[9a-B21-1~10] FS.1 Focused Session "AI Electronics"

9:30 AM - 9:45 AM JST(12:30 AM - 12:45 AM UTC)Presentation by Applicant for JSAP Young Scientists Presentation Award

[9a-B21-3] Redox Reactions and Effective Features in Cu-Doped Ionic Liquid Physical Reservoirs

〇Masato Yamanoue^1,2, Shima Hisashi², Nokami Toshiki³, Mochida Tomoyuki⁴, Naitoh Yasuhisa², Akinaga Hiroyuki², Yumeng Zheng¹, Kinoshita Kentaro¹ (1.Tokyo Univ. of Sci, 2.AIST, 3.Tottori Univ., 4.Kobe Univ.)

Keywords:

physical reservoir computing,Redox reaction,feature selection

This study aims to clarify the relationship between learning-effective features and redox reactions in Cu-doped ionic liquid physical reservoirs. By combining dimensionality reduction based on partial least squares regression with feature selection using Lasso regression, we evaluated the relationships among learning performance, data dimensionality, and individual redox reactions. The results show that most effective features are distributed near redox current peaks, and more than 80% of the performance is maintained even when the number of features is reduced by more than half.