神经形态工程学
冯·诺依曼建筑
桥接(联网)
计算机科学
生物相容性材料
记忆电阻器
材料科学
计算机体系结构
兴奋性突触后电位
神经科学
人工神经网络
内容寻址存储器
纳米技术
人工智能
高效能源利用
钥匙(锁)
抑制性突触后电位
极限(数学)
能量(信号处理)
突触可塑性
仿生学
建筑
深层神经网络
逻辑门
作者
Hao Chen,Dongyue Huang,Zhiruo Zhang,Song Wang,Yinghui Han,Hui Huang
标识
DOI:10.1002/adom.202503737
摘要
Abstract Conventional vision sensors based on the von Neumann architecture suffer from high latency, energy inefficiency, and redundant data transfer, which limit their ability to meet the growing demands of artificial visual perception. Inspired by the human retina, where excitatory and inhibitory signaling pathways cooperate to enhance contrast, suppress noise, and adapt to dynamic environments, all‐optical modulated bidirectional responsive synaptic (AOMBRS) devices have emerged as a promising solution. In this review, recent advances in AOMBRS devices, covering diverse material platforms and their underlying mechanisms, are summarized. Further, emerging applications in optical logic operations, associative learning, neuromorphic computing, and adaptive vision, which demonstrate the promise of AOMBRS devices in bridging biological vision and artificial intelligence, are highlighted. Finally, the critical challenges related to material stability, device performance, and large‐scale integration are discussed, and future opportunities for advancing next‐generation neuromorphic hardware are outlined.
科研通智能强力驱动
Strongly Powered by AbleSci AI