You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Language Models and Vision Language Models have recently demonstratedunprecedented capabilities in terms of understanding human intentions,reasoning, scene understanding, and planning-like behaviour, in text form,among many others. In this work, we investigate how to embed and leverage suchabilities in Reinforcement Learning (RL) agents. We design a framework thatuses language as the core reasoning tool, exploring how this enables an agentto tackle a series of fundamental RL challenges, such as efficient exploration,reusing experience data, scheduling skills, and learning from observations,which traditionally require separate, vertically designed algorithms. We testour method on a sparse-reward simulated robotic manipulation environment, wherea robot needs to stack a set of objects. We demonstrate substantial performanceimprovements over baselines in exploration efficiency and ability to reuse datafrom offline datasets, and illustrate how to reuse learned skills to solvenovel tasks or imitate videos of human experts.
URL
Affiliations
Abstract
Translation (by gpt-3.5-turbo)
Summary (by gpt-3.5-turbo)
The text was updated successfully, but these errors were encountered: