— In the News —
Google has a massive impact on the tools, applications and research that help steer the data science community. As with prior years, this retrospective by Jeff Dean is amazing in its scope. Includes useful summaries, screenshots, videos and linked references throughout.
— Insight —
"... as we look forward to the next decade a few things seem certain: increasing automation, increasing system complexity, faster processing, more inter-connectivity, and an even greater human and societal dependence on technology. What could possibly go wrong?"
— Tools and Techniques —
Great introduction to overfitting by Alex Hayes that offers insights beyond what's typically covered in introductory machine learning resources. Covers both prediction and inference problems, provides supervised and unsupervised examples of overfitting, and presents a fundamental relationship between train and test error.
This collection of considerations for new data projects is based on Jeremy Howard's decades of experience with projects across a wide variety of industries. This is a post to come back to again and again.
In a 3-part series, Chris Said presents a practical way of determining optimal sample sizes that abandons the notion of statistical significance. This first part is a non-technical overview that ends with a section called, "Three lessons for practitioners." Follow the links for more.
Clever use of GPT-2. Give it a corpus of chess games, represented as text-based moves, and it learns to play! Includes a link to a Colab Notebook where you can try it yourself.
No BiaS is a new podcast about the emerging and ever-shifting terrain of artificial intelligence & machine learning. Didn't win the lottery to get into the hottest ML conference of the year? No sweat! Get the full rundown of NeurIPS 2019 in the latest episode of No BiAS 🎧 with Cheryl Martin & Brent Schneeman.
— Resources —
Awesome collection of resources for learning machine learning. This is very well curated and organized. Includes online books, interactives, courses, key articles and other resources.
— Career —
Consulting work can be lucrative but even in data science, it's not a sure path. This is an insightful post by Ethan Rosenthal about the challenges and opportunities of data science consulting work.