#top links

[todd--chavez]

Mr. Peanutbutter Pro-Fracking, Hasn’t Told Wife.

Beloved TV game show host and California gubernatorial candidate Mr. Peanutbutter recently signed a letter in favor of fracking. His wife, writer Diane Nguyen, doesn’t seem to be in the loop. I mean, if she knew, we’d probably all hear about it. Right?

Also the Top Lynx. Pun on top links.

[bananadata]

Favorite data & AI reads of the week

- Long read Strategic Competition in an Era of Artificial Intelligence

- Optimised world Evolving Floorplans

- Teenage Quantum Computer Major Quantum Computing Advance Made Obsolete by Teenager

- Dataviz of the week: Why Some Accents Don't Work on Google on Alexa or GoogleHome

- Beatles AI Lennon or McCartney? Can statistical analysis solve an authorship puzzle?

- Video of the week: How Robot Hands Are Evolving to Do What Ours Can

If you're looking from some stuff to keep you busy at the office this week, try these:

- Why Americans Spend So Much on Health Care—In 12 Charts by the Washington Post "Americans aren’t buying more health care overall than other countries. But what they are buying is increasingly expensive." - Want Less-Biased Decisions? Use Algorithms. by the Harvard Business Review "So how worried should we be about the modern ascendance of algorithms?" - Data's day of reckoning by O'Reilly "What does it mean to take responsibility for building, maintaining, and managing data, technologies, and services?" - DRAW THIS. by Dan Macnish "Draw This is an instant camera that draws cartoons." - How Goodreads offloads Amazon DynamoDB tables to Amazon S3 and queries them using Amazon Athena by AWS blog "At Goodreads, we’re currently in the process of decomposing our monolithic Rails application into microservices." - [Github] T2F: text to face generation using Deep Learning - [Paper] Blood Bowl: The Next Board Game Challenge for AI

[bananadata]

Top data links this week - Google AI & facial recognition

Google announced last week that it would renounce Project Maven, their joint project with the US military. The company made a billion dollar cut in their turnover to respond to a widespread petition within the company not to use Google Tech to make weapons more accurate. This is an impressive victory for Google employees. The article by Jacobin Magazine tells the story of how the employees organised to oppose the project; and won! As part of their response, Google have also released their AI principles - with what is perhaps their new moto "Be socially beneficial", a nice reminder of the original "do no evil."

Our top data links this week 👇

- Can't be done What tech calls "AI" isn’t really AI

- Political AI How The New York Times Uses Software To Recognize Members of Congress

- Getting the right food Food Discovery with Uber Eats: Building a Query Understanding Engine

- Awesome Dataviz of the week: Working Remotely and Where the Time Goes

- Finally in the USA US government to use facial recognition technology at Mexico border crossing

- Data project of the week: Where Killings Go Unsolved by the Washington Post

Here are some non-GDPR related articles for you to read. - How can Santa keep his lists when GDPR is around? by Worldbuilding "I haven't been notified by Santa and/or his elves that he is collecting data about me. And mind you: my name and surname are my personal data, not to mention data on whether I have been good or naughty." - Microsoft sinks data centre off Orkney by BBC News "The data centre, a white cylinder containing computers, could sit on the sea floor for up to five years." - Machine Learning: how to go from Zero to Hero by FreeCodeCamp "If your understanding of A.I. and Machine Learning is a big question mark, then this is the blog post for you. Here, I gradually increase your Awesomenessicity™ by gluing inspirational videos together with friendly text." - How Policymakers Can Foster Algorithmic Accountability by Center for Data Innovation "Policymakers should reject these proposals and instead support algorithmic decision-making by promoting policies that ensure its robust development and widespread adoption." - Counterintuitive examples in probability by Mathematics "For a simple random walk, the mean number of visits to point b before returning to the origin is equal to 1 for every b≠0." - Reinforcement Learning from scratch by Insight "Contrary to many classical Deep Learning problems that often focus on perception (does this image contain a stop sign?), Deep RL adds the dimension of actions that influence the environment (what is the goal, and how do I get there?). - Machine learning explained with gifs: style transfer by Eliot Andres "Pioneered in 2015, style transfer is a concept that uses transfers the style of a painting to an existing photography, using neural networks." - Kubernetes best practices: upgrading your clusters with zero downtime by Google Cloud Platform Blog "Today is the final installment in a seven-part video and blog series from Google Developer Advocate Sandeep Dinesh on how to get the most out of your Kubernetes environment."

[bananadata]

Top data links this week: GDPR party, airbnb image recognition and royal wedding AI

In case you hadn't noticed, the GDPR is upon us. Emailers and data processors everywhere are freaking out. Companies like Touch, or Dataiku for that matter, have been prepping for this for months to ensure they don't make it to the GDPR Half of Shame. And most of them are not ready. But the question remains, why is everyone going insane and sending out shovels of privacy policy emails, or even resubscribe emails? Especially since according to experts, most of these emails are actually unnecessary, and some even illegal. Of course, that doesn't mean you shouldn't read them though and take this opportunity to clean up your inbox! Our biggest concern of course: what impact will GDPR (and the so-called right to an explanation) have on machine learning?

Top 6 data links below 👇

- Image Deep Learning Categorizing Listing Photos at Airbnb

- Crazy ML When algorithms surprise us

- NSFW Sex Toys and Artificial Intelligence — The Past, Present, and the Future

- Dataviz of the week: Average age and tenure of Congress and Senate Members

- Royal Wedding Sky uses machine learning to spot celebrities at the Royal Wedding

- Video of the week: Ramen generated using the same neural network framework that generated new celebrity faces

Here are some non-GDPR related articles for you to read. - Imaginary Soundscape "Take a walk in the soundscapes imagined by AI" - Machine learning is helping computers spot arguments online before they happen by The Verge "The software was preprogrammed to look for certain features that past research has shown correlates with a conversational mood. For example, signs that a discussion will go well include gratitude ("Thanks for your help"), greetings ("How’s your day going?"), hedges ("I think that"), and, of course, the liberal use of the word "please."" - The US military is funding an effort to catch deepfakes and other AI trickery by MIT Tech Review "This summer, under a project funded by the Defense Advanced Research Projects Agency (DARPA), the world’s leading digital forensics experts will gather for an AI fakery contest." - Why is machine learning 'hard'? by Stanford.edu "What is unique about machine learning is that it is ‘exponentially’ harder to figure out what is wrong when things don’t work as expected." - 3D Face Reconstruction with Position Map Regression Networks by Heartbeat "Position Map Regression Networks (PRN) is a method to jointly regress dense alignment and 3D face shape in an end-to-end manner." - Human in the loop: Machine learning and AI for the people by ZDNet "HITL combines two common ML variants, supervised and unsupervised learning." - [Video] Microsoft demonstrates its own chat bot that can call users.by Tom Warren - [Paper] Self-Attention Generative Adversarial Networks

[bananadata]

Top links for the week: doctors & AI, computing AI, Trex and subway dataviz

News flash: according to Bloomberg, Data Scientist is the hottest job in America! If this isn't news to you, this is still something you can share with your non data friends, or on Twitter. So it's official, the Math Men have overthrown the Mad Men according to a New Yorker article - a look at how math geeks made it to the top of the business food chain. And if you still want to hear from them, check out this article on what Alex Korchinski learned from his Math degree, not about math, but about life.

Our top links this week 👇

- AI Gold Rush How tech can turn doctors into clerical workers

- Too much data, too little storage AI and Compute

- Robot future What the History of Math Can Teach Us about the Future of AI

- Dataviz of the week How 2 M.T.A. Decisions Pushed the Subway Into Crisis

- Think about it TF-rex: Playing Google's T-rex game with TensorFlow

- Video of the week The Artist Working to Make Artificial Intelligence Less White

🎼 Gimme gimme more 🎶 - Does the brain store information in discrete or analog form? by MIT Tech Review "For engineers, the question of whether to store information in analog or discrete form is easy to answer. Discrete data storage has clear advantages, not least of which is that it is much more robust against degradation." - Growing up with AI: How can families play and learn with their new smart toys and companions? by MIT Media Lab "After reflecting on my own childhood, I can’t help but wonder what this experience will look like for today’s children, who are not just growing up with the web, but are also the first generation of kids to grow up with artificial intelligence (AI) in their daily lives." - The AI Doctor Will See You Now by the Wall Street Journal "As machines learn from at times millions of humans, doctors are gaining the ability to better identify disease and even predict it before it becomes catastrophic." - The Markov Property, Chain, Reward Process and Decision Process by Xavier Geeinck "To illustrate this with an example, think of playing Tic-Tac-Toe. When we are able to take a decision based on the current state, rather than needing to know the whole history, then we say that we satisfy the conditions of the Markov Property." - Simple and comprehensive tutorials in TensorFlow by easy-tensorflow

[bananadata]

This week’s top AI and data science links: OpenAI vs DOTA, traffic data and the future of data storage

OpenAI was center stage this week when their bot beat top human player at the Dota 2 world championship. This is a big deal because multiplayer online battle arena games and real-time strategy games were seen as being way beyond the abilities of an AI today - with their extremely complex sets of rules and possible actions. 

OpenAI released an article further developing how their bot managed to achieve this through self-play alone.

If you're thinking "wow how did AI get to be this good while I wasn't watching?" you're not the only one. And it turns out the achievement, while impressive, may not be the groundbreaking event the press is raging about. In an article published by the NY Times, the team from OpenAi actually goes into some of their failed experiments - and explains why they believe that building AI that follows instructions set by humans may be a way to control super intelligence in the future.

Our very favorite links this week:

- Long read Eliminating the human

- Data storage trends How hardware drives the shape of databases to come

- Rush hour data viz Leaving town at rush hour? Here's how far you're likely to get from America's largest cities.

- Google Research Making visible watermarks more effective

- Here's an AI goodie: Generating expressive 3D facial animations from audio

Never enough! Here are more links that we enjoyed reading this week, just for you. - Why Machine Learning Is A Metaphor For Life by Adit Deshpande - Romance Novels, Generated by Artificial Intelligence by Elle O'Brien - Inside the Increasingly Complex Algorithms That Get Packages to Your Door by MIT Tech Review - Design makes AI smarter by Elaine Lee - Why Elon Musk is Wrong about AI by Daniel Jeffries - How To Survive Corporate Politics as A Data Scientist by SeattleDataGuy - Killings of Blacks by Whites Are Far More Likely to Be Ruled 'Justifiable' by TheUpshot - God or the divine is referenced in every state constitution by Fact Tank - How We Built a Lifetime Eclipse Predictor by Source - Useless Data Comparisons by Flowing Data - Does Your State Care More About Batman or Puppies? by DiscoverGeek - An AI Dreamed Up Street Scenes, and They're Surprisingly Good by MIT TechReview - Understanding overfitting: an inaccurate meme in supervised learning by Memo's Island - Data Science Rosetta Stone: Classification in Python, R, MATLAB, SAS, & Julia by Heaton Research - [Paper] Some Philosophical Problems from the standpoint of Artificial Intelligence - [Paper] How to become a Bayesian in eight easy steps: An annotated reading list - [Paper] A Primer on Neural Network Models for Natural Language Processing - [Paper] Neural Nets Can Learn Function Type Signatures From Binaries

[bananadata]

Our favorite Data Science Links this week: bullshit, safe crime prediction & okcupid data analysis

RESPECT - Calling Bullshit - In the Age of Big Data

Cry out against Deep Learning - An Adversarial Review of "Adversarial Generation of Natural Language"

Protecting Privacy - Safe Crime Prediction

Dating Insights - Undressed: What's the Deal With the Age Gap in Relationships?

Here's a sweet dataviz! - Bavericks & Heretics - Ideas rejected, later proven correct

More data science links for you to read if you have a little time on your hands. - Human-Level AI Is Right Around the Corner—or Hundreds of Years Away by IEEE Spectrum - Secret Algorithms Threaten the Rule of Law by MIT tech review - Three Reasons 'Wonder Woman' Has Already Made History by FiveThirtyEight - So, bots you say… by Alex Ramirez - Learning to Cooperate, Compete, and Communicate by open.ai - How big data can help you pick better wine by Quid - Scientists slash computations for deep learning by phys.org - DAWN: Infrastructure for Usable Machine Learning by Stanford - Tensorflow I Love You, But You're Bringing Me Down by Nate Harada - Foundations for deep learning by Pauli Space - Neural networks Exercises (Part-1) by Rexercises - Google Brain Residency by Tiny Clouds - the levity of serverlessness by tecznotes - A Query Had To Go by activity.club - Google's AI-Building AI Is a Step Toward Self-Improving AI by SingularityHub - Why Apple is struggling to become an artificial-intelligence powerhouse by Washington Post

[bananadata]

What we’re reading in data science this week: saving the world from AI, job recommender, and greek laws

Research reveals that positively shaping the development of artificial intelligence is the most pressing problem for humanity today: its scale is huge, and it's still fairly neglected despite being solvable. Indeed, AI will most likely be able to beat us at everything by 2016, so we better start working now. For companies, this means that they will have to take into account the risks of algorithmic decision, along with supervision and issues raised by scalability. For our day-to-day life, interesting initiatives are on the rise. The Anti AI AI wearable device for example sends a shiver down your spine when you are interacting with an AI. The Algorithmic justice League and Joy Buolamwini are also actively fighting against bias in visual recognition algorithms. You can join the MIT project right now. And discuss whether skin can be to dark for technology.

Also, this: - How AI Can Keep Accelerating After Moore's Law by MIT Technology Review - [Oldie] Statistical Modeling, the two cultures - Applying deep learning to real-world problems by merantix - What does it mean to ask for an "explainable" algorithm? by Freedom to Tinker - AlphaGo, in context by Andrej Karpathy - Everything that Works Works Because it's Bayesian: Why Deep Nets Generalize? by inFERENCe - An Overview of Multi-Task Learning in Deep Neural Networks by Sebastian Ruder - How to Evaluate the Skill of Deep Learning Models by Machine Learning Mastery - Winning the Battle for Riddler Nation; An Agent-Based Modelling Approach to the Solution by Curtis Miller - Selection Bias in Online Experimentation by Airbnb Engineering & Data Science - How Data Science Helps Power Worldwide Delivery of Netflix Content by the Netflix Tech Blog - [Podcast] How Artificial Intelligence Might Change Your World by Vault Analytics - [Resource] Essential Cheat Sheets for Machine Learning and Deep Learning Engineers by Kailash Ahirwar - [Paper] Living Together: Mind and Machine Intelligence - [Github] A TensorFlow Implementation of Tacotron: A Fully End-to-End Text-To-Speech Synthesis Model - [Github] folk music modelling with LSTM