RISnet‐LiteMix: A Lightweight MLP‐Mixer‐Based Network for RIS Phase Shift Optimization

ABSTRACT Recent deep learning‐based reconfigurable intelligent surfaces (RIS) phase control models achieve competitive performance, but their scalability is constrained by rapidly increasing computational complexity as the number of reflective elements grows. We propose RISnet‐LiteMix, a lightweight MLP‐Mixer‐based RIS phase control model that captures inter‐user interactions with significantly lower computational complexity. In particular, the LiteMixer block in RISnet‐LiteMix applies dimensional compression or expansion across different perspectives, thereby reducing computational burden while enabling effective learning of channel representations. Consequently, RISnet‐LiteMix reduces the number of parameters by 70% and computational complexity by 47% relative to existing MLP‐Mixer‐based models, while maintaining comparable transmission performance.