Work

Edge detection and graph coloring by network of oscillatory neuromorphic devices comprising silicon nanosheet-gated diodes

Year: 2024

Type: article

Abstract: Oscillatory neural networks (ONNs) can handle image processing and combinational optimization problems with high biological plausibility. Coupled oscillators in ONNs mimic the neuronal oscillations and communication in the human brain. In this study, we demonstrate edge detection and graph coloring using a network of oscillatory neuromorphic devices comprising silicon nanosheet (NS)-gated diodes and series resistors. Silicon NS-gated diodes modulate the energy band structure in the channel through electrostatic doping, which tunes the amplitude and frequency of the oscillatory neuromorphic devices. Coupled oscillatory neuromorphic devices with capacitors can be in- or out-of-phase, enabling edge detection. Horizontal and vertical edge detections were performed using the coupling configuration. Furthermore, an ONN using oscillatory neuromorphic devices with resistive and capacitive couplings can solve graph coloring problems by representing the coloring solutions as phase differences. This study provides a method for achieving a compact silicon-based ONN that acts as a solver machine, as well as an image processor.

Source: Nano Research

DOI: 10.26599/nr.2025.94907214

Authors: Yunwoo Shin, Hyojoo Heo, Kyoungah Cho, Sangsig Kim

Institutions: Korea University, Korea University

Language: English

Topic: Advanced Memory and Neural Computing

Subfield: Electrical and Electronic Engineering

Field: Engineering

Domain: Physical Sciences

Open Access status: hybrid

Estimated APC paid: $5,290