Semantic-Aware Task Clustering for Federated Cooperative Multi-Task Semantic Communication

TL;DR

Problem: extend task-oriented SemCom to distributed multi-user networks while avoiding destructive cross-task transfer. Approach: a Clustered-Federated-CMT-SemCom framework that uses InfoMax-based local training and semantic-aware clustering (via Jensen-Shannon divergence) to form task clusters and cluster-wise encoder aggregation on a parameter server. Contributions: (i) FL-based cooperative multi-task SemCom for distributed users, (ii) semantic clustering yielding disjoint task groups without data sharing, and (iii) empirical validation in a LEO satellite MNIST setup showing accuracy gains and faster convergence. Findings: significant improvements over unclustered FL and independent training, with practical reductions in communication. Significance: enables scalable, task-aware SemCom across non-terrestrial networks.

Abstract

Task-oriented semantic communication (SemCom) prioritizes task execution over accurate symbol reconstruction and is well-suited to emerging intelligent applications. Cooperative multi-task SemCom (CMT-SemCom) further improves task execution performance. However, [1] demonstrates that cooperative multi-tasking can be either constructive or destructive. Moreover, the existing CMT-SemCom framework is not directly applicable to distributed multi-user scenarios, such as non-terrestrial satellite networks, where each satellite employs an individual semantic encoder. In this paper, we extend our earlier CMT-SemCom framework to distributed settings by proposing a federated learning (FL) based CMT-SemCom that enables cooperative multi-tasking across distributed users. Moreover, to address performance degradation caused by negative information transfer among heterogeneous tasks, we propose a semantic-aware task clustering method integrated in the FL process to ensure constructive cooperation based on an information-theoretic approach. Unlike common clustering methods that rely on high-dimensional data or feature space similarity, our proposed approach operates in the low-dimensional semantic domain to identify meaningful task relationships. Simulation results based on a LEO satellite network setup demonstrate the effectiveness of our approach and performance gain over unclustered FL and individual single-task SemCom.

Figures (3)

• Figure 1: Illustration of the proposed Clustered-Federated-CMT-SemCom framework applied to LEO satellite network.
• Figure 2: Performance comparison of the proposed framework vs. the benchmarks: Unclustered FL-CMT-SemCom and Individual training on the test set.
• Figure 3: Comparison of the overall network performance of the approaches.