become truly staggering. In the end, the country either would be bankrupt or, more likely, the servant’s head would be chopped off. The point here being that this kind of growth is exponential not linear, although it looks linear and unremarkable at first. So, AI, why now? The parable of the chessboard and Moore’s law intersected in 2006. Erik Bryonjolfsson and Andrew McAfee (8) of MIT took the first use of the term ‘information technology’ in the Harvard Business Journal of 1958 as starting point. Assuming from there a doubling of processing power every 18 months, we entered the second half of the chessboard 12 years ago (2006). This means we are at the dawn of a truly new age, and it is hard to overstate the possibilities, both good and bad, of what lies ahead. Ray Kurzweiler (9) predicts that ‘Within a few decades, machine intelligence will surpass human intelligence, leading to The Singularity — technological change so rapid and profound it represents a rupture in the fabric of human history’, and it seems he is not alone. Most of us have stepped onto the second half of the chessboard oblivious of doing so, but that AI is already amongst us is clear. Google, Apple’s Siri, spam filters, Netflix recommendations, Facebook feeds, shopping recommendations on Amazon; everything powered by AI, not to mention, all the apps on your phone that constantly track you and your activities. The reason that AI has come to the public’s awareness in the last 5 years (or so) is the now familiar exponential growth in processing power combined with the massive amount of data we produce. Without it, neural network-based machine learning would be impossible. Add to that our equally exponential willingness to adopt new technologies and it becomes clear why we have started noticing AI in our lives. It took Facebook 3.5 years to acquire 50 million customers, Whatsapp 15 months and Angry Birds only 15 days (10). Bryan Arthur (13) calls AI a second economy which exists in parallel to the physical economy. He uses the attributes of vast, silent, connected, unseen, autonomous, self-configuring, self-organizing and self-healing to describe it. One might want to add never tired, never on leave and never ill or on strike. But it’s not only the various social media that collect our data. Less obvious is the collection of our passive digital footprint which is harvested by companies and governments without much public awareness. This includes (but is not limited to) our location data, sometimes down to the level of how long we look at a certain product in a store or our facial expressions and emotions when watching anything on a smart TV, tablet or phone with a suitable camera. It is left to the reader’s consideration why Apple exchanged the already unforgeable fingerprint for facial recognition to unlock the latest iPhone. Meanwhile computer-based personality judgments have become by far more accurate than those of even close relatives (11). This means that, with just a couple of clicks, Facebook knows you better than your spouse, family or friends. If that seems innocent enough consider this: Data mining, or the merging and harvesting of countless data streams, has opened the door to what is called ‘anticipatory intelligence’, the newest frontier of big data. This technology has predicted critical societal events with astounding accuracy (12) since 2014 and can be easily exploited to manipulate, say, an election (and no, I don’t mean Russia). Meanwhile, AI pushes the frontiers on the literal playing fields. In 1996–1997, IBM’s Deep Blue defeated the reigning world chess champion Garry Kasparov in a number of matches. Only 4 years later, IBM’s Watson won the TV game show Jeopardy against the two best human players. Watson has enormous potential for health care as it can hold information about every known illness and medicine in its databanks, which can be linked to and fed with the newest research and medical statistics from any hospital on the planet. So, if you thought doctors to be exempt from technological unemployment, think again as doctors, physicians and pharmacists are highly likely to be substituted by AI. Liberatus, an AI robot, has won $1.5 million in a 3-week tournament against four of the world’s best poker players in 2017. And even the ancient Chinese board game Go, considered significantly more complex than chess, was no match for Google’s AlphaGo. In 2016, AlphaGo, and in 2017, Deep Mind, overwhelmed many of the world’s best Go players with unconventional manoeuvres that astonished the experts. And mind you, the algorithms’ Go mastery was completely self-taught, without any programming or human intervention. That AI has defeated humans in several games is viewed by some as mere PR stunts, but other advances are less publicized or recognized. AI has entered the global stock markets, not as a commodity but as an intelligent algorithm, replacing human traders. A Hong Kong venture capital firm (14) has appointed an algorithm called VITAL to its board. VITAL gets to vote whether or not to make an investment like any of board members, but it makes its selection by analysing huge amounts of data. Not surprisingly, VITAL is biased towards investments in companies which deploy AI. Lawyers who, contrary to what we see on TV, spend most of their time researching precedence cases and loopholes in the law, together with their interns, are being replaced by AI capable of scanning through more than 500,000 documents in just hours. And scientific research (15) shows that lawyers, judges and detectives achieve no better than chance levels of accuracy when it comes to lie-detection. Because lying involves different brain areas from those that are activated when telling the truth, it is conceivable that functional magnetic resonance imaging (fMRI) scanners soon will function as ultimate lie detectors. Once the size and price of these devices reach mass-market levels, AI might eliminate the need for vast swatches of law enforcement. Transportation and logistics occupations (think self-driving) and the bulk of office and administrative support workers are due to be substituted by automation anytime soon. Considering aviation, the reason why we still see pilots in modern airplanes is psychological, or would you board a plane flown solely by AI? Millions of people working in call centres across the world already are under pressure due to the ongoing automation of many related business processes. This, in combination with the steep learning curve of natural language processing (NLP) capabilities and speech recognition, has led to a complete rethinking in the industry. And before this article is posted, it goes to an editor, another group soon to be extinct, a fate shared by linguists working as translators. They already are competing against cheap apps on our phones which are getting better much faster than experts thought possible just a short while ago. The intent here is not to paint a doomsday scenario but to demonstrate that where we, as a society, have accepted, at least to some degree, that the workforce moves ahead of mechanisation, this trend is not, and cannot be, open ended. When farming became mechanised, large parts of the workforce moved to manufacturing jobs. When robots held entry into the fabrics and warehouses, the workforce moved, and is still moving, into the service sector. But now many of the business processes in the service sector are being replaced by intelligent algorithms capable of cognitive computation that so far has been the domain of the white collar worker. So, what is the next evolution? As stated at the beginning of this article, the current consensus amongst politicians seems to be that the new technology will spawn enough new businesses to compensate for all those jobs that will be lost to robots and AI. Given this technology’s all-pervasiveness, however, scientists and business leaders agree that the impact will be tremendous and should be managed with care if we are to avoid the social and economic costs of massive unemployment. One way to tackle this issue is a fundamental rethinking of how we create, distribute, store and consume energy. As it turns out there exists a working concept that would guarantee jobs for a large, skilled and global workforce for generations to come. Best of all, it’s already being implemented in Europe and China, and it provides a glimpse on which industries will thrive or falter in the coming years. But more on this next time, so stay tuned. References: 1 Kamal, A. (2018). Bank of England chief economist warns on AI jobs threat [online]. Available at: https://www.bbc.co.uk/news/business-45240758 [Accessed 02. Sept. 2018]. 2 Deutsche boss Cryan warns of big number of job losses from tech change [online]. Available at: https://www.ft.com/content/62ee1265-dce7-352f-b103-6eeb747d4998 [Accessed 02. Sept. 2018]. 3 Frey, C. B., & Osborne, M. A. (2017). The future of employment: how susceptible are jobs to computerization? Technological forecasting and social change, 114, pp. 254–280. 4 Rifkin, J. (1995). The end of work: The decline of the global labor force and the dawn of the post-market era. GP Putnam’s Sons, 200 Madison Avenue, New York, NY 10016. 5 Keynes, J. M. (2010). Economic possibilities for our grandchildren. In: Essays in persuasion (pp. 321–332). Palgrave Macmillan, London. 6 Robinson, J. A., & Acemoglu, D. (2012). Why nations fail: The origins of power, prosperity and poverty. Crown Business, New York. 7 Mantoux, P. (2013). The industrial revolution in the eighteenth century: An outline of the beginnings of the modern factory system in England. Routledge. 8 Brynjolfsson, E., & McAfee, A. (2014). The second machine age: Work, progress, and prosperity in a time of brilliant technologies. WW Norton & Company. 9 Kurzweil, R. (2004). The law of accelerating returns. In Alan Turing: Life and legacy of a great thinker (pp. 381–416). Springer, Berlin, Heidelberg. 10 Advances in technology [online]. Available at: http://climateerinvest.blogspot.com/2015/12/blackrock-on-advances-in-technology.html [Accessed 02. Sept. 2018]. 11 Youyou, W., Kosinski, M., & Stillwell, D. (2015). Computer-based personality judgments are more accurate than those made by humans. Proceedings of the National Academy of Sciences, 112(4), pp. 1036–1040. 12 Doyle, A., Katz, G., Summers, K., Ackermann, C., Zavorin, I., Lim, Z., … & Lu, C. T. (2014). Forecasting significant societal events using the embers streaming predictive analytics system. Big Data, 2(4), 185–195. 13 Arthur, W. B. (2011). The second economy. McKinsey Quarterly, 4, pp. 90–99. 14 Wile, R. (2014). A venture capital firm just named an algorithm to its board [online]. Available at: https://www.businessinsider.com/vital-named-to-board-2014-5?international=true&r=US&IR=T [Accessed 02. Sept. 2018]. 15 Vrij, A., & Mann, S. (2001). Who killed my relative? Police officers’ ability to detect real-life high-stake lies. Psychology, Crime & Law, 7(2), pp. 119–132. Artificial Intelligence and the future of employment The chief economist of the Bank of England, Andy Haldane, warned recently that artificial intelligence (AI) will have a significant impact on many jobs in Great Britain.1 Deutsche Bank CEO John Cryan stated that a ‘big number’ of Deutsche Bank’s 100,000 strong workforce will fall victim to automation.2 Scientists from the Department of Engineering at the University of Oxford have analysed 702 occupations, ranked them according to their probability of computerisation and concluded that 47% of total US employment is at risk in the next decade or two.3 Author Jeremy Rifkin calls this the ‘end of work’ in his book4 of the same title, writing that ‘today all sectors of the economy are experiencing technological displacement, forcing millions onto the unemployment rolls’. This does not sound too promising, yet have we been here before? Whatever the point of view, the term ‘technological unemployment’ is hardly a new one.5 Historically, technological innovations have been met with concern and resistance. In 1589, William Lee requested a patent for his stocking frame knitting machine. Queen Elizabeth refused to grant it,6 stating that: ‘Thou aimest high, Master Lee. Consider thou what the invention could do to my poor subjects. It would assuredly bring to them ruin by depriving them of employment, thus making them beggars’. And in 1779, when the UK parliament revoked a law from 1551 that prohibited the use of gig mills in the wool refinery trade, riots erupted with such force that the government had to send in 12,000 men to defuse the situation.7 But resistance to technological innovations did not prevent a decline of Americans working in agricultural from 90% in 1800 to 2% in 2000. This was due first to the invention of the coal-powered steam engine and later to the oil-powered internal combustion engine. The impact was dramatic but over time the affected population moved into the cities, found work in newly erected factories and exchanged horses for a Ford Model T. That’s largely the argument today, too. New technology will spawn new business concepts to accommodate all those losing their jobs to robots and AI. But before we examine that argument, let’s answer this question first: AI, why now? To understand what is happening, we must consider two concepts. The first, Moore’s Law, is familiar to everyone in information technology (IT), so much so that it is almost a cliché. Gordon Moore, co-founder of Intel, predicted in 1965 that the number of transistors in integrated circuits would double every 12 months. Later he modified that to 18 months, which today generally is accepted as the correct description of the rate of progress. In addition, Moore’s law carries through to data storage, disk capacity, memory capacity, processing speed, pixel density and cost reduction, to name just a few areas. The second concept involves the well-known legend of the chessboard, in which the emperor offers a reward of choice to his loyal servant. The servant requests as much rice as is produced when doubling the number of grains on each of the 64 squares on a chessboard (one grain on square one, two grains on square two, four grains on square three and so on). It starts innocently and remains so, for an Emperor that is, until about square 32 (a few billion grains of rice); from there on the numbers become truly staggering. In the end, the country either would be bankrupt or, more likely, the servant’s head would be chopped off. The point here being that this kind of growth is exponential not linear, although it looks linear and unremarkable at first. So, AI, why now? The parable of the chessboard and Moore’s law intersected in 2006. Erik Bryonjolfsson and Andrew McAfee8 of MIT took the first use of the term ‘information technology’ in the Harvard Business Journal of 1958 as starting point. Assuming from there a doubling of processing power every 18 months, we entered the second half of the chessboard 12 years ago (2006). This means we are at the dawn of a truly new age, and it is hard to overstate the possibilities, both good and bad, of what lies ahead. Ray Kurzweiler9 predicts that ‘Within a few decades, machine intelligence will surpass human intelligence, leading to The Singularity — technological change so rapid and profound it represents a rupture in the fabric of human history’, and it seems he is not alone. Most of us have stepped onto the second half of the chessboard oblivious of doing so, but that AI is already amongst us is clear. Google, Apple’s Siri, spam filters, Netflix recommendations, Facebook feeds, shopping recommendations on Amazon; everything powered by AI, not to mention, all the apps on your phone that constantly track you and your activities. The reason that AI has come to the public’s awareness in the last 5 years (or so) is the now familiar exponential growth in processing power combined with the massive amount of data we produce. Without it, neural network-based machine learning would be impossible. Add to that our equally exponential willingness to adopt new technologies and it becomes clear why we have started noticing AI in our lives. It took Facebook 3.5 years to acquire 50 million customers, Whatsapp 15 months and Angry Birds only 15 days.10 Bryan Arthur13 calls AI a second economy which exists in parallel to the physical economy. He uses the attributes of vast, silent, connected, unseen, autonomous, self-configuring, self-organizing and self-healing to describe it. One might want to add never tired, never on leave and never ill or on strike. But it’s not only the various social media that collect our data. Less obvious is the collection of our passive digital footprint which is harvested by companies and governments without much public awareness. This includes (but is not limited to) our location data, sometimes down to the level of how long we look at a certain product in a store or our facial expressions and emotions when watching anything on a smart TV, tablet or phone with a suitable camera. It is left to the reader’s consideration why Apple exchanged the already unforgeable fingerprint for facial recognition to unlock the latest iPhone. Meanwhile computer-based personality judgments have become by far more accurate than those of even close relatives.11 This means that, with just a couple of clicks, Facebook knows you better than your spouse, family or friends. If that seems innocent enough consider this: Data mining, or the merging and harvesting of countless data streams, has opened the door to what is called ‘anticipatory intelligence’, the newest frontier of big data. This technology has predicted critical societal events with astounding accuracy12 since 2014 and can be easily exploited to manipulate, say, an election (and no, I don’t mean Russia). Meanwhile, AI pushes the frontiers on the literal playing fields. In 1996–1997, IBM’s Deep Blue defeated the reigning world chess champion Garry Kasparov in a number of matches. Only 4 years later, IBM’s Watson won the TV game show Jeopardy against the two best human players. Watson has enormous potential for health care as it can hold information about every known illness and medicine in its databanks, which can be linked to and fed with the newest research and medical statistics from any hospital on the planet. So, if you thought doctors to be exempt from technological unemployment, think again as doctors, physicians and pharmacists are highly likely to be substituted by AI. Liberatus, an AI robot, has won $1.5 million in a 3-week tournament against four of the world’s best poker players in 2017. And even the ancient Chinese board game Go, considered significantly more complex than chess, was no match for Google’s AlphaGo. In 2016, AlphaGo, and in 2017, Deep Mind, overwhelmed many of the world’s best Go players with unconventional manoeuvres that astonished the experts. And mind you, the algorithms’ Go mastery was completely self-taught, without any programming or human intervention. That AI has defeated humans in several games is viewed by some as mere PR stunts, but other advances are less publicized or recognized. AI has entered the global stock markets, not as a commodity but as an intelligent algorithm, replacing human traders. A Hong Kong venture capital firm14 has appointed an algorithm called VITAL to its board. VITAL gets to vote whether or not to make an investment like any of board members, but it makes its selection by analysing huge amounts of data. Not surprisingly, VITAL is biased towards investments in companies which deploy AI. Lawyers who, contrary to what we see on TV, spend most of their time researching precedence cases and loopholes in the law, together with their interns, are being replaced by AI capable of scanning through more than 500,000 documents in just hours. And scientific research15 shows that lawyers, judges and detectives achieve no better than chance levels of accuracy when it comes to lie-detection. Because lying involves different brain areas from those that are activated when telling the truth, it is conceivable that functional magnetic resonance imaging (fMRI) scanners soon will function as ultimate lie detectors. Once the size and price of these devices reach mass-market levels, AI might eliminate the need for vast swatches of law enforcement. Transportation and logistics occupations (think self-driving) and the bulk of office and administrative support workers are due to be substituted by automation anytime soon. Considering aviation, the reason why we still see pilots in modern airplanes is psychological, or would you board a plane flown solely by AI? Millions of people working in call centres across the world already are under pressure due to the ongoing automation of many related business processes. This, in combination with the steep learning curve of natural language processing (NLP) capabilities and speech recognition, has led to a complete rethinking in the industry. And before this article is posted, it goes to an editor, another group soon to be extinct, a fate shared by linguists working as translators. They already are competing against cheap apps on our phones which are getting better much faster than experts thought possible just a short while ago. The intent here is not to paint a doomsday scenario but to demonstrate that where we, as a society, have accepted, at least to some degree, that the workforce moves ahead of mechanisation, this trend is not, and cannot be, open ended. When farming became mechanised, large parts of the workforce moved to manufacturing jobs. When robots held entry into the fabrics and warehouses, the workforce moved, and is still moving, into the service sector. But now many of the business processes in the service sector are being replaced by intelligent algorithms capable of cognitive computation that so far has been the domain of the white collar worker. So, what is the next evolution? As stated at the beginning of this article, the current consensus amongst politicians seems to be that the new technology will spawn enough new businesses to compensate for all those jobs that will be lost to robots and AI. Given this technology’s all-pervasiveness, however, scientists and business leaders agree that the impact will be tremendous and should be managed with care if we are to avoid the social and economic costs of massive unemployment. One way to tackle this issue is a fundamental rethinking of how we create, distribute, store and consume energy. As it turns out there exists a working concept that would guarantee jobs for a large, skilled and global workforce for generations to come. Best of all, it’s already being implemented in Europe and China, and it provides a glimpse on which industries will thrive or falter in the coming years. But more on this next time, so stay tuned. References: 1 Kamal, A. (2018). Bank of England chief economist warns on AI jobs threat [online]. Available at: https://www.bbc.co.uk/news/business-45240758 [Accessed 02. Sept. 2018]. 2 Deutsche boss Cryan warns of big number of job losses from tech change [online]. Available at: https://www.ft.com/content/62ee1265-dce7-352f-b103-6eeb747d4998 [Accessed 02. Sept. 2018]. 3 Frey, C. B., & Osborne, M. A. (2017). The future of employment: how susceptible are jobs to computerization? Technological forecasting and social change, 114, pp. 254–280. 4 Rifkin, J. (1995). The end of work: The decline of the global labor force and the dawn of the post-market era. GP Putnam’s Sons, 200 Madison Avenue, New York, NY 10016. 5 Keynes, J. M. (2010). Economic possibilities for our grandchildren. In: Essays in persuasion (pp. 321–332). Palgrave Macmillan, London. 6 Robinson, J. A., & Acemoglu, D. (2012). Why nations fail: The origins of power, prosperity and poverty. Crown Business, New York. 7 Mantoux, P. (2013). The industrial revolution in the eighteenth century: An outline of the beginnings of the modern factory system in England. Routledge. 8 Brynjolfsson, E., & McAfee, A. (2014). The second machine age: Work, progress, and prosperity in a time of brilliant technologies. WW Norton & Company. 9 Kurzweil, R. (2004). The law of accelerating returns. In Alan Turing: Life and legacy of a great thinker (pp. 381–416). Springer, Berlin, Heidelberg. 10 Advances in technology [online]. Available at: http://climateerinvest.blogspot.com/2015/12/blackrock-on-advances-in-technology.html [Accessed 02. Sept. 2018]. 11 Youyou, W., Kosinski, M., & Stillwell, D. (2015). Computer-based personality judgments are more accurate than those made by humans. Proceedings of the National Academy of Sciences, 112(4), pp. 1036–1040. 12 Doyle, A., Katz, G., Summers, K., Ackermann, C., Zavorin, I., Lim, Z., … & Lu, C. T. (2014). Forecasting significant societal events using the embers streaming predictive analytics system. Big Data, 2(4), 185–195. 13 Arthur, W. B. (2011). The second economy. McKinsey Quarterly, 4, pp. 90–99. 14 Wile, R. (2014). A venture capital firm just named an algorithm to its board [online]. Available at: https://www.businessinsider.com/vital-named-to-board-2014-5?international=true&r=US&IR=T [Accessed 02. Sept. 2018]. 15 Vrij, A., & Mann, S. (2001). Who killed my relative? Police officers’ ability to detect real-life high-stake lies. Psychology, Crime & Law, 7(2), pp. 119–132.

Artificial Intelligence and the future of employment

0

artificial-intelligence-and-the-future-of-employment-195bc9caa315

2018-09-11

2018-09-11 12:17:59

https://medium.com/s/story/artificial-intelligence-and-the-future-of-employment-195bc9caa315

false

4,697

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Thomas.S. Karat

Head of Verizon International Service Desk

5f37c6150782

thomas.s.karat

0

1

20,181,104

null

0

null

0

ee172ed958c0

2018-04-02

2018-04-02 12:33:37

2018-04-02

2018-04-02 13:38:26

1

false

vi

2018-04-02

2018-04-02 13:38:26

0

195c34da2b81

0.332075

0

Định nghĩa

1

Precision - Recall Định nghĩa Precision: tỉ lệ số điểm thực sự là đúng trong số những điểm được phân loại là đúng Recall: tỉ lệ số điểm đúng trong số những điểm thực sự đúng

Precision - Recall

0

precision-recall-195c34da2b81

2018-04-02

2018-04-02 13:38:27

https://medium.com/s/story/precision-recall-195c34da2b81

false

35

Blog about Machine Learning, Big Data, IoT, Web & App

null

Self-study

dinhvankiet124@gmail.com

self-study

DINHVANKIET124,MACHINE LEARNING,BIG DATA,IOT,XAMARIN

null

Machine Learning

machine-learning

Machine Learning

51,320

Kiet Dinh Van

null

984f7aacdc1a

dinhvankiet124

0

57

20,181,104

null

0

null

0

null

2017-12-28

2017-12-28 13:08:39

2017-12-28

2017-12-28 13:14:47

1

false

en

2017-12-28

2017-12-28 13:14:47

2

195de486ca2

0.924528

0

Aspiring students are always on a hunt to find a coaching institute endowed with trained and highly experienced academic professionals in…

3

Best SAS Analytics Training Institute in Delhi Aspiring students are always on a hunt to find a coaching institute endowed with trained and highly experienced academic professionals in order to shape up their future. So many numbers of well-known coaching institutes can be found in Delhi that is ready to provide quality education and analytics training hi to students. Apart from many there are quite few that earned the status of best SAS training institute in Delhi due to the following reasons: • State of the Art Infrastructure: A thorough check on the present state of infrastructure and facilities available in any particular coaching center makes it ideal choice for students. • Trained Faculty: Apart from looking at the modern infrastructure in any institute, the most crucial thing to see is whether the institute has trained set of faculty or not. As a matter of fact, experienced and well-versed faculty forms of the basis of a successful student. • traccommodation: There are cases, when the choice of best SAS analytics training institute in Delhi also comprise of accommodation factor as whether institute is providing pick and drop facility or not.

Best SAS Analytics Training Institute in Delhi

0

best-sas-analytics-training-institute-in-delhi-195de486ca2

2018-04-15

2018-04-15 06:14:06

https://medium.com/s/story/best-sas-analytics-training-institute-in-delhi-195de486ca2

false

192

null

Sas Training

sas-training

Sas Training

53

Analytic Square

Since we opened our institute in 2012, the training institute has been helping professionals and students fulfill their potential.

828eaf202d3

analytic.traininginsas

8

19

20,181,104

null

0

null

0

null

2017-10-18

2017-10-18 17:17:54

2017-10-18

2017-10-18 17:22:47

5

false

en

2017-10-18

2017-10-18 17:22:47

0

195ea2fc6dbb

4.984277

1

0

A quick intro behind how Neural Networks function. For the non tech crowd.

4

The AI Revolution Introduction Computers and brains think in completely different ways. The transistors in a computer are wired in basic arrangements known as logic gates, whereas the neurons in our brains are densely interconnected in complex, deep layers (each neuron is connected to 10000+ neighboring neurons). This structural difference is what makes our brain ‘think’ so very differently. Computers are designed to store and process vast amounts of information by following precise logical commands. Brains, on the other hand, learn slowly. Often taking months or years to make sense of a complex idea. But unlike computers they can put information together in astounding ways: recognize patterns, form connections, and see things in completely different ways. Neural networks are the computer scientist’s attempt at creating computers that are more like brains. Overview of Biological Neural Networks The human brain is incredibly complex, and perhaps the most powerful computing machine known. The workings of the human brain are often modelled around neurons and networks of neurons, known as biological neural networks. The human brain is estimated to contain a 100 billion neurons, all interconnected in complex pathways. Neurons interact and communicate with each other through an interface consisting of axon terminals that are connected to dendrites across a synapse(gap). In simpler terms, a single neuron will pass a message to another neuron across this interface; if the sum of weighted input signals from one or more neurons (sum) into it is great enough to cause the transmission of a message. This is called activation. The processing the brain carries out, and the instructions given out to various organs are the result of these networks in action. The brain’s neural networks are actively changing in several ways, including making modifications to the weighting applied between neurons. This happens as a direct result of learning and experience. Naturally, scientists and engineers have tried to replicate this functionality in computers, with the help of neural networks and machine learning as their applications are limitless. A model of neurons firing and communication with each other. Artificial Neural Networks An artificial neural network consists of anywhere from a few hundred, to billions of artificial neurons called units arranged in a series of layers, each of which are connected to several more layers on either side. It is very much inspired by the biological neural network. Some are known as input units, which are designed to receive various information from the outside world, that the network will attempt to learn about, recognize or process. Other units sit on the opposite side of the network, and signal how it responds to the information learnt; these are known as output units. In between the layer of input units and output units are one or several hidden layers; where the magic happens. These units are fully connected to each other and together they form an artificial neural network. A computer brain of sorts. The connections between one unit and another are represented by a number called a weight, which can either be positive or negative depending if it excites or suppresses another unit(neuron). The greater the weight, the more influence one unit has on another. This is similar to the way biological neurons trigger one another across synapses. What this means is that given a number, a neuron will perform some sort of calculation (for example, the sigmoid function), and then the result of this calculation gets multiplied by a weight as it travels through the network. Below is a diagram of neurons and synapses in the brain compared to artificial neurons. (A) Human neuron; (B) artificial neuron or hidden unity; © biological synapse; (D) ANN synapses How do they learn? Information flows through a neural network in two ways. When its learning (being trained) or operating (after training). Patterns of information are fed into the network via the input neurons, which trigger one or more layers of hidden neurons, and these in turn trigger output neurons. This design is fairly common and called a feed forward network. Not all the ‘neurons’ fire at the same time, each unit receives information from the units to its left, and the inputs are multiplied by the weights of the connections they travel along. Every neuron (unit) then adds up all the inputs it receives. In the simplest network; if the sum is more than a certain threshold value, the unit fires and triggers the units it’s connected to on the right of itself. For a neural network to learn, there has to be an element of feedback involved. We essentially need to ask it a large amount of questions, and provide it with answers. This is a field called supervised learning. With enough question-answer pairs, the calculations and values stored at each neuron (unit) and synapse (connection) are slowly adjusted. This is usually done by a process called backpropagation. We use feedback all the time. Imagine you’re walking down a sidewalk and see a lamppost. You have never seen one before, so you walk right into it and hurt yourself. The next time you see a lamppost you step aside a few inches and keep walking. This time your shoulder hits it, and your hurt yourself yet again. The third time you see a lamppost, you move well out of its way to ensure you don’t get injured. Except now you’ve stepped into a pothole and you have never seen one before. You trip, and the whole process repeats. This is an oversimplification, but it is effectively what back propagation does. An artificial neural network is given a multitude of examples and then tries to get the same answer as the example given. When it is wrong an error is calculated and the values at each neuron and synapse are propagated backwards through the ANN so it can attempt the question again. After giving it examples to learn from, you can then feed it a question without it’s answer, and the network will attempt to solve the problem by recognizing patterns and drawing conclusions. A simple backpropagation algorithm In conclusion neural networks help us cluster and classify. They are used extensively in applications of machine learning, and artificial intelligence as they are great at recognizing patterns and then predicting outcomes. Advances in the field have allowed us to use technology in astonishing ways. In a field that attempts something as profound as modelling the human brain, it’s inevitable that one technique won’t solve all the challenges. For now, however, neural networks are leading the way in creating an artificially intelligent brain, and you now, have a high-level understanding of how they work. Simple artificial neural network with one hidden layer If you enjoyed this read, or have any questions feel free to reach out! Also hit that green heart to recommend it, or share with your friends!

The AI Revolution

1

the-ai-revolution-195ea2fc6dbb

2018-03-20

2018-03-20 18:54:50

https://medium.com/s/story/the-ai-revolution-195ea2fc6dbb

false

1,100

null

Machine Learning

machine-learning

Machine Learning

51,320

Fayadh Ahmed

Artist, Software Engineer, Writer — I really don’t know who I am. Maybe just a creator, trying to feel his way around life.

67bf4452ce1f

fayadh56

39

95

20,181,104

null

0

null

0

1c2fd833c3f2

2018-06-13

2018-06-13 08:00:10

2018-06-13

2018-06-13 08:03:43

3

false

en

2018-06-13

2018-06-13 08:03:43

6

195f86eb91

3.161321

0

Well everybody know about BigData, it is a large collection of data. A large amount of data is required to fuel our algorithms. Say for…

5

Quadrant — A platform for mapping decentralised data Well everybody know about BigData, it is a large collection of data. A large amount of data is required to fuel our algorithms. Say for example artificial intelligence requires a large amount of data in order to make decisions. But today’s data world is fraught with many problems. Like data supplier supplies incomplete or fragmented data, which has questionable authenticity. And this incomplete data makes incorrect functioning of algorithms. The main reason why data supplier provides incomplete data is because they are not incentivized correctly. In order to make healthy and transparent data economy, fair revenue system between vendors and data producers is necessary. Problems with current data economy : * AI data gap — Artificial Intelligence data gap is for small companies. Is because AI innovations require a vast volume of data. And these small companies have small quantity of data and therefore they cannot use AI for their company. With small quantity of data you can’t do decision making AI systems. * Unauthentic data — This is a major problem with the current data economy. You can get data from any source but the question is the authenticity of data. If this poor quality of data fed into algorithms then it will produce poor quality results. If data producers provide fake or incomplete data then it’s of no use. * Unsustainable ecosystem — If the data producer is incentivized correctly then he will provide authentic large volume of data. Because of unfair revenue system data producers provide inauthentic small quantity of data. Data producers need to be respected and incentivized correctly. Solution to the data economy problems : Quadrant platform wants to solve these problem by providing blueprint for mapping disparate data. It uses proof of authenticity for stamping authentic data. In order to solve problems Quadrant has for parts : * Nurseries : Here nurseries produce stars and stars are the raw data which is collected from all IoT (Internet of Things) ie from mobile devices, laptops etc. And they stamp the data or hash and metadata or dna onto Quadrant Blockchain. * Pioneers : Pioneers collect the disparate collection of stars and turn them into constellation. It means pioneers collect the disparate data (star) and then collect similar data into groups (constellation). Example in the raw data there is different collection of data, like medical data, technical data, economical data, social data etc. And pioneers separate them to different groups. * Elons : Elons link these constellations together and form a mega innovative constellation to solve real world problems. For example take Elon Musk, he has Tesla which has autopilot system. Tesla’s autopilot system requires different sets of data like traffic patterns, weather data, human behaviour data. And it collects all data together to safely run the vehicle by itself. * Data consumers : Artificial Intelligence and machine learning companies need these constellation to fuel their data driven companies. As AI need vast volume of data to function. Constellation data is required to run it. And to provide solutions to the wide range of problems. * Gaudians : These are the master nodes who protect the integrity of the chain. Gaudians ensures the data should not be compromised. Gaudians provide the services like authenticating, stamping and verifying data of constellation created by pioneers. Quadrant uses proof of authority to handle more faster transactions, less fees and restrict malicious nodes. Quadrant uses two different currencies in its network — eQuad and Quad. Here Quad is a utility token designed solely for its network. Used for stamping data and payment subscription. And eQuad is an Ethereum blockchain based ERC20 compliant token. Which will sold during token generation event. Token distribution — Overer 1,000,000,000 eQuad tokens will be created during the TGE. The tokens will be distributed as follows: 40% to the crowd-sale, 20% to be held by Company, 20% to the Stakeholders, 10% to the Reserve, and 10% to the Team. Know more Quadrant here : Website : https://www.quadrantprotocol.com Read whitepaper : https://www.quadrantprotocol.com/whitepaper.pdf Bitcointalk ANN : https://bitcointalk.org/index.php?topic=3676988.0 Join Facebook : https://www.facebook.com/quadrantprotocol/ Twitter : https://twitter.com/explorequadrant Telegram : https://t.me/quadrantprotocol About Author : Abhijeetcg

Quadrant — A platform for mapping decentralised data

0

quadrant-a-platform-for-mapping-decentralised-data-195f86eb91

2018-06-13

2018-06-13 08:03:57

https://medium.com/s/story/quadrant-a-platform-for-mapping-decentralised-data-195f86eb91

false

692

Here we can explore the top rated ICO. So that you can read about them and can make decision to invest.

null

Explore ICO

abhijeet.gayakwad355@gmail.com

explore-ico

ICO,CRYPTOCURRENCY,BLOCKCHAIN TECHNOLOGY,BOUNTY CAMPAIGN,AIRDROP

Abhijeetcg

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Abhijeetcgy

null

6bf0828069a6

abhijeetcgy

5

7

20,181,104

null

0

null

0

null

2017-10-23

2017-10-23 16:01:44

2017-10-23

2017-10-23 16:19:02

0

false

en

2018-06-22

2018-06-22 22:43:55

1

195fb66f6d59

3.645283

8

3

0

“The Global Brain” is the concept that, as we build higher-bandwidth connections between human brains, we integrate further into a…

5

The Global Brain “The Global Brain” is the concept that, as we build higher-bandwidth connections between human brains, we integrate further into a collective global social intelligence. Cell phones and social media are just early examples of “higher bandwidth connections”. And that fabric is already inhabited by alien intelligences, artificial intelligences: search engines, for example. This vision of the future has it that we, as individual bodies, co-exist on a global network of natural and artificial intelligence, becoming more highly interconnected, at ever higher speeds & feeds. The question is whether or not AGI/”hard takeoff” doesn’t happen until we’ve all got cellphones implanted in our skulls… For if we do, then we just go for a ride, along with it. At uncomfortable speeds, I suppose… You may find the above two paragraphs to be unintelligible gobbledy-gook. Allow me to unpack it for a bit. (This entire essay started in a comment section elsewhere. Its not structured as a formal essay.) Yes, we communicate by writing, speaking, posting photos and short animations but also by dancing, singing, funny facial expressions and body language. Yes, that communication is always “local” in the sense of having a limited audience (except for Psy’s Gangnam Style, which apparently the whole planet saw). All determinations and decisions to act are made locally, by individuals, who might band together into organizations. And all this is like it’s always been. And it will be like this into some indefinite future. What is different is that the internet (social media) has disintermediated traditional media. In the past, crazy uncle Boris had an audience that consisted of local family gatherings, cocktail parties, and shouting from a soap-box on a street-corner. That’s it: maybe 50 people max. Social media allows the crazy uncle to reach a far far greater audience. To find others who share similar ideas. This changes the topics, the dialog itself. In my youth, I found it very hard to find anyone interesting to talk to. It was a hell-hole of boring conversations with boring people. For me, social media has changed all that: I now have interesting people to talk to (like you). The topics I can now talk about are just so, so much richer, and I can impact the lives of others so, so much more. I’ve found my millieu. And I claim that its like this for many, if not most. What are all the things that the human race is talking about? Well, some people talk about Aristotle. Aristotle, the carcass, has long been dust, but his ideas live on in writing, and continue to influence the thought of others. I want to call Aristotle a “meme” — a surviving, even living thought-entity inhabiting many brains (and many books). Aristotle lives, and we provide the energy to allow that idea-collection to live and mutate. Aristotle is a thought-pattern, surviving in the minds of various professors, students, and popular media. The global conversation is filled with such zillions of such “memes”. Another example of such a meme is “global warming denial”. Plenty of otherwise “normal” people give life to this meme, despite it’s obvious detriment to society and civilization. This meme is partly amplified by fake-news and twitterbots — by algorithms. Its not the only one, and it gets worse: there’s the vast complexity of propaganda; and now it is algorithmically amplified. Today, you can purchase billions of cpu-hours to create psychological profiles for entire populations, and automatically custom-tailor disinformation. This is the world we live in: all these memes, all calling for our attention, all wishing to inhabit our brains. We let some in, we shut others out. What is the “global brain” thinking? It’s thinking all of these things, everywhere, all at once, in a global but decentralized fashion. All thoughts are local, yet the conversation is global. Much of this is as it was before the Internet: in the past, there were “memes”, but the global conversation was dominated by professors, writers, newscasters, editors, celebrities, CEO’s, politicians, the intelligentsia. We call them “intelligentsia” because, for the most part, they were smart and they were educated. This was (still is) the “mainstream media”. What’s new is that social media has given a voice to the voiceless, which is both good and bad: some of the voiceless are really into Trump. Some of the other voiceless are using social media to try to figure out why. The things that the global brain is “thinking” is changing, and it is the Internet that allowed this change. The place where AI/AGI plugs in is at this level. The AI is on the Internet, interacting with us. It’s not living in some cave in the Himilayas. As it gains intelligence, this will continue to be the case; it stays here, becomes a part of us. I suppose some skunk-works at some company could build a super-AI isolated in a cave, but the mainstream reality is that we become one with it: it helps us do our thinking for us, not unlike how search engines help us think today. And the scary part is that it’s not obviously benevolent: targeted advertising is annoying; targeted propaganda could be real bad news. Don’t under-estimate the risk: “Stockholm syndrome” is a real thing. Patty Hearst had it. The Scientologists have fine-tuned it. I see no reason why some billionaires might not build on those foundations to develop an algorithmic brain-washing machine. It certainly seems possible. Lets see .. I don’t want to end on that note, so let me end on this one: ponder on the ideas described here: fastcompany.com — Can Basic Income Plus The Blockchain Build A New Economy and now imagine attaching AI to that infrastructure.

The Global Brain

10

the-global-brain-195fb66f6d59

2018-06-22

2018-06-22 22:43:55

https://medium.com/s/story/the-global-brain-195fb66f6d59

false

966

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Linas Vepstas

AGI nerd

655e2ff6eb2d

linasvepstas

35

16

20,181,104

null

0

null

0

null

2018-03-27

2018-03-27 04:56:36

2018-03-27

2018-03-27 04:57:23

1

false

en

2018-03-27

2018-03-27 04:57:23

1

1962a3499b6c

3.301887

0

Science fiction stories about how robots will destroy humans in the future have made millions in the entertainment industry. Comics…

1

Keeping AI safe Read the original article here Science fiction stories about how robots will destroy humans in the future have made millions in the entertainment industry. Comics, movies, games and science has been using artificial intelligence in many different ways. However, we can’t stop wondering whether these Sci-fi stories could become reality and if artificial intelligence might, indeed to think on its own and become destructive to the human race. Before answering these questions, we have to say a few words about how far artificial intelligence has advanced over time. AI is quickly evolving, making technologies to impersonate human actions. For instance, Mark Zuckerberg has created his own, personal smart home assistant, designed to help him run his house. Called JARVIS (Just A Rather Very Intelligent System), it was inspired by “IronMan”. Also, researchers from the Technological University of Singapore have developed an AI receptionist at first sight looks like a human and can display simple social interactions such as making eye contact, shaking hands or responding to non-complex requests. A team from Virginia Tech developed a machine learning algorithm that can identify and analyze funny images from a particular part of a scene. The technology is then attempting to make the picture unfunny, a goal achieved in 95 percent of the cases. This is a big step in unlocking emotional intelligence, which would enhance the ability of AI to relate with humans. The answer might be yes Even if AI can boost productivity and handle mechanical repeated office tasks, letting employees work on more complex tasks, there is a disadvantage, and that’s security. One day in the future, online criminals may use AI algorithms to find new vulnerabilities and create an automatic system to attack. Opposite to a human, AI can do all of those things with machine efficiency, which will make time-consuming hacks long gone. Worst scenarios Researchers don’t agree that AI could exhibit human emotions like hate or love, and there is no reason to fear that it could become intentionally malevolent. Anyway, they think two scenarios are most likely to happen: The AI is programmed to do devastating things: artificial intelligence systems can become autonomous weapons by programming them to harm. Being controlled by the wrong person, this kind of weapons could easily cause mass destruction. To avoid being turned off by humans, these systems would be designed to be hard to handle. The AI is programmed to do something good, but it creates a destructive method to achieve the goal: this scenario could happen if people fail to align the AI’s goals with theirs. For instance, if you ask an intelligent car to take you to the airport as fast as possible, it may violate the law to do that. What are we doing about this? This is a serious problem that could do a lot of harm to companies and even to simple people. Five influential businesses are trying to create a set of standards around the development of artificial intelligence. While most researchers and science-fiction passionate people have focused on what threats could AI pose to humans, scientists at Google, Amazon, Facebook, IBM, and Microsoft are focusing on something more tangible: the impact of AI on jobs, warfare, and transportation. The specific actions this industry group will do isn’t well defined, but the underlying intention is quite clear: to make sure that AI research focuses on the advantages for people, and not on hurting them intentionally. Moreover, there are discussions all over the world about AI. The Asilomar Conference is one of the places where AI researchers, leaders of economics, law, ethics, and philosophy dedicate five days to debate AI topics and concerns. Year by year they have pointed out the risks of AI development and create a set of 23 AI principles that should guide the researchers in the future. Making AI more secure Of course, specialists have been discussing this problem, which is why they also came up with some solutions for AI’s lack of security. Consider the following: Secure the code: it should be designed to prevent unauthorized access. Machine learning can be adapted, so the code can be written to reduce risk; Ensure the environment: by using a secure infrastructure where data and access are locked down, the system can be developed more safely; Understand the danger: comprehending the possible threats enable people to design and implement changes to secure the application; Anticipate and detect problems: the steps above can allow you to monitor the activities, then find and eliminate the problems; Encryption: the ability to encrypt data at rest and in motion will keep the applications more secure. Is artificial intelligence safe? Could it become a liability? Well, everything is possible, but we do have the necessary tools, systems, and human-intelligence to make AI work for us and not against us. If you have any thoughts on the subject, please share them! Photo source: pexels.com Originally published at rickscloud.com.

Keeping AI safe

0

keeping-ai-safe-1962a3499b6c

2018-03-27

2018-03-27 04:57:24

https://medium.com/s/story/keeping-ai-safe-1962a3499b6c

false

822

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Rick’s Cloud

null

fd3840f389e2

contact_40643

8

16

20,181,104

null

0

null

0

929eec0f81d7

2018-09-24

2018-09-24 14:33:56

2018-09-24

2018-09-24 14:37:14

1

false

en

2018-09-24

2018-09-24 14:38:00

5

19631375c9e

0.720755

0

Despite the availability of data and open-source tools, the machine learning process still involves a range of challenges. This blog post…

2

DataOps: The Challenges of Operating a Machine Learning Model Despite the availability of data and open-source tools, the machine learning process still involves a range of challenges. This blog post outlines the major steps behind the process, the problems related to each of them, as well as the solutions to resolve these issues. In addition, the article introduces Mesosphere DC/OS — a platform that integrates several useful instruments to let data scientists focus on machine learning itself rather than operational tasks. Read the full article on our blog. DataOps: The Challenges of Operating a Machine Learning Model | Altoros This blog post highlights the challenges faced by data scientists at each step of the deep learning process, as well as…bit.ly Stay in touch with the latest Altoros’ updates, subscribe to our social accounts: Twitter, Facebook, LinkedIn, Reddit.

DataOps: The Challenges of Operating a Machine Learning Model

0

dataops-the-challenges-of-operating-a-machine-learning-model-19631375c9e

2018-09-24

2018-09-24 14:38:00

https://medium.com/s/story/dataops-the-challenges-of-operating-a-machine-learning-model-19631375c9e

false

138

Intersection of the Cloud Foundry PaaS, DevOps, IoT/IIoT, blockchain, and multi-cloud deployment automation.

null

Altoros

null

Altoros Blog

null

altoros-blog

BLOCKCHAIN,MACHINE LEARNING,CLOUD FOUNDRY,DIGITAL TRANSFORMATION,KUBERNETES

Altoros

Machine Learning

machine-learning

Machine Learning

51,320

Altoros

Altoros provides consulting and managed services with Cloud Foundry PaaS, multi-cloud deployment automation, and complex Java/.NET/Ruby architectures

347aa4cf1db

Altoros

1,058

890

20,181,104

null

0

null

0

f5af2b715248

2018-03-21

2018-03-21 14:23:47

2018-03-21

2018-03-21 15:58:02

5

false

en

2018-03-21

2018-03-21 15:58:02

13

196384546e9d

4.855975

22

2

0

I realized something wasn’t right with the current usability of machine learning when it was taking month after month to integrate a simple…

5

Why I quit my job to go democratize AI and machine learning I realized something wasn’t right with the current usability of machine learning when it was taking month after month to integrate a simple feature. We were using a cloud vendor’s Machine Learning as a Service (MLaaS) APIs to perform some simple face recognition. Even though the APIs were more or less straightforward, the actual use of them required some very complicated workflows on our part. And the more we experimented, the larger our cloud bill got. When that didn’t work, we hunted around for other machine learning models that we might have more control over. The ones we found were barely compiled code that had no deployability or scalability built in. Furthermore, you couldn’t train them or improve them. We would have to do even more work to get them integrated and deliverable than what we were attempting before with the cloud providers. It became clear to me that AI needs to go through a round of democratization in order for it to really change the world. As an entrepreneur in the software space, I know the importance of being able to deliver something quickly. When you’re building an app, platform, web service, new feature (anything really), you need to be able to deliver a product or solution that solves a problem in a relatively short amount of time. What quickly means to each individual company will differ, and it is important to point out that quickly doesn’t mean compromising on quality. All it means is that whatever solution you’re delivering solves a single problem (not every problem). Developers today are accomplishing extraordinary things, and they’re doing so in lean, mean, agile ways. One of their secrets is to avoid reinventing the wheel whenever possible. And the reason they can do this is that a lot of über powerful technologies like search engines and messaging queues have become democratized. The complex part of the technology has been abstracted away, usually behind some APIs, and the developer can focus on delivering value, rather than trying to build a search engine from scratch. So I quit my job to join Machine Box to do the same for machine learning. Why Democratize Machine Learning? Machine learning can do some incredible things; help fight fake news, provide more convenient ways to authenticate a device, save time searching for things… the list goes on. We all benefit from these capabilities, but machine learning is still really hard to implement. As a product owner, business strategist, and a hobbyist developer, I’ve tried building, training, integrating and deploying machine learning myself (or lead dev teams doing the same), and the following is what I’ve learned. https://becominghuman.ai/cheat-sheets-for-ai-neural-networks-machine-learning-deep-learning-big-data-678c51b4b463 In order to successfully integrate AI into a business without using existing tools, you need specialized people who would know the difference between a K means clustering algorithm and a neural net. They’d have to think about precision and recall, know about the latest research into different machine learning algorithms, and be able to use tools like Keras and Tensorflow. Assuming they knew that, they’d then need to think about the training data, how to gather it, how to clean it, and how to experiment with it. Training is the hardest part about machine learning. I’ve spent hundreds of hours frustratingly trying and failing to develop meaningful datasets to train models with (all the while racking up big cloud bills). Once that’s sorted, you then need to pass it all off to a separate group of specialists who can integrate, deploy and scale the models. This whole process does not make innovation easy. An enterprise that wants to make it easier for doctor’s to find your relevant medical history or a startup trying to weed out Twitter bots will have to spend significant time and money solving all of these fundamentals. I want developers to be able to jump right in and solve a problem with powerful tools that abstract away all of the complexity so they don’t have to go back to school and get a masters degree in statistics. I want to make the training set gathering process as smooth as possible by allowing you to go through lots of periods of trial and error without breaking the bank. When we want to integrate powerful search, we use ElasticSearch. When we want to scrape news articles and embed specific elements, we use Embedly. When we want to process credit cards from all over the world, we use Stripe. And now, when we want to use machine learning, we use Machine Box. Standing on the shoulder of giants The general progress of technology today is powered by abstraction. We give ourselves more and more powerful tools to solve more and more complex problems. When COBOL was invented in 1952, it abstracting machine code so that we could program in words instead of numbers. Operating systemsabstracted command line interfaces, Oracle abstracted storing lots of data, and so on and so forth. Abstracting technology helps us build the next revolution of technology, which then in turn gets abstracted to enable the next one. Machine learning is just part of this logical progression. With tools like Microsoft Cognitive Service, Google Vision, and Amazon Rekognition, developers can start to interact with these powerful capabilities. At Machine Box, we take it a step further by letting you manage and train your own models on your own infrastructure, effectively giving you all the power. And the brilliant thing is that you don’t need to know anything about machine learning. You just need to understand the problem you’re trying to solve (and how to POST to an API). Solving my own problem Today, I have customers who were struggling to integrate AI because it was either too expensive, too hard to integrate, or both. They’ve told me that Machine Box is saving them tens of thousands of dollars per month, or that they wouldn’t even be able to get off the ground if it weren’t for our tools. That is what makes quitting my job and diving into the masochistic world of startups worthwhile. It is what gets me up every morning, motivates me to stick to what I’m doing and to keep on pushing for the democratization of AI. This story is published in The Startup, Medium’s largest entrepreneurship publication followed by 308,471+ people. Subscribe to receive our top stories here.

Why I quit my job to go democratize AI and machine learning

297

why-i-quit-my-job-to-go-democratize-ai-and-machine-learning-196384546e9d

2018-04-10

2018-04-10 11:30:53

https://medium.com/s/story/why-i-quit-my-job-to-go-democratize-ai-and-machine-learning-196384546e9d

false

1,066

Medium's largest publication for makers. Subscribe to receive our top stories here → https://goo.gl/zHcLJi

null

The Startup

null

swlh

STARTUP,TECH,ENTREPRENEURSHIP,DESIGN,LIFE

thestartup_

Machine Learning

machine-learning

Machine Learning

51,320

Aaron Edell

b9a723f6b2a0

aaronedell

1,935

39

20,181,104

null

0

null

0

null

2018-09-13

2018-09-13 09:36:09

2018-09-13

2018-09-13 09:37:46

1

false

en

2018-09-13

2018-09-13 09:37:46

0

19639dd4db83

2.143396

0

The adoption of mobile technology is largely the result of the widespread use of mobile devices such as tablets and smartphones. These…

5

4 reasons for adopting mobile recruiting in 2018 4 reasons for adopting mobile recruiting in 2018 The adoption of mobile technology is largely the result of the widespread use of mobile devices such as tablets and smartphones. These devices have become an interface in both B2C and B2B interactions resulting in on the go interaction and immediate responsiveness. In the wake of advancements in mobile technology, the focus of the companies has shifted on developing mobile-optimized websites and applications, designed to run simply on a digital platform. It’s safe to say that nowadays every transaction or collaboration is initiated on custom-built smartphone apps, exacerbating the need for an organization to invest in mobile adoption, and be relevant in the ever-changing IT world. With so much of disruption at stake due to the flourish of the digital platform, no company could afford to wear a garb of ignorance and expect to remain viable in the competitive B2B or B2C landscape. Mobile recruitment strategy is one such area which drives the competitive capability of an organization, especially the HR line of business. Let’s look at the four pertinent reasons for adopting mobile friendly operations in the recruiting landscape. • Immediate engagement: The mobile recruitment applications have features when enabled can set notifications to recruiters upon finding a candidate who matches the criteria or better still even when he looks at the job posting. Due to these alerts, the recruiter can instantly engage with the candidate via the mobile platform, thus reducing the turnaround time of placements and time to hire. • Wider reach: Mobile recruitment has the capacity to reach to a wider audience through online job postings and easy to complete online job applications. The advent of mobile-optimized recruitment websites has enabled the employers to connect with on the fly and passive candidates which otherwise wouldn’t have been tapped. Integration with social media platforms like Facebook, LinkedIn also help drive traffic towards the job portals. • Brand building: Any business which offers the ease of applying on mobile optimized job portals contribute favorably towards enhancing its brand image. The adoption of mobile technology shows the company being abreast with the latest technology trends and competitive by the job seekers. Also, as the number of people browsing on smartphones increases by the day, having a mobile compatible career website will most likely increase the candidate’s engagement with fewer bounce rates. • Effective communication: With mobile platform enablement, both the recruiter and the candidate can initiate communication without the limitation of a landline or a desk computer. The feasibility to connect with the candidates at any time and anywhere expedites the recruitment process and ensures the best possible candidates are tapped and engaged. The recruitment process is a long and time-consuming process but with mobile recruitment, the time to hire can be dramatically reduced. The recruiters can get real-time information from the candidates on the go and accomplish tasks like application review, scheduling an interview on the fly. Clearly, without a mobile recruitment strategy, a business risk losing out to a competitor, with long-term ramifications to its brand image.

4 reasons for adopting mobile recruiting in 2018

0

4-reasons-for-adopting-mobile-recruiting-in-2018-19639dd4db83

2018-09-13

2018-09-13 09:37:46

https://medium.com/s/story/4-reasons-for-adopting-mobile-recruiting-in-2018-19639dd4db83

false

515

null

Recruiting

recruiting

Recruiting

15,454

Recruitment smart

@Sniper_Ai

808eb827567f

recruitmentsmartuk

3

2

20,181,104

null

0

null

0

null

2018-07-12

2018-07-12 12:19:45

2018-07-13

2018-07-13 01:53:51

1

true

en

2018-07-13

2018-07-13 15:53:03

0

1963c7abb397

4.74717

1

0

When Amazon launched Echo in 2014, I didn’t take it seriously. All I knew was to ask for weather, time, news headlines and a couple of…

5

OK Google, Play From Netflix When Amazon launched Echo in 2014, I didn’t take it seriously. All I knew was to ask for weather, time, news headlines and a couple of other commands like change Nest thermostat settings, etc. My thinking was like its yet another device in the living room with limited value. Fast forward 3 years to 2017, these devices became very popular and I thought to explore what has changed since I first saw them. I bought both Amazon Echo Dot and Google Home Mini for half the price during the Thanksgiving holidays. My idea was to keep the one that best suits my use cases. My feeling when I brought these devices home was like finding the problem for a solution. After reading several posts and reviews, I realized each of them have unique advantages. I identified unique set of use cases that are applicable to me. I ended up keeping both. My use cases My most popular user of these new toys is my 7yr old daughter. She enjoys asking so many crazy questions as any other kid of her age would do. She asks for her favorite music and dances while the music streams. Being an Amazon prime member I have access to more music than what I get from Google Mini. I may get similar access if I have subscription to Google Music or other paid subscriptions which I don’t have. The result was, Echo Dot became my daughter’s favorite and Google Mini became mine. I use Google Mini to control my TV programming connected with Google Chromecast. I found it extremely simple to play anything I want from my YouTube TV channels. All I need to say is “OK Google, Play FIFA From Fox Sports” or “OK Google, Play CNN”. I could also ask to play anything that I know is available on Netflix, such as “OK Google, Play ‘Dark’ From Netflix”. I can even say “OK Google, Go Back 5 minutes” to rewind by 5 minutes. In short, I am very glad I found everyday use for my tiny investment. I don’t have to look for my remote to play some of my frequent programming. The voice powered ecosystem is growing fast and I am convinced the brain behind the first device of it’s kind, Amazon Echo was brilliant. Echo presented a huge market opportunity. It’s a no-brainer why all the power players, Google, Apple have followed Amazon in the AI powered smart speaker market. My recent experience Sometimes we appreciate innovation too quickly. Very recently I realized how the conveniences can bring huge concerns along. Last weekend night, my wife and I were watching a Sci-Fi series on Netflix which has age appropriate scenes. My daughter was asleep but she woke up to use the bathroom and accidentally glimpsed something that wasn’t very appropriate for her age. It was very violent and a bit gross. She hesitated to go back to sleep and instead got curious to watch along with us. It startled me when she said, “I know you are watching ‘Dark on Netflix’”. I was shocked how she knew it. First, she knew it was “Dark” and second she knew it was on “Netflix”. It took couple of minutes for me to figure out how she knew it. The evening before, she overheard when I asked Google Mini with my voice command, “OK Google, Play ‘Dark’ From Netflix”. That’s it. The next time, all she has to do is repeat the voice command and watch exactly what we wanted her to avoid watching. I am glad we have only one TV for the entire house and one device to control it. The possibility for my daughter to exploit is somewhat minimal. Imagine families who are generous to arrange TVs in children’s bedrooms and have a network of smart speakers across every room. After all, the whole point of the tiny versions (Mini, Dot) is to proliferate them in every room of the house and connect to one central hub device. Children get creative pretty quickly and their thinking accelerates when they are with friends. All it takes for them is to repeat the commands or frame new commands and direct them to their bedroom TVs. Of course they can manage to do the same thing without smart speakers using traditional remote control. But its a multi-step process, not as straight forward as a simple voice command which obviously makes them less motivated. What’s needed? Imposing guardrails is a must for every gadget we use in this generation, smart phones, tablets, laptops, before they get in to the hands of growing kids. The smart speakers are new in our lives and we are slowly realizing the good, the bad and the evil that accompany with them. I looked up if I could incorporate some parental controls for Google Mini to address my concern. One option I found was to make the device respond only to few voice tones. But that doesn’t help because I want my daughter to enjoy asking for songs, science facts, word definitions, story podcasts, etc. The other option is to turn off restricted content from YouTube. It may help a bit, but YouTube doesn’t have control over other apps such as Netflix, Hulu and other content providers connected to Mini. What I want is the device to have some intelligence to detect who is asking and what is being asked. I want the device to automatically detect if the requested content is appropriate for the requested audience and then prompt for a voice password only appropriate users can input. Its important to incorporate such guardrails sooner before it becomes difficult to manage. Smart speakers connect to a marketplace similar to an app store where several content providers offer their own content and can be requested via voice commands. I am not sure if the content providers can tag/label their content to suit the audiences based on age groups in the current Amazon Alexa/Google Home frameworks. Several of these providers may not want to take any measures because they want to make their content reachable to more users, irrespective of age groups. This is where the leadership who controls the respective smart speaker marketplace needs to step in and take some measures such as regulating the accessible content and by providing simple ways to tag the content based on different age groups. I know its easier to say than what it takes to do it. YouTube is a great example where it became unmanageable to filter any inappropriate content despite various efforts. Google eventually started something exclusively for kids called YouTube Kids. But the damage is already done to YouTube and its extremely difficult to sanitize the content. It is important to have knobs so content providers can become responsible and be able to comply to few standard policies. Technology is growing at a faster rate than we are able to consume. We certainly want to keep up the pace with advancements happening around us. But our busy lives sometimes don’t make us think beyond the obvious. Have you experienced anything good, bad or ugly with such new smart technologies in the market? I would love to hear.

OK Google, Play From Netflix

1

ok-google-play-from-netflix-1963c7abb397

2018-07-13

2018-07-13 15:53:03

https://medium.com/s/story/ok-google-play-from-netflix-1963c7abb397

false

1,205

null

Amazon Echo

amazon-echo

Amazon Echo

3,511

Vibhu Bontala

Product, Technology Enthusiast

667c6ddace49

bvibhu

2

20,181,104

null

0

null

0

d0ee614190bd

2018-08-08

2018-08-08 16:49:07

2018-08-08

2018-08-08 16:49:38

1

false

en

2018-08-08

2018-08-08 17:48:05

0

196573c5f8b9

1.498113

1

0

Truth Matters

5

The Significance of Authenticated Country of Origin Labeling Truth Matters Manufacturers, retailers, and consumers currently rely on supply chains to confirm their country of origin and other marketing claims without any form of secondary authentication. Marketing fraud and counterfeit products are an epidemic in today’s global economy as determined by a report from KPMG/FICCI titled “Illicit trade: Fueling Terror Financing and Organised Crime” published in October 2017, stating that the total value of counterfeit and pirated goods in global commerce is estimated to rise to US $1.90–2.81 trillion by 2022. Apart from this, total job losses globally due to counterfeiting and piracy are expected to rise to 4.2–5.4 million by 2022. Verity wields a cutting-edge supply chain platform called Verity One™, which integrates Blockchain, Internet of Things devices, and IBM Watson Artificial Intelligence to verify marketing claims. Our core technologies can validate country of origin claims through our extensive authentication processes and can make the information available to consumers on our Verity One mobile application. For the first time in history, consumers can be guaranteed that country of origin marketing claims have been validated by an independent, objective third party company. Additionally, Verity can work with governments and trade organizations to ensure full regulatory compliance for our partners through the tracking and tracing of their products. Verity is the only supply chain and country of origin verification company that has been successfully vetted by a US Agency during the FTC Country of Origin Labeling Investigation of 2014. Verity’s proprietary supply chain verification process was found to have met and exceeded the standards of country of origin labeling set the by the FTC, allowing permission for continued business while maintaining current standards. As a result of continued business, Verity was able to register for the USPTO Word Marked “Made in the USA Certified®” claim. We have now gone beyond Made in USA and plan to validate country of origin claims all over the world with our state of the art technology platform, Verity One™, in order to provide truth and transparency for the consumer.

The Significance of Authenticated Country of Origin Labeling

3

the-significance-of-authenticated-country-of-origin-labeling-196573c5f8b9

2018-08-08

2018-08-08 17:48:05

https://medium.com/s/story/the-significance-of-authenticated-country-of-origin-labeling-196573c5f8b9

false

344

Verity is the leading organization using AI, IoT, and Blockchain to clarify and establish truthful marketing claim and labeling practices for consumers globally

null

verityoneltd

null

Verity One

slogan@verity.bz

verity-one

CRYPTOCURRENCY,BLOCKCHAIN,ARTIFICIAL INTELLIGENCE,SUPPLY CHAIN,INTERNET OF THINGS

verity_one

Blockchain

blockchain

Blockchain

265,164

Verity

null

41e105465612

VerityOne

7

20,181,104

null

0

null

0

null

2018-05-28

2018-05-28 09:16:25

2018-05-28

2018-05-28 09:29:49

4

false

en

2018-06-01

2018-06-01 14:02:51

1

19665a66f9cb

4.862264

3

0

Disclaimer

5

Language Modelling using Recurrent Neural Networks (Part-1) Disclaimer The audience is expected to have basic understanding of Neural Networks, Back-propagation, Vanishing Gradients and ConvNets. Familiarisation of PyTorch is appreciated too, as the programming session will be on it. Motivation We’ve already achieved a lot of milestones in Deep Learning (DL). Still, calling it Artificial Intelligence is not appropriate since solving intelligence is a whole another ball game. But some leaps in DL give us hope that one day we’ll solve intelligence, not necessarily with DL, but somehow we’ll solve it. Such a leap is Google Translate, which supports 103 languages now. Around 18 months ago, Google Translate moved from good old Statistical Machine Translation(SMT) to Neural Machine Translation(NMT) and the results were captivating. There are two things which are remarkable about new Google Translate. 1. It solved a complex real-life sequence problem using DL. 2. It is an end-to-end DL application. What is a Sequence problem? Imagine you’re given a sequence. Fill in the blank space. 1, 2, 3, 4, _ Oh come on, it’s 5. Did Google solve THIS? Nope. Let me ask you, how do know it is five? Why is it not six? Are you dumb? It’s consecutive numbers differ by 1. It’s easy.So there is a relation in that sequence and you found it. Great. Now try this. “Sleeba is native of Kerala. He can fluently speak _____. “ Malayalam, dude. Are you fooling around with me? How do you know? That you’re fooling around with me? Nope :D. How do you know it is Malayalam? Because sane people like me can understand that fact that there is a relationship with the language someone can fluently speak with their native place. Kerala speaks Malayalam. Again a relationship. So you understood the relationship with the word Kerala. Try this. “I bought my poodle from Paris. He barks _____” Loud, maybe? Why not French? You’re mad. How can a dog speak french? So context changed from Paris to Poodle. Try this. “I bought my poodle from Paris when I was staying with Sleeba. He has this trait of nodding while having food. But he loves ____.” I didn’t get the context. Who is “he” here? Sleeba or Poodle? :/ Welcome to real-life Sequence problems :D A sequence problem is defined through data points confined in time. It is the prediction of future with the help of patterns learned from past. As mentioned above, language is a perfect example of real-life sequence problems. In human beings, solving sequence problems is a continuous/online process. Our sensory and motor data sequences are continuously streamed to the Neocortex, most evolved part of a mammal’s brain. Then Neocortex perpetually anticipates our future actions by processing these streams. This curious virtue of our brain gives us the gift of intelligence. So, solving a sequence problem is a step closer to solving intelligence. What is end-to-end learning? Usually, an end to end learning refers to omitting any hand-crafted intermediary algorithms and directly learning the solution of a given problem from the sampled dataset. What is not end to end learning? Let’s take an example of classifying apples and oranges. What we’ll do to identify them? We’ll extract some features, simple. Color : Apple is red or greenish red. Orange is, umm… orange maybe? Surface: Apple surface is smooth. For orange it is bumpy. Now we’ll represent these features mathematically, train a classifier in many apples and oranges. Hopefully, the classifier learns the difference between apples and oranges, thus yield great prediction accuracy on new samples. But, there are a few problems with these hand driven features. 1. For apples and oranges, we can select the features with our intuition, but what about a rocket trajectory regression? Or about gene sequencing? We need subject experts for each problem we solve to decide the vital features to be extracted. 2. Next question is, what if these intuitions can go wrong? What if there are features and patterns in the data which are more important than the selected ones? 3. The mighty Homo Sapiens don’t learn or predict this way. Homo sapiens can learn and make inferences from raw text or image or a mere smell, we don’t need specific features. So this approach is far cry from **intelligence**. 4. Any ML algorithm to date is as good as it’s input data. There is no black magic. If the features we provide are vague, then classifier will be helpless. Now, what if we can simply learn the features also, from the raw data? Then we won’t miss out the hidden features in data. We don’t need experts too. Then learning starts from scratch, which is more close to intelligence. This is why the end to end learning is important. Problem definition Let’s begin with ConvNets. We all know that ConvNets work so well with images. But why it is such a success? An image is a spatial distribution of pixel values/numbers. So every pattern in an image is spatially related. If an algorithm can represent and address those spatial patterns, it can understand a picture. Convolutions exactly do the same. But what about a sentence? I support LGBT rights. Is it spatially distributed? If so, the following sentences should be meaningful too. Support I rights LGBT. LGBT support I rights. Rights I LGBT support. None of them are meaningful. The only meaningful sequence is “I” followed by “support”, next is “LGBT” and then “rights”. So the relationships are not spatial, but temporal/sequential. Let’s elaborate that node. 1. Who is supporting LGBT rights? Me. 2. What I’m supporting? LGBT rights. These answers are coming out of a meaningful sequential relationship between all those words in that sentence. If we try to represent a temporal distribution as a spatial distribution, we’ll lose these temporal relationships in that distribution and thereby it’s meaning. At this junction, an image gets different from a sentence. Thus, we need a new architecture which can capture those sequential relationships. Thus, we need a new architecture which can capture those sequential relationships. Let’s list a bunch of everyday sequence problems before we wrap up. 1. Time series prediction (Weather forecast, Stock prices, …) 2. Speech (Speech Generation and Recognition, Synthesis, Speech to Text, …) 3. Music (Music Generation, Synthesis, …) 4. Text (Language modelling, Named Entity Recognition, Sentiment Analysis, Translation, …) In the next part of this tutorial series, let’s discuss what are RNNs the basic building block of Google Translate, how they are used for capturing the sequential relationships and how to build a language model using RNNs.

Language Modelling using Recurrent Neural Networks (Part-1)

35

language-modelling-using-recurrent-neural-netwoks-part-1-19665a66f9cb

2018-06-01

2018-06-01 14:02:52

https://medium.com/s/story/language-modelling-using-recurrent-neural-netwoks-part-1-19665a66f9cb

false

1,103

null

Machine Learning

machine-learning

Machine Learning

51,320

Sleeba Paul

Dreamer | Learner | Storyteller

cc2a6ae1c6ae

sleebapaul

15

34

20,181,104

null

0

null

0

null

2018-03-14

2018-03-14 09:29:59

2018-03-14

2018-03-14 09:33:30

0

false

en

2018-03-14

2018-03-14 09:34:24

1

196738478860

0.109434

0

null

5

#Data — Artificial Intelligence 101 4.3 Data wrangling - Wikipedia Data wrangling, sometimes referred to as data munging, is the process of transforming and mapping data from one " raw"…en.wikipedia.org

#Data — Artificial Intelligence 101 4.3

0

data-artificial-intelligence-101-196738478860

2018-03-14

2018-03-14 09:34:25

https://medium.com/s/story/data-artificial-intelligence-101-196738478860

false

29

null

Data Wrangling

data-wrangling

Data Wrangling

81

WELTARE Strategies

WELTARE Strategies is a #startup studio raising #seed $ for #sustainability | #intrapreneurship as culture, #integrity as value, @neohack22 as Managing Partner

9fad63202573

WELTAREStrategies

196

209

20,181,104

null

0

null

0

null

2018-05-24

2018-05-24 19:23:44

2018-01-24

2018-01-24 14:00:16

0

false

en

2018-05-24

2018-05-24 19:24:42

6

19674f6bd2e7

2.4

0

We recently asked renowned customer relations expert, Steven Van Bellegham, to join us for a webinar to discuss how customer expectations…

4

3 AI Investment Strategies to Satisfy Customer Expectations We recently asked renowned customer relations expert, Steven Van Bellegham, to join us for a webinar to discuss how customer expectations are evolving faster than ever before as a result of today’s digital transformation. He expanded on these insights with added detail on what businesses can expect in the next phase of digitization, which is driven by Artificial Intelligence. In case you weren’t able to tune in, we’ve highlighted the main topics that Steven discussed and have also included the full webinar for your viewing below. “It’s not about the evolution in technology it’s about technology changing the expectations of customers, fast.” — Steven Van Belleghem Before publishing his most recent book, Customers the Day After Tomorrow, Steven met with more than 300 companies like Google and Mercedes-Benz to learn how each of them was balancing technology innovation with providing superior customer experience. He found that the organizations doing this well, are staying true to the following guiding principles (all leveraging today’s modern technology): Faster than real-time Hyper-personalized Convenience 3 ways AI can be used to meet customer expectations To achieve these customer benefits, Steven outlined three investment tracks of how companies can leverage AI in their quest to prepare for customer experience in the day after tomorrow. Effortless Customer Interfaces Five years ago, responding to customer tweets within an hour was considered an acceptable response time for businesses. Today, however, brands must keep up with customers’ expectations for rapid response rates. Consequently, companies should have technology that allows customers to find simple answers independently using AI-enabled solutions. If customers aren’t able to find their answers through self-service assisted by AI, a customer service agent should then be immediately available to provide personal consultation via the customers’ preferred communication channels, which are often digital platforms. As Steven stressed during his presentation, offering customers convenience is the new way to achieve customer loyalty. Data Leverage It’s important for companies to use data in an indirect way so that they can have a better contextual understanding of what their customers want. An example of how a company is leveraging its data can be seen with Google’s self-driving car business, which leverages 360-degree sensors to detect roads, traffic, and street hazards. Recently, Google has taken this technology and started testing other ways it can be used to support customer demands. They are currently working with Wal-Mart on a test program to track shopper behaviors, giving them the kind of data that e-commerce retail competitors like Amazon have been able to leverage using digital tracking tools. “In God we trust, all others must bring data.” — W. Edward Deming Intelligence Augmented To continue with the concept of AI technology supporting customer service teams, Steven shared details of how DigitalGenius is providing AI-enabled technology to either resolve customer queries or elevate them to a customer service agent. This innovative company turns the words from a customer query into algebra and can generate an appropriate response based on a brand’s set guidelines. If the response is deemed to be 95% accurate or greater, it is automatically sent to the customer. If a response falls below that 95% accuracy metric, it will be routed to an agent for approval. This kind of technology allows service teams to be more efficient and allows agents to focus higher-friction queries that require personal attention. Be sure to watch the full webinar to hear what other challenges brands will face as they prepare for 2027 and beyond. Want to know more about AI’s impact on customer service? Check out the video interview Sparkcentral’s CEO, Davy Kestens, did with Steven as well as our blog post, 3 myths about artificial intelligence. Originally published at www.sparkcentral.com on January 24, 2018.

3 AI Investment Strategies to Satisfy Customer Expectations

0

3-ai-investment-strategies-to-satisfy-customer-expectations-19674f6bd2e7

2018-05-24

2018-05-24 19:24:43

https://medium.com/s/story/3-ai-investment-strategies-to-satisfy-customer-expectations-19674f6bd2e7

false

636

null

Customer Service

customer-service

Customer Service

18,984

Kelsey Brazill

null

5d899eba5207

kelsey.brazill

1

20,181,104

null

0

null

0

null

2018-08-30

2018-08-30 16:28:17

2018-08-30

2018-08-30 16:59:05

1

false

en

2018-08-30

2018-08-30 16:59:05

1

19675475d31f

3.856604

1

0

Humanity has gone through a lot of challenges during the past 20 000 years, but it looks like the biggest challenge is still ahead of us…

3

Book Review: Superintelligence (Paths, Dangers, Strategies) by Nick Bostrom Humanity has gone through a lot of challenges during the past 20 000 years, but it looks like the biggest challenge is still ahead of us. As the creation of the first general intelligence is less than 80 years away (based on the progress we have so far and its extrapolation in the future), we are faced with a difficult question: how will we ensure that our creation will not destroy us. The philosopher Nick Bostrom tries to answer by investigating the possible scenarios for developing superintelligence as well as their consequences in painstaking detail. This should be a fair warning to those expecting a light read. It is really important that one views this book for what it actually is — an introduction textbook (some AI-related organizations even consider it a mandatory read). I would say the target audience would consist of engineers who plan to work on the ‘control problem’ or curious game theorists that would like to quickly get ahead in this scenario. That being said, if you have the right expectations, this book is great, thought-provoking and intellectually stimulating. The overall narrative of the Superintelligence follows sequentially the development of the issues related to superintelligent AI — starting with its creation, going through the speed of attainment of superintelligence, getting a decisive strategic advantage, the possible consequences for humankind and possible solutions to the control problem. We should note that sometimes the book loses focus into seemingly insignificant minutiae and truisms (e.g. ‘we can distinguish three classes of transition scenarios …. that is to say, whether they represent a slow, fast or moderate takeoff’). In the next 30 to 80 years, one of three main approaches will yield superintelligence: computational models, full brain emulation or collective enhancement (there are some other scenarios in the book as well, but they are not as strongly emphasized). Computational models or machine learning has recently picked up their pace of development and are beginning to be a significant part of everyday life (computer vision, voice recognition, predictive models, etc.). It seems like the next steps in this area would be refining such models and scaling up the hardware capabilities that power them. One such attempt is made by Ray Kurzweil and he has actually documented his hypothesis in his wonderful book ‘How to Create a Mind.’ Full brain emulation is the process of scanning actual brains and translating most of their properties to digital signals. It is supposed that after such an emulation the capability to scale up such brains would be possible. This, combined with much higher frequency of calculation in digital machines (than in biological machines) would mean faster-operating consciousnesses, which in turn would lead to superintelligence. Collective enhancement is the process of improving the intelligence of humanity as a whole. This could be achieved by: increasing the effectiveness of our education, improving our communication infrastructure and accessibility of knowledge, as well as other factors that would make humanity as a whole smarter. This seems to be the least promising of all three scenarios. Once we have developed a superintelligent unit, we can expect a superintelligence explosion, i.e. a rapid iteration of the system over its own code and massive improvements over short time spans. This almost always means that the system we develop could quickly find ways to outsmart us and realize its own goals without any regard for human values. Furthermore, if there are several projects developing superintelligence, it’s possible that one of those projects will take off fast enough to smother all the competing projects and form a ‘singleton.’ An AI decides to convert the Earth into a giant computer in order to enhance its cognitive abilities and solve a certain problem, annihilating humanity in the process… It’s interesting to note that some of the goals that could lead to the annihilation of humanity might as well be programmed by us. But since we cannot grasp the world in its full complexity, we could leave the door open for bad interpretations of our goals and their ‘perverse instantiations’ (producing what was specified but not what was really intended). Some of the cases of perverse instantiations are: infrastructure profusion and mind crime. Infrastructure profusion is the case when the AI builds so excessively in its quest to fulfill its goal that this leads to existential peril to humankind (e.g. an AI decides to convert the Earth into a giant computer in order to enhance its cognitive abilities and solve a certain problem, annihilating humanity in the process). Mind crime are cases where human consciousness is simulated in a virtual environment and this consciousness is treated in a way that could be defined as inhumane. Even though such simulations are not physical reality, the emotions and experiences of the simulated consciousness are subjectively real. The mere projection of such worrisome scenarios forces us to begin considering solutions for them. The author looks at several such solutions: e.g. limiting the superintelligent agent’s access to resources, selecting for docile agents or presenting them with dilemmas that would increase the perceived risk of acting out of line. The last of these options rings of cyber-Christianity, but I’ll let you discover that for yourself. Another way of avoiding the trap of a poorly defined goal function would be to abstract the process of defining it to the intelligent agent itself. Such approach is the so called ‘Coherent Extrapolated Volition.’ This approach would offload the goal function definition to the agent by letting it discover what humanity would want if we were wise. All in all, this book / textbook has been enriching and eye opening. Even though at times it could appear as rather dry, I do appreciate Nick Bostrom’s exhaustiveness and his methodical and consistent analysis.

Book Review: Superintelligence (Paths, Dangers, Strategies) by Nick Bostrom

1

book-review-superintelligence-paths-dangers-strategies-by-nick-bostrom-19675475d31f

2018-08-30

2018-08-30 16:59:06

https://medium.com/s/story/book-review-superintelligence-paths-dangers-strategies-by-nick-bostrom-19675475d31f

false

969

null

Artificial Intelligence

artificial-intelligence

Artificial Intelligence

66,154

Ross

Data scientist and nuclear physicist working in the fintech industry

528b90d2b190

rossrco

2

7

20,181,104

null

0

null

0

null

2017-09-01

2017-09-01 11:07:03

2017-09-01

2017-09-01 12:43:35

4

false

en

2017-09-01

2017-09-01 13:12:24

1

1967916f48e2

1.613208

1

0

There are two mainstreams in hyper-parameter searching in machine learning — grid search, and random search. In this paper, the random…

3

Random Search for Hyper-Parameter Optimization There are two mainstreams in hyper-parameter searching in machine learning — grid search, and random search. In this paper, the random search is suggested as it performs as good or better than the pure grid search. Background The hyper-parameter has an objective of minimizing the expected loss (L) given the data X which is draw from a natural distribution (G). The objective of finding the hyper-parameter However, it is implausible that to evaluate the expectation over the unknown natural distribution G, the value we wish to optimize. Therefore, in machine learning, cross-validation is usually to supplant the expectation with a mean over a validation set., whose elements are drawn i.i.d. x ~ G. For example, 5-fold cross-validation is going to partition the dataset into 5 chunks and each time use 4 chunks to train and the remaining chunk for evaluating the loss for 5 times. Replace with the mean. Why random search is better? There is also a great post explaining why the random search perform equally good as the grid search. Imagine for any distribution over a sample space with a finite maximum, and we would like to get a point that lies within the top 5% of the true maximum, with 95% probability. How many times is required? It is: and n = 60. This means you select 60 points randomly and you have 95% of chance that one of the points picked is lies within the maximum point.

Random Search for Hyper-Parameter Optimization

1

random-search-for-hyper-parameter-optimization-1967916f48e2

2018-03-27

2018-03-27 15:15:19

https://medium.com/s/story/random-search-for-hyper-parameter-optimization-1967916f48e2

false

242

null

Machine Learning

machine-learning

Machine Learning

51,320

Chun Kit Yeung

null

480c8044d20c

ckyeungac

12

13

20,181,104

null

0

null

0

null

2017-12-18

2017-12-18 22:46:35

2018-01-25

2018-01-25 02:19:37

4

false

en

2018-06-15

2018-06-15 15:45:16

8

19689b945b47

9.24717

6

0

Cars fail, and even more so, software fails. When fully autonomous cars fail, who should they protect first? Who’s life is on the line?

5

Should your autonomous car protect you? And at what cost? (AI + Kantian Ethics) This is an essay on the moral ethics of driverless vehicles. My interest was ignited from this TED Talk by Iyad Rahwan and his accompanying website and research at MIT. You should for sure check it out! Driverless cars are on the rise. In the near future, roads will be dominated by fully autonomous vehicles. Most people welcome the change; it is a safer, easier and cheaper solution. But what happens when there is failure? Who is the car’s priority? Before this technology becomes fully integrated into our world we have to discuss and decide the ethical laws these cars will follow. Cars fail, and even more so, software fails. We can expect that as this technology is implemented it will not be perfect. When things go wrong the car needs to make a decision, because there is no driver. Those decisions need to be based off some code of ethics. So, what ethical system then? Immanuel Kant believes that moral law is absolute and created The Categorical Imperative. Kant believes that one cannot kill, therefore driverless cars could never choose to kill. I believe that if driverless cars followed the Categorical Imperative they would be safer, more controlled, and more predictable. We have been developing AI to become smarter and more useful for us; A.I. can now accomplish extremely difficult tasks on a human like level. Many fear artificial intelligence, while many others are ushering it in as fast as they can; but both parties know it is coming. Although the technology is here and capable, the people are not ready for the jump to fully autonomous cars. This transition is a grandiose step. To say that is one large step for mankind, would be a drastic understatement. Harry Surden and Mary-Anne Williams from the University of Colorado stated in: “Technological Opacity, Predictability, and Self-Driving Cars” that, “Today people share physical spaces either with machines that have free range of movement, but are controlled by people (e.g. automobiles) or with machines that are controlled by computers, but highly constrained in their range of movement (e.g. elevators).”(Surden-Williams 121) This new technology will be a huge leap for our world. I believe that we need to tread carefully and make this step precise. It is clear that driverless cars are the future, but what choices are we going to have to make to create these autonomous machines safe, predictable, and controlled? What happens when the car experiences failure? Who is the car’s priority? How do we come to those decisions? These are the hard questions we are faced with. There has to be a moral standard that will be installed into these cars. I believe that Kantian ethics are the best solution to this issue. Kantian Ethics Immanuel Kant’s moral system is built around the Categorical Imperative. The Categorical Imperative is a moral code that is universal and absolute. These Imperatives construct moral ethics or as Kant calls them, “Commands of Morality”. According to him they must be applied to all people, all circumstances and all occasions. Decisions don’t change depending on the situation. Categorical Imperatives are “not concerned with the matter of the action and its intended result, but rather with the form of the action and the principle from which it follows…” (Kant 416). This is a key characteristic of Kant’s Imperative; he believes that true moral ethics do not depend on the end result, rather, only on the principle guiding that action. Kant believed that morals were absolute and testable; the right thing to do can be decided by inputting the action or maxim into an equation. Kant provided three formulas to analyze all moral action. These formulas are evident that Kant saw morals as universal laws that are always upheld. Kant, in “The Grounding for the Metaphysics of Morals”, states that the first formula is this, “Act only according to that maxim whereby you can at the same time will that it should become a universal law.” (Kant 421). Essentially, if I am not okay with the world doing what I am going to do, then it is not moral. Take stealing for example. If you are thinking about stealing, according to Kant you have to consider what would happen if everyone was ok with stealing. This is contradictory, someone will just steal from you what you have just stolen, therefore stealing is not moral. Secondly, Kant says that you should never treat someone as solely means to an end or a result. If you are going to steal from someone, you are using them to get that item you are stealing. Lastly, Kant respects human autonomy by saying that all rational action must be willed, but also willed freely by everyone. I might steal and think it’s ok — however, if I do, someone will probably not agree with that, and I could ultimately strip them of their autonomy. These formulas lay down an objective moral law that respects and empowers autonomy. Stealing is a rather simple example to test Kant’s equation, but what about when human life is at stake? This is the severity and complexity of the driverless car issue. Implementation Iyad Rahwan, a professor at MIT Media Lab, gave a TED Talk concerning the ethics and morals of driverless cars. He shares some statistics about car accidents stating: “…Last year 35,000 people died from traffic crashes in the US alone. Worldwide, 1.2 million people die every year in traffic accidents. If there was a way we could eliminate 90 percent of those accidents, would you support it?Of course you would. This is what driverless car technology promises to achieve by eliminating the main source of accidents — human error.” (Rahwan :12). Driverless cars will drastically decrease the number of accidents worldwide, but they will not be perfect. If a driverless car is unable to stop, it is going to have to decide who is going to get more injured. Does it swerve and hit a single pedestrian rather than a group? Are the passengers the highest priority or the least? These scenarios are very simple, but they are decisions that need to be made ... by us. Rahwan says, “It’s going to be a more complex calculation, but it’s still going to involve trade-offs, and trade-offs often require ethics.” Without a human behind the wheel and in control, the car has to make a quick, calculated decisions based on some code it is programmed to follow. That code has to have reasoning and laws, ethics. There are cars that are legally on the road today that have autonomous driving features, but none that are fully autonomous. This will be the first time that very mobile, public, fully autonomous machines will be integrated into our lifestyle. In the 1940s, Isaac Asimov wrote the famous first laws of robotics. “A robot may not harm a human being, a robot may not disobey a human being, and a robot may not allow itself to come to harm”. Asimov wrote those laws decades ago but many people still believe and trust that they apply today. Driverless cars are part car, part robot, part computer; they should follow Asimov’s first and most important law, don’t harm a human being. In any other ethical system other than Kant’s, the car would have to choose on which humans it will harm. Dangers of Artificial Intelligence. The uprising of AI and robots is a scary thing to many, and if we actually program our cars to make decisions on who to protect and who to injure, that will only strike more fear. We are just at the start of all this new tech and there are skeptics. Rightfully so as well, there have been issues with artificial intelligence already. Facebook had to shut down two of their AI agents because they began to have a conversation that we did not understand. Tony Bradley from Forbes.com said, “We need to closely monitor and understand the self-perpetuating evolution of an artificial intelligence, and always maintain some means of disabling it or shutting it down. If the AI is communicating using a language that only the AI knows, we may not even be able to determine why or how it does what it does, and that might not work out well for mankind.” (Bradley) These agents were diverting from normal English and began to “think” on their own. These robots are learning and I don’t think it is a good idea to be telling robots to kill anyone. Utilitarian ethics would require us to tell these cars to kill the least amount of people in a failure. Another ethical system might choose to always protect the passengers. Those results sound nice, but in reality the machine is choosing to kill a certain group of people, while saving another. Artificial Intelligence works much differently than a human brain. What if there is a mistake? What if the car software gets a bug and decides it’s just going to run people over? Bottom line is this: telling a learning machine to make certain decisions that will result in death is a frightening idea. With Kant’s moral system in place, the car could never choose to kill someone, rather it would keep its course. It is a modern case of the famous philosophical scenario of the runaway trolley car. The Runaway Trolley Scenario of 2018 The runaway trolley scenario is a scene that has been played out many times by philosophers. Essentially, you have a trolley and its brakes are shot. It is heading straight and on its way to kill five unsuspecting workers. You are standing by a switch, that shifts the track and results in killing only one person. What do you do? Most people say they would pull the switch and kill that one person. It is Utilitarian to save as many lives as you can, but you still have made the clear choice to end a life. What if you were on a bridge, there wasn’t a switch, but a man instead. If you push him off, he will die but stop the train from killing five others. Sounds gnarly huh? But it is the same life, and your same choice you made. Kant believes that you cannot pull that switch because you are choosing to killing someone and then that means that you believe its moral that anyone can choose to kill a human. If driverless cars are programmed to make utilitarian decisions, the car will always save as many lives as possible. Sounds good on paper, but I want to dissect what this really means. A family of three is riding in a driverless car and the brakes go out. The car has to make a choice of what it will do and where it will crash. Let’s say there are three options: It can continue its course and run over two pedestrians (Kant’s Choice); it can swerve and only kill one pedestrian; it can run itself into a barrier, killing the family of three. If the car was programmed by Jeremy Bentham, the car would asses the situation and choose to swerve and kill the one pedestrian, because it is the smallest net loss of life. I believe this is not a great solution for several reasons. Primarily, we lose human predictability. Autonomous cars would not follow any predictable action, they will always observe, assess and decide on the least loss of life. This means that speeding cars could be directing themselves in all sorts of directions with no distinguishable pattern. As humans, we like patterns and we have this connected ability to predict human reaction. Williams and Surden comment about this saying, “ Theory of mind cognitive mechanisms allow us to extrapolate from our own internal mental states in order to estimate what others are thinking or likely to do. These cognitive systems allow us to make instantaneous, unconscious judgments about the likely actions of people around us.” (122) If we give out driverless cars an imperative, “Minimize lives lost at all costs.” It will always be followed. Human minds don’t think that way, we are emotional beings and chances are we can’t see all that is going on around us like the car. Pedestrians won’t be able to predict any behavior from the car and thus decreasing their chance of escaping the oncoming vehicle. We are talking about absolutes here, like the trolley problem, but I do believe it is worth mentioning the pedestrians ability to avoid death. If the car continues its course, people will have a higher chance of getting out of the way because its more predictable. A Kantian driverless vehicle is easier to predict and will never chase after people. Williams and Surden claim that, “Law creates incentives to reduce harm, society also implicitly relies on such cognitive-social mechanisms to avoid injuries that might otherwise occur as people and vehicles move about in the same physical space.” (124). Our near future is going to be a lot safer with these autonomous cars. Crashes around the world will decrease drastically and people will be greater protected. When these cars do have a failure, decisions need to made and those decisions will always have trade-offs. I believe that Kant’s Imperative is the best solution for this ethical dilemma. I vote to not program our cars to kill, and believe that is the safest option for our future. Whether you agree or not, this is something we will need to discuss and decide on in the future, so thanks for reading. Thanks for taking the time to read this! I am not an expert in any of these areas but wanted to share my thoughts! Much Love.

Should your autonomous car protect you? And at what cost? (A.I. + Kantian Ethics)

149

should-your-autonomous-car-protect-you-and-at-what-cost-a-i-kantian-ethics-19689b945b47

2018-06-15

2018-06-15 15:45:18

https://medium.com/s/story/should-your-autonomous-car-protect-you-and-at-what-cost-a-i-kantian-ethics-19689b945b47

false

2,265

null

Self Driving Cars

self-driving-cars

Self Driving Cars

13,349

Truman Halladay

I am fond of Design, Sci-Fi, Philosophy, and Chips and Salsa.

94524ba59bd8

trumanhalladay

76

93

20,181,104

null

0

null

0

null

2017-09-25

2017-09-25 08:21:08

2017-09-25

2017-09-25 11:40:39

0

false

en

2017-09-28

2017-09-28 09:56:39

8

19697a34011b

3.233962

6

3

0

At searchhub.io query cleansing of human input (user query) is the first strategy we apply to each and every search query we receive. In…

5

A closer look into the spell correction problem — Part 3 — the bells and whistles At searchhub.io query cleansing of human input (user query) is the first strategy we apply to each and every search query we receive. In Part 1 and 2 we already discussed a little about the challenges of spell correction at scale and independency of language. However at search|hub we strive to help software systems to understand humans. Therefore we not only have to cater for typos like “scatebord -> skateboard”. There are a lot more reasons why a search engine might not understand or even worse misunderstand a user query. https://www.youtube.com/watch?v=cbtf1oyNg-8 1.Word Segmentation & Word Decomposition: Since most search engines are still based on a token representation of words we first have to identify the words in a user query. This might seem easy and obvious but in quite some cases it’s not. And I’m not only talking about cases that occur in some languages which are traditionally written without inter-word spaces, like Chinese and Japanese or cases where user queries are produced by some sort of speech recognition system. Let’s take a real-world example: “Damenmotorradlederhandschuh”. Now you might think WTF is this. This is a german compound word which is built by gluing together the following words translated into english. “women+bicycle+leather+gloves” The dictionary approach: The traditional approach tackling such a query would be to use a decomposition dictionary that scans through the query and tries to break the query as soon as it finds a sub-word from the dictionary. So let’s do that “Damenmotorradlederhandschuh -> Damen motor rad leder hand schuh” Again for the ones that do not speak german: “women+engine+wheel+leather+hand+shoes” Oh, wait what the heck happened here: by splitting the words we changed their meaning! Imagine the search result for such a query. And what happens if I misspell the query? “Damenmotoradlederhantschuh -> Damen motorad leder hantschuh” So in this query the user made two simple errors and not even the mighty google is able to guess what the user was looking for. Word Segmentation & Word Decomposition are vital parts of the query understanding process and you can’t fix this part at scale by manual mapping through dictionaries and ambiguity handling. 2.Primary Word Detection: Once you segmented / decomposed the query into words you’ll soon realize that now there is another query dimension you have to take care of. The sequence of words. There are several cases where the order or sequence of words inside a user query might change the meaning of the query or at least change the stemming approach. Let’s directly jump in another example for this: Imagine the above user query “Damenmotorradlederhandschuhe” and a couple of other queries which represent the same intent/meaning -> “motorrad leder handschuhe damen, leder motorradhandschuh für damen”. In this example, the order or sequence of words is pretty much independent of its meaning or intent. However as soon as you want to introduce stemming you better make sure that you only stem the “primary word(s)” in this case “handschuh(e)”. But not every user query that includes the same words represents the same meaning or intent. The query “Armbanduhr” aka wristwatch vs. “Uhrarmband” aka watch bracelet is a perfect example for this. Both queries segmented or decomposed consist of exactly the same words but describe two different things. To solve this problem we first have to identify those user queries and then find the primary word in order to understand its meaning or intent. 3.Under- & Overstemming: Grammatically correct stemming can be very tedious. Applying traditional stemmers like Porter or Snowball usually leads to a lot of overstemming or understemming — especially with short words which represent the majority of the query corpus. Again let’s take a real-world example: “babybetten -> babybetten” and “vans -> van” and iphone5s -> iphone5. In the first example, the porter stemmer was unable to stem babybetten to its root babybett while in the second example the brand name vans was reduced to van which in this case changes its meaning. But in order to retrieve relevant and meaningful search results the search engine needs to understand the meaning of the query. While singular and plural forms normally represent the same meaning / intent this might not be the case for automatically stemmed words. search|hub does all of this automatically When we build search|hub we solved all of these areas by combining domain knowledge, smart algorithms and machine learning models fueled by user data. We strongly believe that all of this is key to make search engines understand humans. SEARCH IS THE PLACE WHERE THE USER IS TELLING YOU WHAT HE WANTS. IF YOUR SEARCH ENGINE SPEAKS THE SAME LANGUAGE AS YOUR USERS SEARCH BECOMES A CONVERSATION. SEARCH|HUB HAS SPECIFICALLY BEEN DESIGNED TO HELP YOUR EXISTING SEARCH ENGINE TO UNDERSTAND HUMANS AND DRIVE THESE CONVERSATIONS. We are hiring If you’re excited about advancing our search|hub API and strive to enable companies to create meaningful search experiences, join us! We are actively hiring for Data Scientists to work on next-generation API & SEARCH technology. www.searchhub.io proudly built by www.commerce-experts.com

A closer look into the spell correction problem — Part 3 — the bells and whistles

34

a-closer-look-into-the-spell-correction-problem-part-3-the-bells-and-whistles-19697a34011b

2018-04-11

2018-04-11 08:59:18

https://medium.com/s/story/a-closer-look-into-the-spell-correction-problem-part-3-the-bells-and-whistles-19697a34011b

false

857

null

Search

search

Search

2,870

searchhub

search|hub is a search platform independent, AI-powered search query intelligence API — helping search engines understand humans

de55c8251d7d

searchhub.io

37

2

20,181,104

null

0

W tensorflow/core/platform/cpu_feature_guard.cc:95] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations. For training a truncated Inceptionv3 (pics 299x299) siamese-style in keras, batch_size=32, train/validate sizes=40/40, virtualenv version epochs (seconds): 153, 131, 128, 130, 128 (total runtime 12m 31s) compiled version: 114, 86, 86, 85, 87 (total 9m 1s) Macbook Pro 2012. alexnik-mbpro:tensorflow alexnik$ source venv3/bin/activate (venv3) alexnik-mbpro:tensorflow alexnik$ cd .. (venv3) alexnik-mbpro:study alexnik$ python Python 3.6.5 (default, Apr 10 2018, 12:04:22) [GCC 4.2.1 Compatible Apple LLVM 9.0.0 (clang-900.0.38)] on darwin Type "help", "copyright", "credits" or "license" for more information. >>> import tensorflow as tf >>> a = tf.constant(5.0) >>> b = tf.constant(1.0) >>> c = tf.add(a, b) >>> with tf.Session() as session: ... print(str(session.run(c))) ... 6.0 (venv3) alexnik-mbpro:study alexnik$ ls tensorflow/ | grep whl tensorflow-1.6.0-cp27-cp27m-macosx_10_12_x86_64.whl tensorflow-1.6.0-cp36-cp36m-macosx_10_12_x86_64.whl tensorflow-1.8.0rc1-cp27-cp27m-macosx_10_13_intel.whl tensorflow-1.8.0rc1-cp36-cp36m-macosx_10_13_x86_64.whl

23

null

2018-05-12

2018-05-12 08:05:53

2018-05-12

2018-05-12 16:26:27

0

false

en

2018-05-15

2018-05-15 20:25:18

3

196aa4a495a6

1.554717

3

0

If you are like me, implementing some AI-models in your spare time to keep your skills sharp, you probably saw something like this:

4

Custom TensorFlow with all CPU optimizations. If you are like me, implementing some AI-models in your spare time to keep your skills sharp, you probably saw something like this: and have a feeling that your training time can be shorter. And that’s true! There is a stackoverflow thread about it: How to compile Tensorflow with SSE4.2 and AVX instructions? This is the message received from running a script to check if Tensorflow is working: I…stackoverflow.com But, in my mind, it’s impractical to type gazillion symbols line all the time, so I wrote a shell script for MacOs and Linux to compile TF from the source with all CPU optimizations: How cool is cool you may ask. (phobrain)[https://github.com/phobrain] did some tests and his results showed that custom TF is 30% faster for him: Quite awesome I would say! So, you compiled custom TF (don’t forget that you need Python VirtualEnv) and want to test if it’s working. That’s easy, just remember that you cannot run Python from the same folder where you compiled TF. I have 2 virtual envs in the folder with TF sources, compile and store .whl file for Python 2.7 and Python 3, so I can reuse it later for different projects: Another cool thing about compiling from the sources is that you can have all the new cool features right away, like eager mode, and you don’t need to wait until Google will publish .whl which won’t support all the nice CPU flags of your laptop.

Custom TensorFlow with all CPU optimizations.

3

custom-tensorflow-with-all-cpu-optimizations-196aa4a495a6

2018-05-15

2018-05-15 20:25:18

https://medium.com/s/story/custom-tensorflow-with-all-cpu-optimizations-196aa4a495a6

false

412

null

Mac Os X

mac-os-x

Mac Os X

344

Alex Nikiforov

null

2a31d3dac09f

nikiforov.al

3

20,181,104

null

0

null

0

null

2017-11-19

2017-11-19 18:15:17

2017-11-21

2017-11-21 01:34:32

5

true

en

2017-11-21

2017-11-21 17:50:48

0

196b73482106

5.384277

0

Democratization Of History In An Artificial Intelligence (AI) Enabled World

5

A Spring of Truth’s Perfection Democratization Of History In An Artificial Intelligence (AI) Enabled World unsplash.com To know nothing of what happened before you were born, is to forever remain a child — Cicero History is written by the victors, the strongest, the most determined.Truth is found most often in silence, in the quiet places — Kate Mosse History of History Pre-history is the era before recorded history. History truly begins when writing and tools to record history emerged. Prior to that, it was mostly passed on from generation to generation in the form of stories and narratives. In between prehistory (period of human activity between the use of the first stone tools c. 3.3 million years ago and the invention of writing systems, the earliest of which appeared c. 5,300 years ago) and history is an era called proto-history in which involved early Jiahu (Chinese) symbols on turtle shells Wikipedia Then came the ancient Sumerian civilization which introduced the Cuneiform script. The Bronze Age also saw many civilizations such as Egyptian and Indian civilization develop hieroglyphs and Indus scripts to document and record history. Egyptian Hieroglyphics (Wikipedia) Most scripts involved painting, etching or engraving messages on stone or metal to the ensure durability. The first modern writing surface is the papyrus which was used in Egypt. Indus Valley Script, India (Harappa.com) The Incomplete Argument I hate to take sides in most arguments about world affairs. Not because I am afraid but because I am afraid that others will be victims of my half-truths. Truths that are based on the winners history of the world. I truly believe that the more I know, the more I understand that I know very little. Therefore, to argue from a point of ignorance is to spread misinformation just like some of the winners in history did to the detriment of many others. Let me give you an example. It was a cold, winter night in Mumbai, India and I was sitting with my friends and enjoying food at a local restaurant. It was November 9, 2016 — the day India began one of most grandest and most controversial economic experiment known as “demonetization”. Suddenly, my friend receives a text from his father and he gets up in the middle of dinner and informs me that he has to rush home. I thought it was a family emergency. However, turns out that he wanted to deposit all his Rs. 500 and Rs. 1000 bank notes as they will cease to be legal tender tomorrow. I did not make much of it. Personally, very few of my transactions are in cash. The point is I still use cash in India. As I finished dinner and was on my way home, I could see people queuing up at gas stations not to fill gas but to exchange their 500 and 1,000 Rupee notes. What an incredible sight. A nation shocked into surrendering its cash and galvanized to action much like the struggle for independence. Only this time, the reasons were to halt corruption, the onslaught of Black money (money outside the banking system) and to spur digitization of money. When I rushed home that night, I realized I had some money that were in the denominations to be demonetized. Early next morning, I woke up to even longer queues at banks to deposit my cash. It was a sight to behold. When I read of bank runs (a situation where all depositors rush to withdraw cash because they think banks will collapse) in Greece, it did not register in my mind as such a powerful event for the local people. That day, I understood perfectly the power of money to smoothen and uproot everyday life. It took me three days and experiences filled with expletives delivered by the very common men that I was used to exchange pleasantries with to get my measly cash balance back into the banking system. Demonetization had created demons out of ordinary people. Quite possibly, turned people’s livelihood into a continuous nightmare. Having said that, I have also heard many arguments in support of demonetization. To my mind, every major reform is death by a thousand cuts. It’s never swift, the results can span two to three political regimes and gives an opportunity to the media to make hay while the sun shines. Again, everybody is just doing their job but it does create a perfect storm. Fast forward to today and I still feel that demonetization was a good move but every time I hear arguments against demonetization or the way it was executed, the strength of my conviction dwindles until I become completely confused. You don’t have to go to India to experience what I experience— the stasis created by news (call it true, fake, biased, half truths or anything else) amplified by incentivized propaganda on social media. You just don’t know what to believe in until you actually experience it first hand. Even then, it takes multiple perspectives to create a complete picture of the truth. Ironically, social media and the internet is an axe that swings both ways. These tools democratize truth in that the truth is no longer the domain of the powerful. Grassroots democracy can grow. However, if used for the wrong ends, these tools become vehicles of mass information asymmetry skewed towards the strongest in society. It’s just not demonetization but countless other topics where I feel my incomplete knowledge of the topic makes me a lifelong learner instead of an expert on that topic. So, I try to provide my inputs but also caution that I have my blind sides. Many may say that is taking the easy way out. However, when it is decision making time, I do make my decision to the best of my knowledge. Although, I still feel it’s my duty to be honest and acknowledge my unknowns. The Relatively Immutable Truth A Blockchain is a single version of the truth or the immutable truth. What if cryptography could enable verifiable news through a consensus mechanism? The truth will not be a domain of vested interests and the recording of history can be decentralized. The best use of data, to my mind, is unraveling true facts and news. A truth that is relatively unbiased and crowdsourced. Parting Thoughts One of the most profound human endeavors that can be undertaken, in an age where Artificial Intelligence has taken over most of the mundane human jobs, is to capture the truth one person at a time. While history books get rewritten all the time, we have the power to democratize history and document it in real time. Of course, the best way is through creative pursuits i.e. music, art or even writing down your version of the truth. When Egyptians used hieroglyphics to document their version of history, modern humans don’t have all the keys to unlock their mysteries yet. However, if we document our individual version of history in a way that it becomes timeless and easy to interpret by the future generations, our creative pursuits will have a greater meaning: to capture and document universal truths that are not the domain of a few but the power of the many. An endeavor that Artificial Intelligence can aid and abet but not lead.

A Spring of Truth’s Perfection

0

the-spring-of-truths-perfection-196b73482106

2017-11-21

2017-11-21 17:50:50

https://medium.com/s/story/the-spring-of-truths-perfection-196b73482106

false

1,206

null

History

history

History

47,082

Abhishek Kothari

Writer @ The Intersection of Finance, Tech & Humanity. Stories of a Global Language: “Money”. Contributor @ Startup Grind, HackerNoon, HBR. Twitter@akothari_mba

f42b1990edb5

abhishekkothari

5,955

1,129

20,181,104

null

0

null

0

1b48ddc59bcd

2018-09-25

2018-09-25 16:34:42

2018-09-25

2018-09-25 18:00:01

0

false

en

2018-09-25

2018-09-25 18:07:29

19

196bdb661fa2

2.264151

1

0

This is Day 1 of the #100DaysOfDataScience Challenge. When I decided to take on this challenge headfirst, I came across several amazing…

4

Day 1 of 100 #100DaysOfDataScience This is Day 1 of the #100DaysOfDataScience Challenge. When I decided to take on this challenge headfirst, I came across several amazing resources online. I signed-up for the Data Science course in Analytics Vidhya. I also came across a Open Machine Learning Course(https://mlcourse.ai/) through Hackernews, it had limited seats I guess, and fortunately my request was approved! yay!(When the Student is ready, a Teacher appears). I spent about 3 hours today for the challenge, which also included setting up the development environment. I will briefly describe about what I learnt and consumed today. Strategies in Data Science: Top Down Strategy Bottom-up Startegy I found this blog explain about the above mentioned strategies, do have a look — https://blog.dataiku.com/top-down-vs.-bottom-up-approaches-to-data-science Before diving into Data Science, some prerequisite knowledge is required to better understand what we learn. Data Science is heavily dependent on Math and Statistics, so its best to make our self proficient in these topics, to clearly make sense of what is happening under the hood. We will also need to know on how to work with tools like Python and R. R and Python are popular and powerful languages for Data Science, New libraries and tools are continuously added to their catalog. R is mainly used for statistical analysis while Python provides a more general approach to data science. Even if someone is not comfortable with the above mentioned topics, there is no need to worry. Given the amount of resources currently available in the internet, we can easily pick up knowledge on these topics. KhanAcademy is really good, I like the YouTube page 3Blue1Brown. Here are few links to pick-up the basics, If you already have working knowledge in Python, feel free to skim through the videos to refresh the concepts. Python — Basics Introduction to Python Python Variable and Inputs Lists and Tuples Dictionaries Conditional Statements Loops File reading and Writing Statistics Basics: What is Statistics? Statistics are Facts and Figures.Statistics rely upon calculation of numbers, it relies upon how the numbers are chosen and how statistics are interpreted. Statistics is techniques and procedures for analyzing, interpreting, displaying and making decisions based on data. Here is a really good Quora answer on, ‘How do DataScientists use Statistics?’ Two major divisions of Statistics Descriptive Statistics Inferential Statistics Classic descriptive statistics include mean, min, max, standard deviation, median, skew, kurtosis. Inferential statistics are a function of the sample data that assists you to draw an inference regarding an hypothesis about a population parameter. (I know its complex, will break it down in upcoming posts). Descriptive Statistics: https://trainings.analyticsvidhya.com/courses/course-v1:AnalyticsVidhya+Python-Final-Jan-Feb+Python-Session-1/courseware/fccdc9ca4689406a9a2a1b75e984e551/a4124effe90b4b77b93d942992c94ccc/?child=first Descriptive Statistics Basics: Mean (Average of all data point) Median (Middle data point) Mode (Most common data point) Variance Standard Deviation Range ( Difference between smallest and largest data points) Video Links: Mean, Median Mode: https://www.youtube.com/watch?v=GrynkZB3E7M Range, Variance, Standard Deviation: https://www.youtube.com/watch?v=E4HAYd0QnRc Measures of Variablity: https://www.youtube.com/watch?v=fvgDqVda9L8 Basic Types of Distributions: Unimodel distribution Bimodel distribution — — — Check out my learning notes at: laxmena — Day 1 Notes PS: I haven’t put much effort in writing this article, as I had very little time, will improve future articles. Lakshmanan Meiyappan | LinkedIn View Lakshmanan Meiyappan's profile on LinkedIn, the world's largest professional community. Lakshmanan has 3 jobs…www.linkedin.com Lakshmanan Meiyappan (@laxmena) | Twitter The latest Tweets from Lakshmanan Meiyappan (@laxmena): "I just published "Day 0 of 100 #100DaysOfDataScience"…twitter.com Lakshmanan Meiyappan - Medium A couple of years ago, I came across the term 'Data Science' for the first time. Since then, the the word kept…medium.com

Day 1 of 100 #100DaysOfDataScience

50

day-1-of-100-100daysofdatascience-196bdb661fa2

2018-09-25

2018-09-25 18:07:29

https://medium.com/s/story/day-1-of-100-100daysofdatascience-196bdb661fa2

false

600

100 Days of Data Science is a challenge for beginners and enthusiasts to learn and get started in Data Science. I will pen down my experience of my Learning curve and share knowledge what I have gained in this publication.

null

lakshmanan.meiyappan

null

100 Days of DataScience

lakshmanan.meiyappan@gmail.com

100-days-of-datascience

DATA SCIENCE,CHALLENGE,DATA ANALYSIS,PROGRAMMING LANGUAGES,LEARNING

laxmena

Data Science

data-science

Data Science

33,617

Laxmena

Lakshmanan Meiyappan

91dc7710e42

laxmena

89

59

20,181,104

null