Saved in:
Bibliographic Details
Main Authors: Chen, Xiaocong, Wang, Siyu, Yu, Tong, Yao, Lina
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2403.17646
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918078513676288
author Chen, Xiaocong
Wang, Siyu
Yu, Tong
Yao, Lina
author_facet Chen, Xiaocong
Wang, Siyu
Yu, Tong
Yao, Lina
contents Offline reinforcement learning (RL) presents distinct challenges as it relies solely on observational data. A central concern in this context is ensuring the safety of the learned policy by quantifying uncertainties associated with various actions and environmental stochasticity. Traditional approaches primarily emphasize mitigating epistemic uncertainty by learning risk-averse policies, often overlooking environmental stochasticity. In this study, we propose an uncertainty-aware distributional offline RL method to simultaneously address both epistemic uncertainty and environmental stochasticity. We propose a model-free offline RL algorithm capable of learning risk-averse policies and characterizing the entire distribution of discounted cumulative rewards, as opposed to merely maximizing the expected value of accumulated discounted returns. Our method is rigorously evaluated through comprehensive experiments in both risk-sensitive and risk-neutral benchmarks, demonstrating its superior performance.
format Preprint
id arxiv_https___arxiv_org_abs_2403_17646
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Diffusion Policies for Risk-Averse Behavior Modeling in Offline Reinforcement Learning
Chen, Xiaocong
Wang, Siyu
Yu, Tong
Yao, Lina
Machine Learning
Offline reinforcement learning (RL) presents distinct challenges as it relies solely on observational data. A central concern in this context is ensuring the safety of the learned policy by quantifying uncertainties associated with various actions and environmental stochasticity. Traditional approaches primarily emphasize mitigating epistemic uncertainty by learning risk-averse policies, often overlooking environmental stochasticity. In this study, we propose an uncertainty-aware distributional offline RL method to simultaneously address both epistemic uncertainty and environmental stochasticity. We propose a model-free offline RL algorithm capable of learning risk-averse policies and characterizing the entire distribution of discounted cumulative rewards, as opposed to merely maximizing the expected value of accumulated discounted returns. Our method is rigorously evaluated through comprehensive experiments in both risk-sensitive and risk-neutral benchmarks, demonstrating its superior performance.
title Diffusion Policies for Risk-Averse Behavior Modeling in Offline Reinforcement Learning
topic Machine Learning
url https://arxiv.org/abs/2403.17646