✳flâneur — a map of the web's best reading

[2010.15638] Abstract Value Iteration for Hierarchical Reinforcement Learning

arxiv.org · 652 words · saved by 1 readers

We propose a novel hierarchical reinforcement learning framework for control with continuous state and action spaces. In our framework, the user specifies subgoal regions which are subsets of states; then, we (i) learn options that serve as transitions between these subgoal regions, and (ii) construct a high-level plan in the resulting abstract decision process (ADP). A key challenge is that the ADP may not be Markov, which we address by proposing two algorithms for planning in the ADP. Our first algorithm is conservative, allowing us to prove theoretical guarantees on its performance, which help inform the design of subgoal regions. Our second algorithm is a practical one that interweaves planning at the abstract level and learning at the concrete level. In our experiments, we demonstrate that our approach outperforms state-of-the-art hierarchical reinforcement learning algorithms on several challenging benchmarks.

[2010.15638] Abstract Value Iteration for Hierarchical Reinforcement Learning Skip to main content arXiv is now an independent nonprofit! Learn more × Search arXiv Press Enter to search · Advanced search --> Computer Science > Machine Learning arXiv:2010.15638 (cs) [Submitted on 29 Oct 2020 ( v1 ), last revised 25 Feb 2021 (this version, v2)] Title: Abstract Value Iteration for Hierarchical Reinforcement Learning Authors: Kishor Jothimurugan , Osbert Bastani , Rajeev Alur View a PDF of the paper titled Abstract Value Iteration for Hierarchical Reinforcement Learning, by Kishor Jot

Explore this link on the map →

related reading