Offline Reinforcement Learning from Datasets with Structured Non-Stationarity

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Ackermann, Johannes, Osa, Takayuki, Sugiyama, Masashi
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916262569836544
author Ackermann, Johannes
Osa, Takayuki
Sugiyama, Masashi
author_facet Ackermann, Johannes
Osa, Takayuki
Sugiyama, Masashi
contents Current Reinforcement Learning (RL) is often limited by the large amount of data needed to learn a successful policy. Offline RL aims to solve this issue by using transitions collected by a different behavior policy. We address a novel Offline RL problem setting in which, while collecting the dataset, the transition and reward functions gradually change between episodes but stay constant within each episode. We propose a method based on Contrastive Predictive Coding that identifies this non-stationarity in the offline dataset, accounts for it when training a policy, and predicts it during evaluation. We analyze our proposed method and show that it performs well in simple continuous control tasks and challenging, high-dimensional locomotion tasks. We show that our method often achieves the oracle performance and performs better than baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2405_14114
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Offline Reinforcement Learning from Datasets with Structured Non-Stationarity
Ackermann, Johannes
Osa, Takayuki
Sugiyama, Masashi
Machine Learning
Artificial Intelligence
Current Reinforcement Learning (RL) is often limited by the large amount of data needed to learn a successful policy. Offline RL aims to solve this issue by using transitions collected by a different behavior policy. We address a novel Offline RL problem setting in which, while collecting the dataset, the transition and reward functions gradually change between episodes but stay constant within each episode. We propose a method based on Contrastive Predictive Coding that identifies this non-stationarity in the offline dataset, accounts for it when training a policy, and predicts it during evaluation. We analyze our proposed method and show that it performs well in simple continuous control tasks and challenging, high-dimensional locomotion tasks. We show that our method often achieves the oracle performance and performs better than baselines.
title Offline Reinforcement Learning from Datasets with Structured Non-Stationarity
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2405.14114