[15p-M_123-1~8] FS.1 Focused Session "AI Electronics"

2:45 PM - 3:00 PM JST(5:45 AM - 6:00 AM UTC)

[15p-M_123-4] Application of Surrogate Models to Logic Tasks in Nanoscale Molecular Networks

Aoto Sasaki¹, Tomoki Misaka¹, Hiroshi Ohyama¹, Wilfred van der Wiel², 〇Takuya Matsumoto¹ (1.Univ. Osaka, 2.Univ. Twente)

Keywords:

Physical reservoir,Molecular network,Surrogate model

“Physical reservoir computing” is expected to be a low-power, high-efficiency information-processing paradigm. In this study, we fabricated nanoscale molecular network devices based on the conductive polymer PEDOT:PSS and aimed to realize information processing by exploiting their nonlinear electrical conduction characteristics. In particular, to address two major challenges in the practical implementation of physical reservoirs—high experimental cost and difficulty in device control—we developed a surrogate model based on deep learning. Using this model, we analyzed the conduction pathways (side-gate effects) arising between multiple input electrodes within the device and explored their application to logic operation tasks.