International Display Workshops General Incorporated Association

10:00 〜 10:20

[LCT9-4L] Machine Learning Model of Liquid Crystal Cell Controllable by Design Parameters

*Makoto Watanabe1,2, Reo Otsuki1, Kiyoshi Kotani1, Yasuhiko Jimbo1 (1. The University of Tokyo (Japan), 2. Silvaco Japan Co., Ltd. (Japan))

Machine learning, Reservoir computing, Liquid crystal display, Macro model, Design parameter

https://doi.org/10.36463/idw.2023.0101

Using one of the machine learning methods “Reservoir computing”, we developed a macro model of liquid crystal cells that follows changes in design parameters. Higher accuracy was obtained by setting the hyperparameters appropriately. This model can be trained very fast, making it suitable for further design space expansion.