Decentralized Federated Learning for Heterogeneous Multi-Task Semantic Communication

2026-08-15

AI Summary

This study addresses negative transfer and over-consensus bias in heterogeneous multi-task collaboration for distributed semantic communication by proposing a personalized framework. The approach employs policy-driven routing to disentangle features and utilizes a communication aggregation protocol to calibrate the consensus matrix, thereby mitigating interference. Furthermore, a closed-form solution for optimal aggregation depth is derived to balance the variance-bias trade-off. This work reveals a topology-mixed U-shaped pattern and establishes a unified Lyapunov drift analysis theory. Experiments on the NYU-v2 dataset demonstrate a 4.77% improvement in global performance, significantly outperforming baselines such as FedAvg and validating the method's effectiveness and robustness in handling heterogeneous collaborative tasks within distributed semantic communication systems.

Abstract

Collaborative training in distributed semantic communication (DSC) networks typically relies on decentralized federated learning (DFL). However, pushing topology-agnostic aggregation into heterogeneous, multi-task environments creates a fundamental bottleneck: it drives negative transfer and overconsensus bias (OCB). This paper introduces a personalized DSC framework that cuts off this cross-task interference. At the node level, a policy-driven multi-path routing mechanism separates task-specific features from shared representations to preserve local fidelity. Across the network, we deploy a "communicationwhile- aggregation" protocol. It calibrates a column-stochastic consensus matrix using task affinities. This limits the system to absorbing complementary knowledge while actively blocking mismatched parameter updates. To bound the convergence, we derive a unified Lyapunov drift analysis. We reveal a strict Ushaped trade-off: deeper topological mixing reduces variance but amplifies structural OCB. Resolving this tension yields a closed-form expression for the optimal aggregation depth. We evaluate the proposed framework on NYU-v2, where the results reveal a clear trade-off between insufficient aggregation and excessive topological mixing. At the analytically derived optimal aggregation depth, our method achieves a 4.77% global relative improvement over the no-aggregation baseline and outperforms decentralized FedAvg, FedAMP, and heuristic max aggregation. We further evaluate the framework on Taskonomy and imperfect wireless links to examine the effects of network-size variation and wireless-link reliability.

Problem

• Decentralized Federated Learning
• Semantic Communication
• Negative Transfer
• Overconsensus Bias
• Heterogeneous Multi-Task

Innovation

• Decentralized Federated Learning
• Semantic Communication
• Personalized Aggregation
• Communication-while-Aggregation
• Lyapunov Drift Analysis