Evolutionary Approximation of Ternary Neurons for On-sensor Printed Neural Networks

article in a collection out of WoS and Scopus

English

Printed electronics offer ultra-low manufacturing costs and the potential for on-demand fabrication of flexible hardware. However, significant intrinsic constraints stemming from their large feature sizes and low integration density pose design challenges that hinder their practicality. In this work, we conduct a holistic exploration of printed neural network accelerators, starting from the analog-to-digital interface---a major area and power sink for sensor processing applications---and extending to networks of ternary neurons and their implementation. We propose bespoke ternary neural networks using approximate popcount and popcount-compare units, developed through a multi-phase evolutionary optimization approach and interfaced with sensors via customizable analog-to-binary converters. Our evaluation results show that the presented designs outperform the state of the art, achieving at least 6x improvement in area and 19x in power. To our knowledge, they represent the first open-source digital printed neural network classifiers capable of operating with existing printed energy harvesters.

Approximate Computing, Electrolyte-gated FET, Printed Electronics, Low-Power Classifiers, Ternary Neural Networks

MRÁZEK, V.; KOKKINIS, A.; PAPANIKOLAOU, P.; VAŠÍČEK, Z.; SIOZIOS, K.; TZIMPRAGOS, G.; TAHOORI, M.; ZERVAKIS, G.

28. 10. 2024

Association for Computing Machinery

New York

979-8-4007-1077-3

2024 IEEE/ACM International Conference on Computer Aided Design (ICCAD)

1

9

9