Bayesian In-Memory Computing with Resistive Memories - Archive ouverte HAL
Communication Dans Un Congrès Année : 2023

Bayesian In-Memory Computing with Resistive Memories

A Pontlevy
  • Fonction : Auteur
J Droulez
  • Fonction : Auteur
R Laurent
  • Fonction : Auteur
D Querlioz

Résumé

This paper explores three approaches using resistive memory for Bayesian near-memory and in-memory computing, leveraging their inherent randomness. The strategies include Bayesian machines for efficient near-memory computing, Bayesian neural networks exploiting randomness of synapses, and Bayesian learning utilizing the Metropolis-Hastings Markov Chain Monte Carlo technique. These methods achieve accuracy competitive with conventional software methods and allow for the evaluation of decision uncertainty.
Fichier non déposé

Dates et versions

hal-04695832 , version 1 (12-09-2024)

Identifiants

Citer

C Turck, D Bonnet, K-E Harabi, T Dalgaty, T Ballet, et al.. Bayesian In-Memory Computing with Resistive Memories. 2023 International Electron Devices Meeting (IEDM), Dec 2023, San Francisco, United States. pp.1 - 4, ⟨10.1109/iedm45741.2023.10413773⟩. ⟨hal-04695832⟩
48 Consultations
0 Téléchargements

Altmetric

Partager

More