Saved in:
Bibliographic Details
Main Authors: Tang, Xiaohang, Marques, Afonso, Kamalaruban, Parameswaran, Bogunovic, Ilija
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2407.18414
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916464804495360
author Tang, Xiaohang
Marques, Afonso
Kamalaruban, Parameswaran
Bogunovic, Ilija
author_facet Tang, Xiaohang
Marques, Afonso
Kamalaruban, Parameswaran
Bogunovic, Ilija
contents Decision Transformer (DT), as one of the representative Reinforcement Learning via Supervised Learning (RvS) methods, has achieved strong performance in offline learning tasks by leveraging the powerful Transformer architecture for sequential decision-making. However, in adversarial environments, these methods can be non-robust, since the return is dependent on the strategies of both the decision-maker and adversary. Training a probabilistic model conditioned on observed return to predict action can fail to generalize, as the trajectories that achieve a return in the dataset might have done so due to a suboptimal behavior adversary. To address this, we propose a worst-case-aware RvS algorithm, the Adversarially Robust Decision Transformer (ARDT), which learns and conditions the policy on in-sample minimax returns-to-go. ARDT aligns the target return with the worst-case return learned through minimax expectile regression, thereby enhancing robustness against powerful test-time adversaries. In experiments conducted on sequential games with full data coverage, ARDT can generate a maximin (Nash Equilibrium) strategy, the solution with the largest adversarial robustness. In large-scale sequential games and continuous adversarial RL environments with partial data coverage, ARDT demonstrates significantly superior robustness to powerful test-time adversaries and attains higher worst-case returns compared to contemporary DT methods.
format Preprint
id arxiv_https___arxiv_org_abs_2407_18414
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Adversarially Robust Decision Transformer
Tang, Xiaohang
Marques, Afonso
Kamalaruban, Parameswaran
Bogunovic, Ilija
Machine Learning
Artificial Intelligence
Decision Transformer (DT), as one of the representative Reinforcement Learning via Supervised Learning (RvS) methods, has achieved strong performance in offline learning tasks by leveraging the powerful Transformer architecture for sequential decision-making. However, in adversarial environments, these methods can be non-robust, since the return is dependent on the strategies of both the decision-maker and adversary. Training a probabilistic model conditioned on observed return to predict action can fail to generalize, as the trajectories that achieve a return in the dataset might have done so due to a suboptimal behavior adversary. To address this, we propose a worst-case-aware RvS algorithm, the Adversarially Robust Decision Transformer (ARDT), which learns and conditions the policy on in-sample minimax returns-to-go. ARDT aligns the target return with the worst-case return learned through minimax expectile regression, thereby enhancing robustness against powerful test-time adversaries. In experiments conducted on sequential games with full data coverage, ARDT can generate a maximin (Nash Equilibrium) strategy, the solution with the largest adversarial robustness. In large-scale sequential games and continuous adversarial RL environments with partial data coverage, ARDT demonstrates significantly superior robustness to powerful test-time adversaries and attains higher worst-case returns compared to contemporary DT methods.
title Adversarially Robust Decision Transformer
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2407.18414