Understanding Inverse Reinforcement Learning under Overparameterization: Non-Asymptotic Analysis and Global Optimality

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Ruijia, Zeng, Siliang, Li, Chenliang, Garcia, Alfredo, Hong, Mingyi
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910889705209856
author Zhang, Ruijia
Zeng, Siliang
Li, Chenliang
Garcia, Alfredo
Hong, Mingyi
author_facet Zhang, Ruijia
Zeng, Siliang
Li, Chenliang
Garcia, Alfredo
Hong, Mingyi
contents The goal of the Inverse reinforcement learning (IRL) task is to identify the underlying reward function and the corresponding optimal policy from a set of expert demonstrations. While most IRL algorithms' theoretical guarantees rely on a linear reward structure, we aim to extend the theoretical understanding of IRL to scenarios where the reward function is parameterized by neural networks. Meanwhile, conventional IRL algorithms usually adopt a nested structure, leading to computational inefficiency, especially in high-dimensional settings. To address this problem, we propose the first two-timescale single-loop IRL algorithm under neural network parameterized reward and provide a non-asymptotic convergence analysis under overparameterization. Although prior optimality results for linear rewards do not apply, we show that our algorithm can identify the globally optimal reward and policy under certain neural network structures. This is the first IRL algorithm with a non-asymptotic convergence guarantee that provably achieves global optimality in neural network settings.
format Preprint
id arxiv_https___arxiv_org_abs_2503_17865
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Understanding Inverse Reinforcement Learning under Overparameterization: Non-Asymptotic Analysis and Global Optimality
Zhang, Ruijia
Zeng, Siliang
Li, Chenliang
Garcia, Alfredo
Hong, Mingyi
Machine Learning
The goal of the Inverse reinforcement learning (IRL) task is to identify the underlying reward function and the corresponding optimal policy from a set of expert demonstrations. While most IRL algorithms' theoretical guarantees rely on a linear reward structure, we aim to extend the theoretical understanding of IRL to scenarios where the reward function is parameterized by neural networks. Meanwhile, conventional IRL algorithms usually adopt a nested structure, leading to computational inefficiency, especially in high-dimensional settings. To address this problem, we propose the first two-timescale single-loop IRL algorithm under neural network parameterized reward and provide a non-asymptotic convergence analysis under overparameterization. Although prior optimality results for linear rewards do not apply, we show that our algorithm can identify the globally optimal reward and policy under certain neural network structures. This is the first IRL algorithm with a non-asymptotic convergence guarantee that provably achieves global optimality in neural network settings.
title Understanding Inverse Reinforcement Learning under Overparameterization: Non-Asymptotic Analysis and Global Optimality
topic Machine Learning
url https://arxiv.org/abs/2503.17865