Collaborative Learning for Less Online Retraining of Neural Receivers
Tianxin Wang, Shuo Wang, Xudong Wang, Geoffrey Ye Li · 2024
Offline-trained neural receivers achieve significant performance gains. Yet, online retraining is required to sustain such gains in a new environment. Instead of retraining whenever a new channel environment arises, a multi-cell collaborative learning framework is designed to enable the neural receivers to generalize to unseen sce-narios, thus preventing frequent retraining. This framework features two key designs: 1) the personalized federated learning paradigm is exploited to strike a generalization-personalization balance, with each model sharing a global representation network and personal-izing the local head network; 2) an online data filtering mechanism is designed to filter out low-impact data samples. According to simulations, the collaboratively-learned receivers outperform the traditional ones by over 3 dB and improves the generalization per-formance by 5.2 dB in the unseen scenarios.