Visualized in-sensor computing

Visualized in-sensor computing

24 April 2024 | Yao Ni, Jiaqi Liu, Hong Han, Qianbo Yu, Lu Yang, Zhipeng Xu, Chengpeng Jiang, Lu Liu & Wentao Xu
This article introduces an electrochromic neuromorphic transistor (ENT) designed for in-sensor computing, which uses color changes to represent synaptic weight. The ENT employs an ion-exchange membrane to regulate ion doping, enabling precise control over color-coded synaptic weight. This innovation allows for visualized pattern-recognition networks and bionic reflex systems, mimicking the color-changing behavior of organisms like the chameleon. The ENT integrates with an artificial whisker to simulate a bionic reflex system, enabling real-time visualization of signal flow in response to environmental stimuli. The device demonstrates rapid reset times and high-fidelity data processing, making it suitable for in-sensor computing tasks. The ENT also supports multimode signal coding, including the International Morse code, and can be used for visualized pattern-recognition networks. The study highlights the potential of the ENT in advancing bio-hybrid interfaces and biomimetic coding paradigms. The device's ability to adaptively regulate ion doping and its visualized synaptic weight changes offer new opportunities for enhancing artificial neural systems. The research provides a framework for developing more sophisticated and bioinspired computing systems.This article introduces an electrochromic neuromorphic transistor (ENT) designed for in-sensor computing, which uses color changes to represent synaptic weight. The ENT employs an ion-exchange membrane to regulate ion doping, enabling precise control over color-coded synaptic weight. This innovation allows for visualized pattern-recognition networks and bionic reflex systems, mimicking the color-changing behavior of organisms like the chameleon. The ENT integrates with an artificial whisker to simulate a bionic reflex system, enabling real-time visualization of signal flow in response to environmental stimuli. The device demonstrates rapid reset times and high-fidelity data processing, making it suitable for in-sensor computing tasks. The ENT also supports multimode signal coding, including the International Morse code, and can be used for visualized pattern-recognition networks. The study highlights the potential of the ENT in advancing bio-hybrid interfaces and biomimetic coding paradigms. The device's ability to adaptively regulate ion doping and its visualized synaptic weight changes offer new opportunities for enhancing artificial neural systems. The research provides a framework for developing more sophisticated and bioinspired computing systems.
Reach us at info@study.space
[slides] Visualized in-sensor computing | StudySpace