Dynamic resistive switching of WOx-based memristor for associative learning activities, on-receptor, and reservoir computing

Citations

WEB OF SCIENCE

7
Citations

SCOPUS

6

초록

The rapid expansion of data driven by the fourth industrial revolution has revealed significant limitations in conventional computing architectures, particularly in their ability to efficiently process vast amounts of data. Neuromorphic computing, which draws inspiration from the brain's parallel processing capabilities and efficiency, presents a promising solution to overcome these limitations. This study introduces a TiN/WOx/Pt memory device capable of emulating both nociceptive and synaptic behaviors, highlighting its potential for neuromorphic computing applications. The device successfully replicates key nociceptive functions, including threshold response, allodynia, and hyperalgesia, through the migration of oxygen ions and vacancies within the interface. Furthermore, it demonstrates a range of synaptic plasticity behaviors, such as spike-number-dependent plasticity, spike-amplitude-dependent plasticity, spike-rate-dependent plasticity, and paired-pulse facilitation. In addition, the device achieves 4-bit multibit reservoir computing with high accuracy, showcasing its ability to perform adaptive learning and nonlinear data processing. These results underline the TiN/WOx/Pt memory device's promise for mimicking biological functions and its significant potential in the development of advanced neuromorphic computing systems.

키워드

Dynamic memristorAssociative learningOn-receptor computingReservoir computingMEMORYNOCICEPTORSBEHAVIOROXIDE
제목
Dynamic resistive switching of WOx-based memristor for associative learning activities, on-receptor, and reservoir computing
저자
Noh, MinseoPark, HyogeunKim, Sungjun
DOI
10.1016/j.chaos.2025.116381
발행일
2025-07
유형
Article
저널명
Chaos, Solitons and Fractals
196
페이지
1 ~ 10