Andre's photo
   

research scientist at DeepMind

research interests

artificial intelligence, machine learning, reinforcement learning

postal address   

6 Pancras Square
London N1C 4AG, UK

email andrebarreto @ google.com
I received my PhD in Computational Systems from Universidade Federal do Rio de Janeiro in 2008 (part of it was done in the Colorado State University). After that I spent two and a half years as a postdoc in the Reasoning and Learning Laboratory at McGill University. In 2013 I became an assistant researcher in the Department of Applied and Computational Mathematics at the National Laboratory for Scientific Computing. In 2016 I joined DeepMind, where I am now a research scientist.

Most of my research focuses on reinforcement learning, a sub-field of artificial intelligence that is concerned with situated agents that learn by interacting with the environment. Currently I am interested in principled ways of decomposing a reinforcement learning problem into simpler tasks whose solutions can be combined to quickly solve the original problem. The papers below give a good summary of what I have been up to in the last few years (more details here):

Successor Features for Transfer in Reinforcement Learning, NIPS 2017.

Transfer in Deep Reinforcement Learning Using Successor Features and Generalised Policy Improvement, ICML 2018.

The Option Keyboard: Combining Skills in Reinforcement Learning, NeurIPS 2019.