[2603.24599] A Learnable SIM Paradigm: Fundamentals, Training Techniques, and Applications

Electrical Engineering and Systems Science > Signal Processing arXiv:2603.24599 (eess) [Submitted on 13 Mar 2026] Title:A Learnable SIM Paradigm: Fundamentals, Training Techniques, and Applications Authors:Hetong Wang, Yashuai Cao, Tiejun Lv View a PDF of the paper titled A Learnable SIM Paradigm: Fundamentals, Training Techniques, and Applications, by Hetong Wang and 2 other authors View PDF HTML (experimental) Abstract:Stacked intelligent metasurfaces (SIMs) represent a breakthrough in wireless hardware by comprising multilayer, programmable metasurfaces capable of analog computing in the electromagnetic (EM) wave domain. By examining their architectural analogies, this article reveals a deeper connection between SIMs and artificial neural networks (ANNs). Leveraging this profound structural similarity, this work introduces a learnable SIM architecture and proposes a learnable SIM-based machine learning (ML) paradigm for sixth-generation (6G)-andbeyond systems. Then, we develop two SIM-empowered wireless signal processing schemes to effectively achieve multi-user signal separation and distinguish communication signals from jamming signals. The use cases highlight that the proposed SIM-enabled signal processing system can significantly enhance spectrum utilization efficiency and anti-jamming capability in a lightweight manner and pave the way for ultra-efficient and intelligent wireless infrastructures. Comments: Subjects: Signal Processing (eess.SP); Artificial Intellige...