Rohin Shah
TalkRL: The Reinforcement Learning Podcast
Is It Possible to Learn the Concept of a Waterfall?
It seems to me the goals here were things that could be modelled and learned. But do you have any comments on if weereto want goals that are harder to model than these things? I was tryng to think examples thi came up with, like irony or dance coreography scoring. Like, how you even begin to to model those things? Do we have to just continue improving our modelling tool kit so that we can make models of these reward functions? Or is there some other strategy? It depends exactly what you mean by improving the modelling tool kit but basically i think the answer is yes.
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
