Exploiting Local Dynamics Regularity for Reusable Skills in Offline Hierarchical RL
The paper discusses advancements in Hierarchical Reinforcement Learning (HRL) by focusing on the reuse of skills through local dynamics regularity. The proposed algorithm, CARL, aims to enhance the efficiency of HRL by aligning action sequences with their corresponding contexts. This approach demonstrates improved performance in complex environments and on benchmark tests.
- ▪Hierarchical Reinforcement Learning aims to solve long-horizon tasks more efficiently by discovering reusable skills.
- ▪The CARL algorithm aligns local transitions with action sequences to identify reusable skills.
- ▪The approach shows qualitative clustering of skills and improved performance on the OGBench benchmark.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Artificial Intelligence arXiv:2605.26371 (cs) [Submitted on 25 May 2026] Title:Exploiting Local Dynamics Regularity for Reusable Skills in Offline Hierarchical RL Authors:Sarthak Dayal, Abhinav Peri, Carl Qi, Claas Voelcker, Alexander Levine, Caleb Chuck, Amy Zhang View a PDF of the paper titled Exploiting Local Dynamics Regularity for Reusable Skills in Offline Hierarchical RL, by Sarthak Dayal and 6 other authors View PDF HTML (experimental) Abstract:Hierarchical Reinforcement Learning (HRL) promises to solve long-horizon Reinforcement Learning (RL) tasks more efficiently than non-hierarchical counterparts by discovering and reusing temporally-extended skills. However, obtaining skills that are actually reusable remains an open challenge.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
