A Finite-Time Analysis of Q-Learning with Neural Network Function Approximation
Lists
Only you see this
Abstract
Q-learning with neural network function approximation (neural Q-learning for short) is among the most prevalent deep reinforcement learning algorithms. Despite its empirical success, the non-asymptotic convergence rate of neural Q-learning remains virtually unknown. In this paper, we present a finite-time analysis of a neural Q-learning algorithm, where the data are generated from a Markov decision process and the action-value function is approximated by a deep ReLU neural network. We prove that neural Q-learning finds the optimal policy with $O(1/\sqrt{T})$ convergence rate if the neural function approximator is sufficiently overparameterized, where $T$ is the number of iterations. To our best knowledge, our result is the first finite-time analysis of neural Q-learning under non-i.i.d. data assumption.
BibTeX
@inproceedings{xu2020finite,
author = {Pan Xu and Quanquan Gu},
title = {{A Finite-Time Analysis of Q-Learning with Neural Network Function Approximation}},
booktitle = {International Conference on Machine Learning},
year = {2020},
url = {https://openreview.net/forum?id=4br2QUQavld},
}