Federated Client Selection under Partial Visibility: A POMDP Approach with Spatio-Temporal Attention

2026-05-12

AI Summary

This work addresses the challenge in federated learning where the server, constrained by communication overhead or client availability, can only observe a subset of clients, making effective participant selection particularly difficult under data heterogeneity. To tackle this issue, the paper formulates the problem for the first time as a partially observable Markov decision process (POMDP) and introduces a reinforcement learning framework enhanced with spatiotemporal attention mechanisms. This framework integrates historical global model states with client identity embeddings to jointly capture temporal context and individual client characteristics. By explicitly accounting for both partial observability and statistical heterogeneity, the proposed method significantly outperforms existing baselines across multiple datasets, achieving faster convergence and improved final model performance.

Abstract

Federated learning relies on effective client selection to alleviate the performance degradation caused by data heterogeneity. Most existing methods assume full visibility of all clients at each communication round. However, in large-scale or edge-based deployments, the server can only access a subset of clients due to communication, mobility, or availability constraints, resulting in partial visibility where only a subset of clients is observable for aggregation in each communication round. In this paper, we formulate federated client selection under partial visibility as a Partially Observable Markov Decision Process (POMDP) and propose a Spatial-Temporal attention-based reinforcement learning framework. By integrating historical global models and client identity embeddings, the proposed method captures both the temporal contexts of training and the persistent characteristics of clients. Experimental results across multiple datasets demonstrate that our approach achieves superior performance compared to existing baselines in heterogeneous and partially visible settings, validating its effectiveness in addressing the challenges of incomplete observations in practical federated learning systems.

Problem

• federated learning
• client selection
• partial visibility
• data heterogeneity
• POMDP

Innovation

• POMDP
• spatio-temporal attention
• federated client selection
• partial visibility
• reinforcement learning