[24p-31A-1~14] FS.1 フォーカストセッション「AIエレクトロニクス」

16:45 〜 17:00英語発表

[24p-31A-14] Analysis of Read Current Fluctuation in Low Resistance State ReRAM by using Fluctuation Pattern Classifier

〇(M2)Zhiyuan Huang¹, Ayumu Yamada¹, Naoko Misawa¹, Chihiro Matsui¹, Ken Takeuchi¹ (1.Univ. Tokyo)

キーワード：

Computation-in-Memory、Neural network、ReRAM

ReRAM is a type of non-volatile memory which is suitable for Computation-in-Memory (CiM)because of its high speed, scalability, and power efficiency. The fluctuation in the ReRAM's readcurrent influences the performance of ReRAM and ReRAM-based CiM. In this work, aconvolutional neural network-based fluctuation pattern classifier trained by synthesized databaseis used to detect different types of low resistance state ReRAM read current. Fluctuation dataunder different measurement conditions is analyzed.