New analogue stop-learning control module using astrocyte for neuromorphic learning

Almasi, Milad, Karimi, Gholamreza, Ranjbar, Mahnaz, and Rahimi Azghadi, Mostafa (2020) New analogue stop-learning control module using astrocyte for neuromorphic learning. IET Circuits, Devices & Systems, 14 (1). pp. 100-106.

[img] PDF (Published Version) - Published Version
Restricted to Repository staff only

View at Publisher Website: https://doi.org/10.1049/iet-cds.2019.029...
 
1


Abstract

Learning algorithms and devices are an essential part of neural networks and neuromorphic architectures. Astrocyte, as an important element in the learning of neural networks, is believed to play a key role in long-term synaptic plasticity and memory. In addition, recent experimental observations indicate that astrocytes are active elements in learning in complex networks. In this study, the authors propose a new analogue astrocyte circuit that consumes fewer numbers of transistors compared to its previous counterparts. The authors then use this astrocyte to design a novel analogue circuit to implement a stop-learning mechanism in a spike-based learning algorithm and present its neuromorphic very large scale integration (VLSI) simulations. Experimental results demonstrate that the designed circuit can precisely implement the learning mechanisms shown by previous studies implementing spike-based learning rules without astrocyte. The proposed circuit proposes the first analogue stop-learning control algorithm that uses astrocytes. It has been designed and simulated in Taiwan semiconductor manufacturing company (TSMC) 0.35 μm complementary metal-oxide semiconductor (CMOS) technology.

Item ID: 62457
Item Type: Article (Research - C1)
ISSN: 1751-859
Related URLs:
Copyright Information: © The Institution of Engineering and Technology 2019
Date Deposited: 15 Mar 2020 22:46
FoR Codes: 09 ENGINEERING > 0906 Electrical and Electronic Engineering > 090601 Circuits and Systems @ 100%
SEO Codes: 97 EXPANDING KNOWLEDGE > 970109 Expanding Knowledge in Engineering @ 50%
97 EXPANDING KNOWLEDGE > 970110 Expanding Knowledge in Technology @ 50%
Downloads: Total: 1
More Statistics

Actions (Repository Staff Only)

Item Control Page Item Control Page