神经形态工程学
冯·诺依曼建筑
瓶颈
计算机科学
计算机体系结构
人工神经网络
带宽(计算)
功率消耗
材料科学
电子工程
人工智能
纳米技术
功率(物理)
工程类
电信
嵌入式系统
物理
量子力学
操作系统
出处
期刊:eScience
[Elsevier]
日期:2023-08-31
卷期号:3 (6): 100178-100178
被引量:16
标识
DOI:10.1016/j.esci.2023.100178
摘要
Optoelectronic artificial synapses (OEASs) are essential for realizing artificial neural networks (ANNs) in next-generation information processing that has high transmission speed, high bandwidth, and low power consumption. Two-dimensional (2D) materials endowed with strong light-matter interactions and atomically thin dangling-bond-free surfaces are candidates for achieving versatile optoelectronics. Developing 2D OEASs for future neuromorphic applications is significant to break the bottleneck of von Neumann architecture and achieve future artificial intelligence systems. This review primarily focuses on recent developments in advanced 2D OEASs, discussing their working mechanism as well as potential applications. Common materials, device structures, and their synthesis and construction methods are also summarized. Finally, the prospects for future 2D OEASs from the perspectives of materials, performance, and applications are briefly described.
科研通智能强力驱动
Strongly Powered by AbleSci AI