Gradient descent-based programming of analog in-memory computing cores
Julian Büchel, Athanasios Vasilopoulos, Benedikt Kersting, Frédéric Odermatt, Kevin Brew, Kang Min Ok, S. Choi, Iqbal Saraf, V. Chan, Timothy M. Philip, Nicole Saulnier, Vijay Narayanan, Manuel Le Gallo, Abu Sebastian
Abstract
The precise programming of crossbar arrays of unit-cells is crucial for obtaining high matrix-vector-multiplication (MVM) accuracy in analog in-memory computing (AIMC) cores. We propose a radically different approach based on directly minimizing the MVM error using gradient descent with synthetic random input data. Our method significantly reduces the MVM error compared with conventional unit-cell by unit-cell iterative programming. It also eliminates the need for high-resolution analog-to-digital converters (ADCs) to read the small unit-cell conductance during programming. Our method improves the experimental inference accuracy of ResNet-9 implemented on two phase-change memory (PCM)-based AIMC cores by 1.26%.