Dec 18, 2019
Historically, the research communities of reinforcement learning (RL) and game theory (GT) have been rather distinct. The formalism used were different, the environments of interest were different, and the questions asked were different as well. But on high level - both approaches look for agent's optimal policy in an environment. And games are environments that have always been of interest to both communities. Recently, the RL/GT "schism" is being bridged as the new challenging games require insights from both communities. There is a need for new algorithms and techniques to tackle these challenging games. DeepMind Edmonton lab is a cold place in Canada, an ideal place to do game research that bridges these two worlds. Not only the lab includes most of the team that developed DeepStack, the lab itself includes bright RL researchers and is led by Rich Sutton. Martin will talk about what has already happened in bridging these communities, and what he believes the future holds.
Spojení sil ČVUT v Praze, Univerzity Karlovy, Akademie věd České republiky a hlavního město Prahy za účelem přeměny naší metropole na centrum umělé inteligence evropského významu.
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Presentations on similar topic, category or speaker