In this study, we develop an enhanced in-context learning (ICL) framework to improve the performance of pilot-based beamforming in multi-user multiple-input single-output (MU-MISO) systems. The proposed scheme integrates the ICL-Transformer backbone with the pilot encoder-decoder network (EDN) and the beamformer EDN. A crucial feature of our ICL network is its ability to handle multiple channel models without retraining, enabled by the construction of model-specific context datasets.
To enhance convergence and robustness, we introduce three key innovations:
(a) a curriculum learning (CL) strategy that smoothly transitions from supervised LMMSE-labeled imitation to unsupervised sum-rate maximization;
(b) a self-evolving mechanism that dynamically expands and refines the context datasets for all channel models during CL-based training;
(c) a mismatch-aware extension that incorporates several mismatches into the general ICL framework, bypassing explicit channel calibrations.
Ablation studies validate the effectiveness of the in-context architecture and enhanced training strategies. Simulation results across diverse communication environments demonstrate that the proposed scheme can rapidly adapt to both seen and unseen channel models without gradient-based parameter updates, and mitigate mismatch issues via intelligent context constructions. Furthermore, our scheme consistently outperforms existing beamforming schemes under pilot-based settings, including the WMMSE benchmark and recent Transformer-based methods.
Blogger's Review: This study significantly enhances beamforming efficiency in MU-MISO systems by introducing self-evolving mechanisms and curriculum learning strategies, showcasing the potential of in-context learning in wireless communications. The framework not only improves adaptability but also addresses channel mismatch issues, warranting further exploration and application.