Data Science & Machine Learning Newsletter #103

Posted on Fri 19 January 2018 in Data Science & Machine Learning Newsletter


Do you want to get updates? Please join Data Science & Machine Learning Newsletter Linked Group


  • Multiple imputation utilising denoising autoencoder for approximate Bayesian inference
    • "Missing data is a widespread problem in machine learning. Bayesian inference is a robust solution to imputing missing values, particularly if multiple imputations are used to model the uncertainty regarding said values."
  • Alibaba's AI Outguns Humans in Reading Test
    • Alibaba has developed an artificial intelligence model that scored better than humans in a Stanford University reading and comprehension test.
  • Robo-Advisers Are Coming to Consulting and Corporate Strategy
    • "Does a robot manage your money? For many of us, the answer is yes. Online and algorithmic investment and financial advice is easy to come by these days, usually under the moniker of “robo-advisor.” Startups such as Wealthfront, Personal Capital, and Betterment launched robo-advisors as industry disruptors, and incumbents, such as Schwab’s (Intelligent Advisor), Vanguard (Personal Advisor Services), Morgan Stanley and BlackRock have joined the fray with their own hybrid machine/advisor solutions."
  • One model to learn them all
    • "Suppose you ask me if I’d like anything to eat. I can say the word ‘banana’ (such that you hear it spoken), send you a text message whereby you see (and read) the word ‘banana,’ show you a picture of a banana, and so on. All of these different modalities (the sound waves, the written word, the visual image) tie back to the same concept – they are different ways of ‘inputting’ the banana concept.... It’s as if we had one concept for the written word ‘banana’, another concept for pictures of bananas, and another concept for the spoken word ‘banana’ – but these weren’t linked in any way. The central question in today’s paper choice is this: Can we create a unified deep learning model to solve tasks across multiple domains?"
  • Uber AI Labs Open Sources Pyro, a Deep Probabilistic Programming Language
    • " Pyro is a tool for deep probabilistic modeling, unifying the best of modern deep learning and Bayesian modeling. The goal of Pyro is to accelerate research and applications of these techniques, and to make them more accessible to the broader AI community."
  • AutoML on AWS
    • "In this article, we present an AWS based framework which allows non technical people to build predictive pipelines in a matter of hours while achieving results that rival solutions handcrafted by data scientists."
  • https://github.com/slundberg/shap
    • Explain the output of any machine learning model using expectations and Shapley values.
  • Deep Learning: AlphaGo Zero Explained In One Picture
    • Alphago Zero Cheat Sheet