Zhang, Ellina, Cai, Jack, Amirsoleimani, Amirali, Rahimi Azghadi, Mostafa, Genov, Roman, and Ahmadi, Majid (2024) Manhattan Rule for Robust In-Situ Training of Memristive Deep Neural Network Accelerators. In: Proceedings of the Midwest Symposium on Circuits and Systems. pp. 1324-1328. From: MWSCAS 2024: IEEE 67th International Midwest Symposium on Circuits and Systems, 11-14 August 2024, Springfield, MA, USA.

PDF (Published Version) - Published Version Restricted to Repository staff only

1

Abstract

In this work, we propose an alternative training approach for memristive circuits - the Manhattan rule training - which utilizes only sign information for weight updates. We present an in-depth analysis in both in-situ and ex-situ settings and show that not only does our method simplify circuit design but it also improves neural network robustness against device non-idealities. Using the MemTorch and our custom in-situ training framework, we implemented the Manhattan rule for MNIST classification and ECG signal detection tasks and achieved close to state-of-the-art performance under noise. Our work also provides a thorough comparison of Manhattan and conventional training methods under the effects of various device non-idealities, giving a crucial benchmark useful for the design of biomedical neural circuits.

Actions (Repository Staff Only)

Item Control Page