Výsledky vyhledávání

Report

Gradient descent-based programming of analog in-memory computing cores

Autor: Büchel, Julian, Vasilopoulos, Athanasios, Kersting, Benedikt, Odermatt, Frederic, Brew, Kevin, Ok, Injo, Choi, Sam, Saraf, Iqbal, Chan, Victor, Philip, Timothy, Saulnier, Nicole, Narayanan, Vijay, Gallo, Manuel Le, Sebastian, Abu

Publikováno v: 2022 International Electron Devices Meeting (IEDM), San Francisco, CA, USA, 2022, pp. 33.1.1-33.1.4

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

Externí odkaz: http://arxiv.org/abs/2305.16647

Zobrazit plný text záznamu

Report

A 64-core mixed-signal in-memory compute chip based on phase-change memory for deep neural network inference

Publikováno v: Nature Electronics 6, 680-693 (2023)

The need to repeatedly shuttle around synaptic weight values from memory to processing units has been a key source of energy inefficiency associated with hardware implementation of artificial neural networks. Analog in-memory computing (AIMC) with sp

Externí odkaz: http://arxiv.org/abs/2212.02872

Zobrazit plný text záznamu

Report

In-memory Realization of In-situ Few-shot Continual Learning with a Dynamically Evolving Explicit Memory

Autor: Karunaratne, Geethan, Hersche, Michael, Langenegger, Jovin, Cherubini, Giovanni, Gallo-Bourdeau, Manuel Le, Egger, Urs, Brew, Kevin, Choi, Sam, OK, INJO, Silvestre, Mary Claire, Li, Ning, Saulnier, Nicole, Chan, Victor, Ahsan, Ishtiaq, Narayanan, Vijay, Benini, Luca, Sebastian, Abu, Rahimi, Abbas

Continually learning new classes from a few training examples without forgetting previous old classes demands a flexible architecture with an inevitably growing portion of storage, in which new examples and classes can be incrementally stored and eff

Externí odkaz: http://arxiv.org/abs/2207.06810

Zobrazit plný text záznamu