AI RESEARCH
Exploiting Local Dynamics Regularity for Reusable Skills in Offline Hierarchical RL
arXiv CS.AI
•
ArXi:2605.26371v1 Announce Type: new Hierarchical Reinforcement Learning (HRL) promises to solve long-horizon Reinforcement Learning (RL) tasks efficiently than non-hierarchical counterparts by discovering and reusing temporally-extended skills. However, obtaining skills that are actually reusable remains an open challenge. Towards this end, we focus on abstractions that exploit the intuition of local dynamics: local transitions in different global contexts require similar kinds of action sequences.