Customized Design of Twist‐Stacking Films Towards Polarization Multiplexing and Full Polarization Detection

Abstract Polarization multiplexing has garnered increasing attention due to its superior performance in sensing, communication and imaging. Extensive studies in this field are mainly focused on metasurfaces, however, some drawbacks including high costs, complicated fabrication processes and the limited information capacity imped their practical applications. Herein, this study proposes a machine learning strategy to fabricate twisted‐stacking functional films with user‐specified polarization‐responsive performances, achieving a robust mapping relationship among structural parameters, polarization states, and transmission characteristics. This work demonstrates a continuous manipulation of the spectra and the output color display by the arbitrary combinations of the input polarization channels and structure/process parameters, enabling accurate forward prediction of output displays and, consequently, allowing for the inverse design with a high information capacity (10 6 ) and information encryption. Moreover, by utilizing a reverse fully connected neural network for comparison and multi‐region joint determination, this work proposes a novel fully polarization detection platform, favoring the identification of arbitrary unknown incident light.