Data Science Weekly - Issue 184
Issue #184 June 1 2017
Editor Picks
A Year of Google and Apple Maps
Google Maps’s quiet transformation, Apple Maps‘s ______?...
AlphaGo, in context
I had a chance to talk to several people about the recent AlphaGo matches with Ke Jie and others. In particular, most of the coverage was a mix of popular science + PR so the most common questions I’ve seen were along the lines of “to what extent is AlphaGo a breakthrough?”, “How do researchers in AI see its victories?” and “what implications do the wins have?”. I thought I might as well serialize some of my thoughts into a post...
Experts Predict When Artificial Intelligence Will Exceed Human Performance
Artificial intelligence is changing the world and doing it at breakneck speed. The promise is that intelligent machines will be able to do every task better and more cheaply than humans. Rightly or wrongly, one industry after another is falling under its spell, even though few have benefited significantly so far. And that raises an interesting question: when will artificial intelligence exceed human performance? More specifically, when will a machine do your job better than you?...
A Message from this week's Sponsor:
JupyterCon is August 22-25 in NYC
Discover how the world’s most data-driven organizations are using Jupyter to analyze their data and share their insights. Co-Chaired by Ferenando Perez (creator of IPython) and Andrew Odewahn (CTO of O'Reilly Media), this official Jupyter Conference will explore the breadth and depth of of the Jupyter Platform. Get Early Price by registering before June 30. Learn more.
Data Science Articles & Videos
What does it mean to ask for an “explainable” algorithm?
One of the standard critiques of using algorithms for decision-making about people, and especially for consequential decisions about access to housing, credit, education, and so on, is that the algorithms don’t provide an “explanation” for their results or the results aren’t “interpretable.” This is a serious issue, but discussions of it are often frustrating. The reason, I think, is that different people mean different things when they ask for an explanation of an algorithm’s results...
This app uses artificial intelligence to turn design mockups into source code
While traditionally it has been the task of front-end developers to transform the work of designers from raw graphical user interface mockups to actual source code, this trend might soon be a thing of the past – courtesy of artificial intelligence...
Everything that Works Works Because it's Bayesian: Why Deep Nets Generalize?
For too long we Bayesians have, quite arrogantly, dismissed deep neural networks as unprincipled, dumb black boxes that lack elegance. We said that highly over-parametrised models fitted via maximum likelihood can't possibly work, they will overfit, won't generalise, etc...
A Brief History of Federal Tax Rates
Last month in Vox, Alvin Chang published a chart of 100 years of federal tax brackets. I’ve recreated the graphic below, substituting a log scale for the y-axis. It readily conveys the Reagan-era simplification of tax brackets, as well as the disappearance of tax brackets for the ultra-rich. (In 1936, the highest tax bracket applied to those making more than $83M in 2013-equivalent dollars!)...
Applying deep learning to real-world problems
I thought it would be helpful for other people who plan to use deep learning in their business to understand some of these tweaks and tricks. In this blog post I want to share three key learnings, which helped us at Merantix when applying deep learning to real-world problems...
How AI Can Keep Accelerating After Moore’s Law
New ideas in chip design look likely to keep software getting smarter...
Bayesian GAN
Generative adversarial networks (GANs) can implicitly learn rich distributions over images, audio, and data which are hard to model with an explicit likelihood. We present a practical Bayesian formulation for unsupervised and semi-supervised learning with GANs. Within this framework, we use stochastic gradient Hamiltonian Monte Carlo to marginalize the weights of the generator and discriminator networks. The resulting approach is straightforward and obtains good performance without any standard interventions such as feature matching, or mini-batch discrimination...
The $1700 great Deep Learning box: Assembly, setup and benchmarks
Building a desktop after a decade of MacBook Airs and cloud servers...
Jobs
Netflix - Los Gatos & Los Angeles, CA We are looking to fill several key roles across our Data Science groups.
Director, Production Science & Algorithms In this role, you will lead a high-impact data science team focused on the digital supply chain at Netflix. The problems this team will work on have a direct impact on the viewing experience of our global member base, including ensuring that the digital assets (video, audio, and subtitle/text files) are of high quality, and developing new algorithms and metrics to improve the perceptual quality of our encoded assets.
Manager, Content Programming Science & Algorithms The ideal candidate for Manager of Content Programming Science & Algorithms is an experienced and entrepreneurial-minded data scientist. This is high-impact and challenging role, and will require both strong leadership and technical prowess.
Senior Data Scientist, Content Science & Algorithms We are looking for an experienced individual who is passionate about data science and enjoys working in a collaborative environment. Members of the Content Science team typically work on one or two projects (e.g. predicting movie viewership) over any six month period.
Training & Resources
Exploring LSTMs
The first time I learned about LSTMs, my eyes glazed over. Not in a good, jelly donut kind of way. It turns out LSTMs are a fairly simple extension to neural networks, and they're behind a lot of the amazing achievements deep learning has made in the past few years. So I'll try to present them as intuitively as possible – in such a way that you could have discovered them yourself...
Conversion rates — you are (most likely) computing them wrong
How hard can it be to compute conversion rate? Take the total number of users that converted and divide them with the total number of users. Done. Except… it’s a lot more complicated when you have any sort of significant time lag...
[Stanford] Lecture Collection | Natural Language Processing with Deep Learning (Winter 2017)
Natural language processing (NLP) deals with the key artificial intelligence technology of understanding complex human language communication. This lecture series provides a thorough introduction to the cutting-edge research in deep learning applied to NLP, an approach that has recently obtained very high performance across many different NLP tasks including question answering and machine translation...
Books
The Joy of x: A Guided Tour of Math, from One to Infinity "Delightful . . . easily digestible chapters include plenty of helpful examples and illustrations. You'll never forget the Pythagorean theorem again!"...
For a detailed list of books covering Data Science, Machine Learning, AI and associated programming languages check out our resources page.
P.S. Looking to hire a Data Scientist? Find an awesome one among our readers! Email us for details on how to post your job :) - All the best, Hannah & Sebastian