Effect of Program Error in Memristive Neural Network With Weight Quantization

Recently, various memory devices have been actively studied as suitable candidates for synaptic devices, which are important memory and computing units in neuromorphic systems. One of the ways to manage these devices is off-chip training, where it is essential to transfer the pretrained weights accu...

Ausführliche Beschreibung

Gespeichert in:

Bibliographische Detailangaben
Veröffentlicht in:	IEEE transactions on electron devices 2022-06, Vol.69 (6), p.3151-3157
Hauptverfasser:	Kim, Tae-Hyeon, Kim, Sungjoon, Hong, Kyungho, Park, Jinwoo, Youn, Sangwook, Lee, Jong-Ho, Park, Byung-Gook, Kim, Hyungjin
Format:	Artikel
Sprache:	eng
Schlagworte:	Aluminum oxide Biological neural networks Convolutional neural networks Errors Memory devices Memristors Neural networks Neuromorphic computing off-chip training program error program time Quantization (signal) quantization aware training (QAT) Random access memory resistive random access memory (RRAM) Switches synaptic device System performance Training variation weight quantization weight transfer
Online-Zugang:	Volltext bestellen
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Schreiben Sie den ersten Kommentar!