Scaling-up Memristor Monte Carlo with magnetic domain-wall physics - Département Systèmes et Circuits Intégrés Numériques Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Scaling-up Memristor Monte Carlo with magnetic domain-wall physics

Résumé

By exploiting the intrinsic random nature of nanoscale devices, Memristor Monte Carlo (MMC) is a promising enabler of edge learning systems. However, due to multiple algorithmic and device-level limitations, existing demonstrations have been restricted to very small neural network models and datasets. We discuss these limitations, and describe how they can be overcome, by mapping the stochastic gradient Langevin dynamics (SGLD) algorithm onto the physics of magnetic domain-wall Memristors to scale-up MMC models by five orders of magnitude. We propose the push-pull pulse programming method that realises SGLD inphysics, and use it to train a domain-wall based ResNet18 on the CIFAR-10 dataset. On this task, we observe no performance degradation relative to a floating point model down to an update precision of between 6 and 7-bits, indicating we have made a step towards a large-scale edge learning system leveraging noisy analogue devices.

Mots clés

Fichier principal
Vignette du fichier
2312.02771v1.pdf (884.21 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04590374 , version 1 (28-05-2024)

Identifiants

  • HAL Id : hal-04590374 , version 1

Citer

Thomas Dalgaty, Shogo Yamada, Anca Molnos, Eiji Kawasaki, Thomas Mesquida, et al.. Scaling-up Memristor Monte Carlo with magnetic domain-wall physics. MLNCP2023 - 37th NeurIPS Machine Learning with New Compute Paradigms workshop, Dec 2023, New Orleans (LA), United States. ⟨hal-04590374⟩
0 Consultations
0 Téléchargements

Partager

Gmail Mastodon Facebook X LinkedIn More