Data Science Weekly - Issue 71
Issue #71 April 2 2015
Editor Picks
Breaking Linear Classifiers on ImageNet
You've probably heard that Convolutional Networks work very well in practice and across a wide range of visual recognition problems. Yet, a second group of seemingly baffling results has emerged that brings up an apparent contradiction. I'm referring to several people who have noticed that it is possible to take an image that a state-of-the-art Convolutional Network thinks is one class (e.g. "panda"), and it is possible to change it almost imperceptibly to the human eye in such a way that the Convolutional Network suddenly classifies the image as any other class of choice (e.g. "gibbon"). We say that we break, or fool ConvNets...
How to Consistently Hire Remarkable Data Scientists
This article is by Jeremy Stanley, Chief Data Scientist and EVP Engineering at Sailthru, where he’s responsible for building intelligence into the company’s marketing personalization platform...
Learning to See Data
For the past year or so genetic scientists at the Albert Einstein College of Medicine in New York have been collaborating with a specialist from another universe: Daniel Kohn, a Brooklyn-based painter and conceptual artist. Mr. Kohn has no training in computers or genetics, and he’s not there to conduct art therapy classes. His role is to help the scientists with a signature 21st-century problem: Big Data overload....
Data Science Articles & Videos
Peter Norvig: How Computers Learn
Vienna Gödel Lecture 2015 with Peter Norvig, Research Director at Google Inc. (talk starts at 10:56)...
Bill Moreau, USOC on Empowering World’s Best Athletes through Analytics.
We discuss how United States Olympic Committee uses Big Data, how athletes respond to Analytical insights, integration of sports medicine into sports performance and sports injury...
Capital One Labs - Data Science at a Bank: Randy Carnevale Interview
We recently caught up with Randy Carnevale, Director of Data Science at Capital One Labs. We were keen to learn more about his background, his move to data science from medical informatics, his choice of going to a financial firm to do data science, what he thinks of data science education, and why he has chosen to work with the Metis Data Science Bootcamp to find up-and-coming data scientists......
The Science of Crawl (Part 3): Priorization
In this post, we look at the challenge of prioritizing which web documents to capture first. To ensure our search engine contains relevant results for our publishers, we need a crawler which continuously discovers new content...
Are people watching Facebook videos?
I was interested in whether or not people actually watch videos that we (news people) post on Facebook. I had a feeling that videos get a lot of views because they are auto-played...
How To Be Data Scientists That Works In Field X
You want to combine your experience in your current professions (called field X) for this article, with your rapidly increasing knowledge of data science. However, you're not sure if data science is more than a mixture of statistics and programming...
Facebook's demo of Memory Networks
Some of the leading minds in AI research are working at Facebook to build intelligent machines. One of the group's more recent advances is a technology called Memory Networks, which enables a machine to perform relatively sophisticated question answering, as in this example of a machine answering questions about a Lord of the Rings synopsis....
Data Mining Problems in Retail
Retail is one of the most important business domains for data science and data mining applications because of its prolific data and numerous optimization problems such as optimal prices, discounts, recommendations, and stock levels that can be solved using data analysis methods...
Tetris AI Environment - Build bots to play against this Tetris Sandbox
A final project for Advanced Machine Learning. Build bots to play against this Tetris Sandbox!...
Jobs
Data Scientist - Craft Coffee - Brooklyn, NY This is a rare opportunity to dive deep into an untouched domain. There's no roadmap for what we're doing in coffee. Until now, the coffee industry has relied on subjective sensory methods, derivative thinking and bullshit marketing. But we know that there are objective truths behind coffee preferences, and we're building the first quantitative model to understand these truths. Our data scientist will work closely with our coffee director and head roaster to figure out what data points really matter (and which ones don't)...
Training & Resources
The Grammar of Data Science: Python vs R
In this post, I will elaborate on my experience switching teams by comparing and contrasting R and Python solutions to some simple data exploration exercises...
scikit-learn 0.16 is out
Faster DBSCAN, LSHForest, out of core PCA & Birch clustering...
Deep Learning: MIT Press Book in Preparation
Yoshua Bengio Deep Learning Book has a new draft...
Books
The Drunkard's Walk: How Randomness Rules Our Lives Excellent book on randomness in day to day lives...
"This smart book will make you think. Academic yet easy to read, it explores how random events shape the world and how human intuition fights that fact...."
For a detailed list of books covering Data Science, Machine Learning, AI and associated programming languages check out our resources page.
P.S. Enjoyed the newsletter? Please forward it along to friends and colleagues - we'd love to have them onboard! - All the best, Hannah & Sebastian