Work

Synstor circuits for real-time adaptive intelligent systems

Year: 2025

Type: article

Abstract: The foundation of current AI systems lies in computers that implement learning and inference procedures sequentially, resulting in learning latency, elevated power consumption, and diminished adapbility to unforeseen environmental changes. We overcome these issues by presenting a brain-inspired synaptic resistor (synstor) circuit, enabling concurrent real-time learning and inference. AI systems utilizing synstor circuits, such as those in drones and morphing wings, achieve significantly faster learning, better performance, greater energy efficiency, and improved adaptability compared to computer-based AI systems. Synstor circuits offer a promising path for AI systems to excel in dynamic and unpredictable real-world environments.

Source: Journal of Intelligent Material Systems and Structures

DOI: 10.1177/1045389x251349347

Author: Yong Chen

Institution: University of California, Los Angeles

Language: English

Topic: Advanced Memory and Neural Computing

Subfield: Electrical and Electronic Engineering

Field: Engineering

Domain: Physical Sciences

SDG: Industry, innovation and infrastructure

Open Access status: hybrid

Estimated APC paid: $4,150