This paper explores how language models can be used in reinforcement learning agents to improve exploration efficiency, data reuse, skill scheduling, and learning from observations. The method is tested in a robotic manipulation environment and shows significant performance improvements.

https://arxiv.org/abs//2307.09668

YouTube: https://www.youtube.com/@ArxivPapers

PODCASTS:

Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016

Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers