FAWAC: Feasibility Informed Advantage Weighted Regression for Persistent Safety in Offline Reinforcement Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Koirala, Prajwal, Jiang, Zhanhong, Sarkar, Soumik, Fleming, Cody
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912153406013440
author Koirala, Prajwal
Jiang, Zhanhong
Sarkar, Soumik
Fleming, Cody
author_facet Koirala, Prajwal
Jiang, Zhanhong
Sarkar, Soumik
Fleming, Cody
contents Safe offline reinforcement learning aims to learn policies that maximize cumulative rewards while adhering to safety constraints, using only offline data for training. A key challenge is balancing safety and performance, particularly when the policy encounters out-of-distribution (OOD) states and actions, which can lead to safety violations or overly conservative behavior during deployment. To address these challenges, we introduce Feasibility Informed Advantage Weighted Actor-Critic (FAWAC), a method that prioritizes persistent safety in constrained Markov decision processes (CMDPs). FAWAC formulates policy optimization with feasibility conditions derived specifically for offline datasets, enabling safe policy updates in non-parametric policy space, followed by projection into parametric space for constrained actor training. By incorporating a cost-advantage term into Advantage Weighted Regression (AWR), FAWAC ensures that the safety constraints are respected while maximizing performance. Additionally, we propose a strategy to address a more challenging class of problems that involves tempting datasets where trajectories are predominantly high-rewarded but unsafe. Empirical evaluations on standard benchmarks demonstrate that FAWAC achieves strong results, effectively balancing safety and performance in learning policies from the static datasets.
format Preprint
id arxiv_https___arxiv_org_abs_2412_08880
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle FAWAC: Feasibility Informed Advantage Weighted Regression for Persistent Safety in Offline Reinforcement Learning
Koirala, Prajwal
Jiang, Zhanhong
Sarkar, Soumik
Fleming, Cody
Machine Learning
Safe offline reinforcement learning aims to learn policies that maximize cumulative rewards while adhering to safety constraints, using only offline data for training. A key challenge is balancing safety and performance, particularly when the policy encounters out-of-distribution (OOD) states and actions, which can lead to safety violations or overly conservative behavior during deployment. To address these challenges, we introduce Feasibility Informed Advantage Weighted Actor-Critic (FAWAC), a method that prioritizes persistent safety in constrained Markov decision processes (CMDPs). FAWAC formulates policy optimization with feasibility conditions derived specifically for offline datasets, enabling safe policy updates in non-parametric policy space, followed by projection into parametric space for constrained actor training. By incorporating a cost-advantage term into Advantage Weighted Regression (AWR), FAWAC ensures that the safety constraints are respected while maximizing performance. Additionally, we propose a strategy to address a more challenging class of problems that involves tempting datasets where trajectories are predominantly high-rewarded but unsafe. Empirical evaluations on standard benchmarks demonstrate that FAWAC achieves strong results, effectively balancing safety and performance in learning policies from the static datasets.
title FAWAC: Feasibility Informed Advantage Weighted Regression for Persistent Safety in Offline Reinforcement Learning
topic Machine Learning
url https://arxiv.org/abs/2412.08880