[14a-K101-1~10] 13.5 Semiconductor devices/ Interconnect/ Integration technologies

12:00 PM - 12:15 PM JST(3:00 AM - 3:15 AM UTC)

[14a-K101-10] Trial for Easing Cryogenic Model Parameter Extraction utilizing Generative AI

〇Takumi Inaba¹, Yusuke Chiashi¹, Hiroshi Oka¹, Minoru Ogura¹, Hidehiro Asai¹, Hiroshi Fuketa¹, Shota Iizuka¹, Kimihiko Kato¹, Shunsuke Shitakata¹, Takahiro Mori¹ (1.AIST)

Keywords:

Deep Learning,Generative AI,Cryo-CMOS

We have previously reported on the acquisition of a large volume of I-V characteristics using a cryogenic 300mm wafer prober and the development of a cryogenic foundation model. This cryogenic foundation model enables transfer learning with a limited set of cryogenic I-V characteristics related to the target device, thereby predicting the unmeasured cryogenic I-V characteristics necessary for circuit design, including extrapolated data. In this presentation, we will discuss the development of a generative AI for extracting model parameters from the predicted cryogenic I-V characteristics, which we are undertaking as a continuation of our previous work.