Sai Krishna Gottipati
TalkRL: The Reinforcement Learning Podcast
00:00
Inductive Bias
Ike: It's very unclear if the usual cconitionnitworks are suitable to these kind of representations. E, like a rook and move all across the n board all at once. And that seems like a sean end is not going to be very suitable for that part. You'll never have enough a data if the game is hard enough. I think iwould really consider it a sort or if an agent can play without any er research. Y is lot of open prolem supeexpoent.
Play episode from 59:29
Transcript
