Saved in:
Bibliographic Details
Main Authors: Yue, Sheng, Hua, Xingyuan, Chen, Lili, Ren, Ju
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2405.17471
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913367965302784
author Yue, Sheng
Hua, Xingyuan
Chen, Lili
Ren, Ju
author_facet Yue, Sheng
Hua, Xingyuan
Chen, Lili
Ren, Ju
contents Federated Reinforcement Learning (FRL) has garnered increasing attention recently. However, due to the intrinsic spatio-temporal non-stationarity of data distributions, the current approaches typically suffer from high interaction and communication costs. In this paper, we introduce a new FRL algorithm, named $\texttt{MFPO}$, that utilizes momentum, importance sampling, and additional server-side adjustment to control the shift of stochastic policy gradients and enhance the efficiency of data utilization. We prove that by proper selection of momentum parameters and interaction frequency, $\texttt{MFPO}$ can achieve $\tilde{\mathcal{O}}(H N^{-1}ε^{-3/2})$ and $\tilde{\mathcal{O}}(ε^{-1})$ interaction and communication complexities ($N$ represents the number of agents), where the interaction complexity achieves linear speedup with the number of agents, and the communication complexity aligns the best achievable of existing first-order FL algorithms. Extensive experiments corroborate the substantial performance gains of $\texttt{MFPO}$ over existing methods on a suite of complex and high-dimensional benchmarks.
format Preprint
id arxiv_https___arxiv_org_abs_2405_17471
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Momentum-Based Federated Reinforcement Learning with Interaction and Communication Efficiency
Yue, Sheng
Hua, Xingyuan
Chen, Lili
Ren, Ju
Machine Learning
Artificial Intelligence
Federated Reinforcement Learning (FRL) has garnered increasing attention recently. However, due to the intrinsic spatio-temporal non-stationarity of data distributions, the current approaches typically suffer from high interaction and communication costs. In this paper, we introduce a new FRL algorithm, named $\texttt{MFPO}$, that utilizes momentum, importance sampling, and additional server-side adjustment to control the shift of stochastic policy gradients and enhance the efficiency of data utilization. We prove that by proper selection of momentum parameters and interaction frequency, $\texttt{MFPO}$ can achieve $\tilde{\mathcal{O}}(H N^{-1}ε^{-3/2})$ and $\tilde{\mathcal{O}}(ε^{-1})$ interaction and communication complexities ($N$ represents the number of agents), where the interaction complexity achieves linear speedup with the number of agents, and the communication complexity aligns the best achievable of existing first-order FL algorithms. Extensive experiments corroborate the substantial performance gains of $\texttt{MFPO}$ over existing methods on a suite of complex and high-dimensional benchmarks.
title Momentum-Based Federated Reinforcement Learning with Interaction and Communication Efficiency
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2405.17471