神经形态工程学
晶体管
材料科学
光电子学
计算机科学
计算机体系结构
电气工程
人工神经网络
人工智能
工程类
电压
作者
Xiao Liu,Liang Chu,Wensheng Yan,Xiaodong Pi
标识
DOI:10.1016/j.xcrp.2024.102079
摘要
SummaryNeuromorphic computing systems based on high-temperature-resistant synaptic devices have emerged as energy-efficient and intelligent strategies for harsh-environment applications, such as fire alarms, nuclear combustion monitoring, and planetary exploration. The synaptic devices can be constructed by two-terminal memristors or three-terminal field-effect transistors. Though the latter exhibits a relatively complex structure, they are advantageous for linear conductance switching, multi-terminal modulation, and emulation of versatile synaptic functionalities. Herein, we will summarize the recent progress of high-temperature-resistant synaptic transistors (HTRSTs). Firstly, an in-depth discussion is conducted regarding their working mechanisms, device structures, and temperature-dependent synaptic characteristics. Then, an overview of the active materials commonly employed in HTRSTs is provided. Several application scenarios of HTRSTs for neuromorphic computing are presented. Finally, a few perspectives and directions for the future development of HTRSTs are outlined.Graphical abstract
科研通智能强力驱动
Strongly Powered by AbleSci AI