Decision-Point Guided Safe Policy Improvement

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sharma, Abhishek, Benac, Leo, Parbhoo, Sonali, Doshi-Velez, Finale
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916435162300416
author Sharma, Abhishek
Benac, Leo
Parbhoo, Sonali
Doshi-Velez, Finale
author_facet Sharma, Abhishek
Benac, Leo
Parbhoo, Sonali
Doshi-Velez, Finale
contents Within batch reinforcement learning, safe policy improvement (SPI) seeks to ensure that the learnt policy performs at least as well as the behavior policy that generated the dataset. The core challenge in SPI is seeking improvements while balancing risk when many state-action pairs may be infrequently visited. In this work, we introduce Decision Points RL (DPRL), an algorithm that restricts the set of state-action pairs (or regions for continuous states) considered for improvement. DPRL ensures high-confidence improvement in densely visited states (i.e. decision points) while still utilizing data from sparsely visited states. By appropriately limiting where and how we may deviate from the behavior policy, we achieve tighter bounds than prior work; specifically, our data-dependent bounds do not scale with the size of the state and action spaces. In addition to the analysis, we demonstrate that DPRL is both safe and performant on synthetic and real datasets.
format Preprint
id arxiv_https___arxiv_org_abs_2410_09361
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Decision-Point Guided Safe Policy Improvement
Sharma, Abhishek
Benac, Leo
Parbhoo, Sonali
Doshi-Velez, Finale
Machine Learning
Within batch reinforcement learning, safe policy improvement (SPI) seeks to ensure that the learnt policy performs at least as well as the behavior policy that generated the dataset. The core challenge in SPI is seeking improvements while balancing risk when many state-action pairs may be infrequently visited. In this work, we introduce Decision Points RL (DPRL), an algorithm that restricts the set of state-action pairs (or regions for continuous states) considered for improvement. DPRL ensures high-confidence improvement in densely visited states (i.e. decision points) while still utilizing data from sparsely visited states. By appropriately limiting where and how we may deviate from the behavior policy, we achieve tighter bounds than prior work; specifically, our data-dependent bounds do not scale with the size of the state and action spaces. In addition to the analysis, we demonstrate that DPRL is both safe and performant on synthetic and real datasets.
title Decision-Point Guided Safe Policy Improvement
topic Machine Learning
url https://arxiv.org/abs/2410.09361