Behavioral Level Simulation Framework to Support Error-Aware CNN Training with In-Memory Computing

Shih Han Chang, Chien Nan Jimmy Liu, Alexandra Kuster

研究成果: Conference contribution同行評審

4 引文 斯高帕斯(Scopus)

摘要

In recent years, in-memory computing (IMC) is a promising technique to solve the bottleneck of data movement in edge AI devices. To perform some simple computation in memory, the IMC designs often adopt analog operations, which may incur inevitable computation errors. To recover the accuracy loss, adding some random disturbance in CNN training is a straightforward approach to make it more tolerant to computation errors. However, random values may be quite different to the real run-time errors caused by non-ideal effects. In this work, we propose a hierarchical simulation framework for the IMC systems to support error-aware CNN training. This framework includes an efficient approach to build accurate IMC behavioral models that reflect real non-ideal effects. By using the behavioral model, an accurate high-level error model can be built efficiently to provide run-time errors for CNN training and error rate verification. As shown in the simulation results, the error-aware CNN training with the proposed models efficiently improves the CNN accuracy in real applications with almost no accuracy loss.

原文English
主出版物標題Proceedings - 2022 18th International Conference on Synthesis, Modeling, Analysis and Simulation Methods, and Applications to Circuit Design, SMACD 2022
發行者Institute of Electrical and Electronics Engineers Inc.
ISBN(電子)9781665467032
DOIs
出版狀態Published - 2022
事件18th International Conference on Synthesis, Modeling, Analysis and Simulation Methods, and Applications to Circuit Design, SMACD 2022 - Villasimius, Italy
持續時間: 12 6月 202215 6月 2022

出版系列

名字Proceedings - 2022 18th International Conference on Synthesis, Modeling, Analysis and Simulation Methods, and Applications to Circuit Design, SMACD 2022

Conference

Conference18th International Conference on Synthesis, Modeling, Analysis and Simulation Methods, and Applications to Circuit Design, SMACD 2022
國家/地區Italy
城市Villasimius
期間12/06/2215/06/22

指紋

深入研究「Behavioral Level Simulation Framework to Support Error-Aware CNN Training with In-Memory Computing」主題。共同形成了獨特的指紋。

引用此