科研速览 · Science Skim继续刷下去 · Keep skimming →
◆ Small2026-04-09· Neuromorphic engineering

Bidirectional All‐Optical Synapses for Neuromorphic Computing and Vision

Zifan Li, Zicheng Zhang, Zhenhui Tao, Fa Zhang, Yuhan Zhang, Mengdan Dun, Guo Fang, Qiye Wang, Fei Xiu, Hai Li, Zhe Zhou, Wei Huang, Juqing Liu

原始摘要(英文原文)· Original abstract
All-optical synapses (AOS) that integrate sensing and processing of photonic signals represent a transformative frontier for optical computing and neuromorphic imaging. Most existing AOSs are limited to unidirectional weight modulation, achieving both excitation and inhibition output through all-optical pathway remains a challenge. Here we report a dual-photon-mediated strategy that achieving bidirectional neuroplasticity within a phosphorescent and photothermal carbon dot hybrid (CDH) neurotransmitter. This regulatory process employs ultraviolet and infrared light manipulation to achieve controllable exciton release and superposition of phosphorescent emission, thereby enabling neurophotonic potentiation and depression output. These features enable the CDH to realize optical logic operation and neurovision with real-time tracking and recognition of light objects, including motion trajectory, velocity, and direction, with a high accuracy of 97%. This work offers an all-optical regulation method to innovate phosphorescent paradigm for all-optical computing and imaging.
读原文 · Read the paper ↗

AI 追问PRO

登录后使用 AI 追问

讨论区

登录后参与讨论

相关论文 · Related

Bidirectional All‐Optical Synapses for Neuromorphic Computing and Vision — 科研速览 Science Skim