Saved in:
Bibliographic Details
Main Authors: Yao, Qingmao, Lei, Zhichao, Chen, Tianyuan, Yuan, Ziyue, Chen, Xuefan, Liu, Jianxiang, Wu, Faguo, Zhang, Xiao
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2506.08417
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918052612800512
author Yao, Qingmao
Lei, Zhichao
Chen, Tianyuan
Yuan, Ziyue
Chen, Xuefan
Liu, Jianxiang
Wu, Faguo
Zhang, Xiao
author_facet Yao, Qingmao
Lei, Zhichao
Chen, Tianyuan
Yuan, Ziyue
Chen, Xuefan
Liu, Jianxiang
Wu, Faguo
Zhang, Xiao
contents Offline Reinforcement Learning (RL) struggles with distributional shifts, leading to the $Q$-value overestimation for out-of-distribution (OOD) actions. Existing methods address this issue by imposing constraints; however, they often become overly conservative when evaluating OOD regions, which constrains the $Q$-function generalization. This over-constraint issue results in poor $Q$-value estimation and hinders policy improvement. In this paper, we introduce a novel approach to achieve better $Q$-value estimation by enhancing $Q$-function generalization in OOD regions within Convex Hull and its Neighborhood (CHN). Under the safety generalization guarantees of the CHN, we propose the Smooth Bellman Operator (SBO), which updates OOD $Q$-values by smoothing them with neighboring in-sample $Q$-values. We theoretically show that SBO approximates true $Q$-values for both in-sample and OOD actions within the CHN. Our practical algorithm, Smooth Q-function OOD Generalization (SQOG), empirically alleviates the over-constraint issue, achieving near-accurate $Q$-value estimation. On the D4RL benchmarks, SQOG outperforms existing state-of-the-art methods in both performance and computational efficiency.
format Preprint
id arxiv_https___arxiv_org_abs_2506_08417
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Offline RL with Smooth OOD Generalization in Convex Hull and its Neighborhood
Yao, Qingmao
Lei, Zhichao
Chen, Tianyuan
Yuan, Ziyue
Chen, Xuefan
Liu, Jianxiang
Wu, Faguo
Zhang, Xiao
Machine Learning
Artificial Intelligence
Offline Reinforcement Learning (RL) struggles with distributional shifts, leading to the $Q$-value overestimation for out-of-distribution (OOD) actions. Existing methods address this issue by imposing constraints; however, they often become overly conservative when evaluating OOD regions, which constrains the $Q$-function generalization. This over-constraint issue results in poor $Q$-value estimation and hinders policy improvement. In this paper, we introduce a novel approach to achieve better $Q$-value estimation by enhancing $Q$-function generalization in OOD regions within Convex Hull and its Neighborhood (CHN). Under the safety generalization guarantees of the CHN, we propose the Smooth Bellman Operator (SBO), which updates OOD $Q$-values by smoothing them with neighboring in-sample $Q$-values. We theoretically show that SBO approximates true $Q$-values for both in-sample and OOD actions within the CHN. Our practical algorithm, Smooth Q-function OOD Generalization (SQOG), empirically alleviates the over-constraint issue, achieving near-accurate $Q$-value estimation. On the D4RL benchmarks, SQOG outperforms existing state-of-the-art methods in both performance and computational efficiency.
title Offline RL with Smooth OOD Generalization in Convex Hull and its Neighborhood
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2506.08417