School of Electronic Engineering and Computer Science

Seminar: Tom Schaul: Challenges of deep reinforcement learning, and the example of AlphaStar

When: Thursday, March 12, 2020, 5:00 PM - 6:00 PM
Where: Peston Lecture Theatre, Mile End, Queen Mary University of London,

Speaker: Tom Schaul (Google DeepMind)

Abstract: This talk is centered on deep reinforcement learning, it's potential and its challenges. I will give a birds-eye overview of the field, and then dive into some of the trickier challenges that are driving research today. In parallel, I will motivate how useful games are for this kind of research. On the solution side, I will describe some of the techniques that have recently helped a deep RL system achieve grandmaster level in StarCraft 2.

Bio: Tom Schaul has been a researcher at DeepMind for 6 years. His research is focused on reinforcement learning with deep neural networks, but includes modular and continual learning, black-box optimization, temporal and state abstractions, off-policy learning about many goals simultaneously, and video-game benchmarks. Tom grew up in Luxembourg and studied computer science in Switzerland (with exchanges at Waterloo and Columbia), where he obtained an MSc from the EPFL in 2005. He holds a PhD from TU Munich (2011), which he did under the supervision of J├╝rgen Schmidhuber at the Swiss AI Lab IDSIA. From 2011 to 2013, he did a postdoc with Yann LeCun at the Courant Institute of NYU.

The talk will be followed by informal drinks and nibbles at The Hub at 6pm.