— In the News —
Currently there are more than 8000 startups and companies listed in Crunchbase that are relying on machine learning for their products and services! Here are 25 that are worth paying attention to. Each entry includes a short description of what they're doing, challenges, opportunities, and funding.
— Sponsored Link —
In this webinar, you will learn practical approaches to maintaining privacy and the vulnerabilities of each approach within the analytics workflow. You’ll walk away with a framework for identifying privacy risks in your own analyses, multiple approaches to preserve privacy, and how to make decisions that balance utility and privacy.
— Tools and Techniques —
If you're doing data science work, you're doing research and in that case, you need research quality data to work with. In this post, Jeff Leek explores what that means exactly, including specific suggestions for getting your data to "research grade."
Great introduction to DataOps by Brad Ito, a Co-Founder and CTO at Retina. This post explores how DataOps is useful and what to consider for implementing it in your own organization. This is a great read that includes a lot of practical suggestions for things like team organization, process, tools, measuring value, etc.
GAMs offer offer a middle ground between simple linear models and complex machine-learning techniques, allowing you to model and understand complex systems. This short, interactive course will teach you how to use these flexible, powerful tools to model data and solve data science problems.
The Sleeping Beauty problem is a thought experiment in decision theory. This post by Matthias Plaue offers a variety of ways to look at it and will definitely get you thinking. Are you a "thirder" or a "halfer?"
— Resources —
Foundations of Machine Learning is a general introduction to machine learning that can serve as a textbook for graduate students and as a reference for researchers. The newly released 2nd edition includes new chapters, appendices, and exercises. Free to download or follow the links for print.
Shervine Amidi is a TA at Stanford who has a knack for creating useful cheatsheets for his students. This set covers the latest Artificial Intelligence course (CS 221). These are well organized and include lots of diagrams. He's also created a popular set for CS 229, Machine Learning >>
— Data Viz —
Nice collection of key visualization books with short descriptions and reviews for each. Includes classics as well as some titles that have recently been released. The descriptions here are useful and don't shy away from constructive criticism along the way. Organized into the following sections: "Jump-starts," "Foundations," "Specialty" and "Not Recommended."
It's no secret that rainbow color maps are considered to be ineffective by visualization experts. Part of the problem is that rainbow color maps tend to discretize continuous data and, as this new EuroVis paper shows, that can happen in unpredictable ways.