神经形态工程学
变性(医学)
阈值电压
材料科学
降级(电信)
光电子学
肖特基二极管
理论(学习稳定性)
神经元变性
神经科学
电压
电子工程
计算机科学
电气工程
晶体管
工程类
生物
人工神经网络
人工智能
机器学习
病理
二极管
医学
疾病
作者
Ao Chen,Zhennan Lin,Guokun Ma,Rui Xiong,Qiming Liu,Hao Wang
标识
DOI:10.1109/led.2024.3375851
摘要
The neuronal degeneration severely disrupts the stability of neuromorphic systems. In this letter, it was experimentally demonstrated that the NbOx device degradation was induced by an unhappy environment, thus the leaky integration-and-fire (LIF) threshold neurons performed the significantly distinguishable frequency characteristics like the degeneration of biological neurons. Furthermore, the mechanism revealed that the threshold voltage increase was induced by broadening the band gap and improving the Schottky barrier height, originating from the reduction of oxygen vacancies. This work realized and solidly understood the neuronal degeneration, benefiting to design feasible strategies for high stability neuromorphic systems.
科研通智能强力驱动
Strongly Powered by AbleSci AI