Sai Krishna Gottipati
TalkRL: The Reinforcement Learning Podcast
00:00
Multi Teacher Approach to Self Play in Deep Reinforcement Learning
A paper that you have currently under review at a conference. The papres calling do as you teach, a multi teacher approached to self play in deep reinforcement learning. Si: We believe that we need multiple teachers o to set divers goals. For that could also help in faster learning of the student agent and also betterer better generalization. And where does this docasticity come from? The randomness in the teachers its random inicializ your random inicialization of the networks.
Play episode from 25:19
Transcript
