Near-Optimal Regret in Linear MDPs with Aggregate Bandit Feedback

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Cassel, Asaf, Luo, Haipeng, Rosenberg, Aviv, Sotnikov, Dmitry
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909201507287040
author Cassel, Asaf
Luo, Haipeng
Rosenberg, Aviv
Sotnikov, Dmitry
author_facet Cassel, Asaf
Luo, Haipeng
Rosenberg, Aviv
Sotnikov, Dmitry
contents In many real-world applications, it is hard to provide a reward signal in each step of a Reinforcement Learning (RL) process and more natural to give feedback when an episode ends. To this end, we study the recently proposed model of RL with Aggregate Bandit Feedback (RL-ABF), where the agent only observes the sum of rewards at the end of an episode instead of each reward individually. Prior work studied RL-ABF only in tabular settings, where the number of states is assumed to be small. In this paper, we extend ABF to linear function approximation and develop two efficient algorithms with near-optimal regret guarantees: a value-based optimistic algorithm built on a new randomization technique with a Q-functions ensemble, and a policy optimization algorithm that uses a novel hedging scheme over the ensemble.
format Preprint
id arxiv_https___arxiv_org_abs_2405_07637
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Near-Optimal Regret in Linear MDPs with Aggregate Bandit Feedback
Cassel, Asaf
Luo, Haipeng
Rosenberg, Aviv
Sotnikov, Dmitry
Machine Learning
In many real-world applications, it is hard to provide a reward signal in each step of a Reinforcement Learning (RL) process and more natural to give feedback when an episode ends. To this end, we study the recently proposed model of RL with Aggregate Bandit Feedback (RL-ABF), where the agent only observes the sum of rewards at the end of an episode instead of each reward individually. Prior work studied RL-ABF only in tabular settings, where the number of states is assumed to be small. In this paper, we extend ABF to linear function approximation and develop two efficient algorithms with near-optimal regret guarantees: a value-based optimistic algorithm built on a new randomization technique with a Q-functions ensemble, and a policy optimization algorithm that uses a novel hedging scheme over the ensemble.
title Near-Optimal Regret in Linear MDPs with Aggregate Bandit Feedback
topic Machine Learning
url https://arxiv.org/abs/2405.07637