Related Experiment Video
Updated: Sep 15, 2025
10:16
Optical Control of Living Cells Electrical Activity by Conjugated Polymers
Published on: January 28, 2016
7.7K
Electrically programmable organic in-display neuromorphic computing
Shilei Dai1, Dingchen Wang1,2, Xu Liu3
1Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong 999077, China.
National Science Review
|July 16, 2025
Summary
This study introduces an electrically programmable in-display neuromorphic computing (EP-IDNC) device using organic semiconductors. This innovation integrates computing and display functions, addressing energy and hardware challenges in smart edge devices.
Area of Science:
- Materials Science
- Computer Engineering
- Artificial Intelligence
Background:
- Smart edge devices face limitations in energy consumption, latency, and hardware costs due to separated memory, processing, and display units.
- Scaling AI models and data exacerbates these challenges for devices like smartphones and AR/VR glasses.
Purpose of the Study:
- To develop an integrated solution for edge AI by collocating computing and display functions.
- To overcome the limitations of current edge devices through in-display neuromorphic computing (IDNC).
Main Methods:
- Development of an all-in-one electrically programmable IDNC (EP-IDNC) device.
- Leveraging the electrochromism of organic semiconductors for device functionality.
- Utilizing a prototype 3x3 EP-IDNC device array for validation.
Main Results:
- The EP-IDNC device exhibits synaptic and neural behaviors, including multi-terminal operability and multi-level weight updates.
- The device visually displays computing results via color changes, enabled by electrochromism.
- Successful validation on tasks like noise reduction and motion object perception.
Conclusions:
- The developed EP-IDNC device successfully integrates neuromorphic computing and display capabilities.
- This technology offers a promising approach to enhance the efficiency and performance of smart edge devices.
- Demonstrated potential applications include advanced driver-assistance systems, such as car steering reminders.
