Q-function Decomposition with Intervention Semantics with Factored Action Spaces

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lee, Junkyu, Gao, Tian, Nelson, Elliot, Liu, Miao, Bhattacharjya, Debarun, Lu, Songtao
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916714527064064
author Lee, Junkyu
Gao, Tian
Nelson, Elliot
Liu, Miao
Bhattacharjya, Debarun
Lu, Songtao
author_facet Lee, Junkyu
Gao, Tian
Nelson, Elliot
Liu, Miao
Bhattacharjya, Debarun
Lu, Songtao
contents Many practical reinforcement learning environments have a discrete factored action space that induces a large combinatorial set of actions, thereby posing significant challenges. Existing approaches leverage the regular structure of the action space and resort to a linear decomposition of Q-functions, which avoids enumerating all combinations of factored actions. In this paper, we consider Q-functions defined over a lower dimensional projected subspace of the original action space, and study the condition for the unbiasedness of decomposed Q-functions using causal effect estimation from the no unobserved confounder setting in causal statistics. This leads to a general scheme which we call action decomposed reinforcement learning that uses the projected Q-functions to approximate the Q-function in standard model-free reinforcement learning algorithms. The proposed approach is shown to improve sample complexity in a model-based reinforcement learning setting. We demonstrate improvements in sample efficiency compared to state-of-the-art baselines in online continuous control environments and a real-world offline sepsis treatment environment.
format Preprint
id arxiv_https___arxiv_org_abs_2504_21326
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Q-function Decomposition with Intervention Semantics with Factored Action Spaces
Lee, Junkyu
Gao, Tian
Nelson, Elliot
Liu, Miao
Bhattacharjya, Debarun
Lu, Songtao
Machine Learning
Artificial Intelligence
Many practical reinforcement learning environments have a discrete factored action space that induces a large combinatorial set of actions, thereby posing significant challenges. Existing approaches leverage the regular structure of the action space and resort to a linear decomposition of Q-functions, which avoids enumerating all combinations of factored actions. In this paper, we consider Q-functions defined over a lower dimensional projected subspace of the original action space, and study the condition for the unbiasedness of decomposed Q-functions using causal effect estimation from the no unobserved confounder setting in causal statistics. This leads to a general scheme which we call action decomposed reinforcement learning that uses the projected Q-functions to approximate the Q-function in standard model-free reinforcement learning algorithms. The proposed approach is shown to improve sample complexity in a model-based reinforcement learning setting. We demonstrate improvements in sample efficiency compared to state-of-the-art baselines in online continuous control environments and a real-world offline sepsis treatment environment.
title Q-function Decomposition with Intervention Semantics with Factored Action Spaces
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2504.21326