Related Experiment Video
Updated: Jun 30, 2026
08:07
Assembly and Characterization of Biomolecular Memristors Consisting of Ion Channel-doped Lipid Membranes
Published on: March 9, 2019
Programmable optoelectronic memristors for energy-efficient adaptive binarized spiking neural networks
Ziyan Zhang1, Jiandong Yan2, Jiahao Gu1
1School of Integrated Circuits Industry, Wang Zheng School of Microelectronics, Changzhou University, Changzhou, Jiangsu 213164, P. R. China. guohuafei@cczu.edu.cn.
Nanoscale
|June 29, 2026
Summary
This study presents a novel optoelectronic memristive platform for brain-inspired vision. The new device enables energy-efficient, self-powered in-sensor processing for neuromorphic computing applications.
Area of Science:
- Materials Science
- Neuroscience
- Computer Engineering
Background:
- Neuromorphic devices aim to mimic brain functions for vision tasks.
- Current limitations include unstable ion dynamics and learning rate sensitivity.
Purpose of the Study:
- To develop an energy-efficient optoelectronic memristive platform for brain-inspired vision.
- To overcome limitations in current neuromorphic device technology.
Main Methods:
- Fabrication of Ag/Sb2Se3/SnO2/FTO memristive devices.
- Characterization of resistive switching, photoresponse, and device stability.
- Integration with a variable-learning-rate binarized spiking neural network.
Main Results:
- Stable bipolar resistive switching with high on/off ratio and retention.
- Fast, polarity-dependent photoresponses enabling in-sensor preprocessing.
- Achieved 89.8% accuracy on Fashion-MNIST with significantly reduced inference energy.
Conclusions:
- The developed platform offers a promising solution for energy-efficient neuromorphic computing.
- In-sensor processing capabilities enhance efficiency and noise suppression.
- This work advances hardware realization of brain-inspired vision systems.
