aiqbench analysis
This research advances the intersection of machine learning and communications by addressing critical failure modes in federated semantic communication, specifically instability under non-IID data and lack of channel adaptation. The proposed FedGenSC framework leverages generative priors to maintain semantic fidelity across heterogeneous link conditions, demonstrating significant performance gains over existing discriminative baselines.
FedGenSC: Federated Generative Semantic Communication with Channel-Aware Adaptation
arXiv:2609.08593v1 Announce Type: cross Abstract: Integrating generative adversarial networks (GANs) into federated semantic communication (SemCom) is a natural progression, as generative priors can recover semantic fidelity under channel distortion that discriminative decoders cannot. However, naive GAN federation introduces three failure modes that prior work has, to the best of our knowledge, neither identified nor resolved: discriminator aggregation instability under non-independent and identically distributed (non-IID) data, semantic drift caused by divergent local embedding spaces, and channel-agnostic generation that cannot adapt to heterogeneous link conditions. We propose federated generative semantic communication (FedGenSC), which mitigates all three by employing a global generator with local-only discriminators, providing cross-client semantic information through a semantic prototype bank, and conditioning generation on the instantaneous signal-to-noise ratio (SNR). Experiments on the Europarl dataset over Rayleigh fading channels (K=10 clients, Dirichlet {\alpha}=0.5) show that FedGenSC under non-IID data outperforms the FedDeepSC baseline across the tested SNR range, achieving up to a 58.2% relative improvement in bilingual evaluation understudy (BLEU)-1 at 18 dB. Ablation studies confirm the independent contribution of each component.
Abstract:Integrating generative adversarial networks (GANs) into federated semantic communication (SemCom) is a natural progression, as generative priors can recover semantic fidelity under channel distortion that discriminative decoders cannot. However, naive GAN federation introduces three failure modes that prior work has, to the best of our knowledge, neither identified nor resolved: discriminator aggregation instability under non-independent and identically distributed (non-IID) data, semantic drift caused by divergent local embedding spaces, and channel-agnostic generation that cannot adapt to heterogeneous link conditions. We propose federated generative semantic communication (FedGenSC), which mitigates all three by employing a global generator with local-only discriminators, providing cross-client semantic information through a semantic prototype bank, and conditioning generation on the instantaneous signal-to-noise ratio (SNR). Experiments on the Europarl dataset over Rayleigh fading channels (K=10 clients, Dirichlet {\alpha}=0.5) show that FedGenSC under non-IID data outperforms the FedDeepSC baseline across the tested SNR range, achieving up to a 58.2% relative improvement in bilingual evaluation understudy (BLEU)-1 at 18 dB. Ablation studies confirm the independent contribution of each component.
6 pages, 3 figures, 4 tables. Accepted for presentation at IEEE GLOBECOM 2026, Macau. This is the author's accepted version; the final published version will be available via IEEE Xplore
Signal Processing (eess.SP); Machine Learning (cs.LG)
arXiv:2609.08593 [eess.SP]
(or arXiv:2609.08593v1 [eess.SP] for this version)
https://doi.org/10.48550/arXiv.2609.08593 arXiv-issued DOI via DataCite (pending registration)
Full text under the publisher's open licence · 218 words