File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Article: Structural plasticity‐based hydrogel optical Willshaw model for one‐shot on‐the‐fly edge learning

TitleStructural plasticity‐based hydrogel optical Willshaw model for one‐shot on‐the‐fly edge learning
Authors
Keywordsassociative memory
hydrogel
one-shot learning
optical neural network
structural plasticity
Willshaw model
Issue Date20-Jan-2023
PublisherWiley-Blackwell
Citation
InfoMat, 2023, v. 5, n. 4 How to Cite?
Abstract

Autonomous one-shot on-the-fly learning copes with the high privacy, small dataset, and in-stream data at the edge. Implementing such learning on digital hardware suffers from the well-known von-Neumann and scaling bottlenecks. The optical neural networks featuring large parallelism, low latency, and high efficiency offer a promising solution. However, ex-situ training of conventional optical networks, where optical path configuration and deep learning model optimization are separated, incurs hardware, energy and time overheads, and defeats the advantages in edge learning. Here, we introduced a bio-inspired material-algorithm co-design to construct a hydrogel-based optical Willshaw model (HOWM), manifesting Hebbian-rule-based structural plasticity for simultaneous optical path configuration and deep learning model optimization thanks to the underlying opto-chemical reactions. We first employed the HOWM as an all optical in-sensor AI processor for one-shot pattern classification, association and denoising. We then leveraged HOWM to function as a ternary content addressable memory (TCAM) of an optical memory augmented neural network (MANN) for one-shot learning the Omniglot dataset. The HOWM empowered one-shot on-the-fly edge learning leads to 1000× boost of energy efficiency and 10× boost of speed, which paves the way for the next-generation autonomous, efficient, and affordable smart edge systems.


Persistent Identifierhttp://hdl.handle.net/10722/338045
ISI Accession Number ID

 

DC FieldValueLanguage
dc.contributor.authorWang, Dingchen-
dc.contributor.authorLiu, Dingyao-
dc.contributor.authorLin, Yinan-
dc.contributor.authorYuan, Anran-
dc.contributor.authorZhang, Woyu-
dc.contributor.authorZhao, Yaping-
dc.contributor.authorWang, Shaocong-
dc.contributor.authorChen, Xi-
dc.contributor.authorChen, Hegan-
dc.contributor.authorZhang, Yi-
dc.contributor.authorJiang, Yang-
dc.contributor.authorShi, Shuhui-
dc.contributor.authorLoong, Kam Chi-
dc.contributor.authorChen, Jia-
dc.contributor.authorWei, Songrui-
dc.contributor.authorWang, Qing-
dc.contributor.authorYu, Hongyu-
dc.contributor.authorXu, Renjing-
dc.contributor.authorShang, Dashan-
dc.contributor.authorZhang, Han-
dc.contributor.authorZhang, Shiming-
dc.contributor.authorWang, Zhongrui-
dc.date.accessioned2024-03-11T10:25:50Z-
dc.date.available2024-03-11T10:25:50Z-
dc.date.issued2023-01-20-
dc.identifier.citationInfoMat, 2023, v. 5, n. 4-
dc.identifier.urihttp://hdl.handle.net/10722/338045-
dc.description.abstract<p>Autonomous one-shot on-the-fly learning copes with the high privacy, small dataset, and in-stream data at the edge. Implementing such learning on digital hardware suffers from the well-known von-Neumann and scaling bottlenecks. The optical neural networks featuring large parallelism, low latency, and high efficiency offer a promising solution. However, ex-situ training of conventional optical networks, where optical path configuration and deep learning model optimization are separated, incurs hardware, energy and time overheads, and defeats the advantages in edge learning. Here, we introduced a bio-inspired material-algorithm co-design to construct a hydrogel-based optical Willshaw model (HOWM), manifesting Hebbian-rule-based structural plasticity for simultaneous optical path configuration and deep learning model optimization thanks to the underlying opto-chemical reactions. We first employed the HOWM as an all optical in-sensor AI processor for one-shot pattern classification, association and denoising. We then leveraged HOWM to function as a ternary content addressable memory (TCAM) of an optical memory augmented neural network (MANN) for one-shot learning the Omniglot dataset. The HOWM empowered one-shot on-the-fly edge learning leads to 1000× boost of energy efficiency and 10× boost of speed, which paves the way for the next-generation autonomous, efficient, and affordable smart edge systems.<br></p>-
dc.languageeng-
dc.publisherWiley-Blackwell-
dc.relation.ispartofInfoMat-
dc.rightsThis work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.-
dc.subjectassociative memory-
dc.subjecthydrogel-
dc.subjectone-shot learning-
dc.subjectoptical neural network-
dc.subjectstructural plasticity-
dc.subjectWillshaw model-
dc.titleStructural plasticity‐based hydrogel optical Willshaw model for one‐shot on‐the‐fly edge learning-
dc.typeArticle-
dc.identifier.doi10.1002/inf2.12399-
dc.identifier.scopuseid_2-s2.0-85146992562-
dc.identifier.volume5-
dc.identifier.issue4-
dc.identifier.eissn2567-3165-
dc.identifier.isiWOS:000919827800001-
dc.identifier.issnl2567-3165-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats