Companies such as Google and Facebook wield enormous power with their data. Along with enabling valuable services, the data these companies control can stifle competition, which raises antitrust concerns. But typical antitrust regulations don't really apply to data. This article in the Economist explores new ideas for regulation.
Ever wonder what happens to all that video data being collected by the body cams worn by police? It's not just to document what happened. Taser intends to help police anticipate what might happen. Imagine that you can find out if someone has a criminal record instantly - or be notified if someone’s demeanor has changed and may now be a threat. Think this won't happen?...
A variety of AIs are already being used to set bail, determine sentences, and even contribute to determinations about guilt or innocence. The scary part is that many of these AIs are proprietary black boxes with creator biases and mistakes baked in. The courts are trying to catch up with the issues but the tech is moving fast.
Spark Capital and The AI Conference are hosting a startup showcase on June 2nd. 5 teams will pitch followed by Q&A with the moderator, John Melas-Kyriazi, Senior Associate, Spark Capital. Mention "DataElixir" & save $50. Submission deadline is 5/12/17.
Great introduction to a family of algorithms that arrange geometric data for efficient search. These types of algorithms, called spatial indices, enable lightning-fast queries for things like "find all buildings in this area" or "find 1000 closest gas stations to this point." This post by Vladimir Agafonkin explores how they work.
If you've been interested in learning about deep learning but not sure where to start, start here! This is a very well organized learning path that's easy to follow with lots of links to curated resources.
This well-researched roundup includes useful descriptions, links and summarized ratings for dozens of the most popular machine learning courses that you can take online. Includes a variety of niches such as market trading, genomics, and recommenders.
Anscome's Quartet is a well-known set of datasets that have nearly identical summary statistics but look very different when plotted. It's an awesome demonstration of why you should always look at your data. This post shows how to create your own version of Anscome's Quartet. This is very well done and got a LOT of attention around the web this week.