ML models often exhibit unexpectedly poor behavior when they are deployed in real-world domains. We identify underspecification as a key reason for these failures. An ML pipeline is underspecified when it can return many predictors with equivalently stron... (more…)

We are always trying to stay up to date with the latest research and publications around the world. This includes browsing the raw ArXiv listing (or the saner Arxiv Sanity Preserver), staying up to date with our nerd Twitter feeds and plenty of other sour... (more…)

Kids taking too long to shower? Build a machine-learning mobile web application that detects how long the shower is turned on by listening for audio cues. (more…)

Niland's approach to Machine Learning Research. Building A Successful Team by Christophe (CTO and founder)...

Posted by Alan Ho, Product Lead and Masoud Mohseni, Technical Lead, Google Research “Nature isn’t classical, damnit, so if you want to m... (more…)