神经形态工程学
冯·诺依曼建筑
瓶颈
材料科学
油藏计算
人工神经网络
人工智能
计算机体系结构
计算机科学
纳米技术
氮化镓
工程类
记忆电阻器
突触重量
突触
变质塑性
神经科学
电子工程
作者
Fangliang Gao,Lihan Li,Shuti Li
出处
期刊:eScience
[Elsevier BV]
日期:2025-12-01
卷期号:: 100519-100519
标识
DOI:10.1016/j.esci.2025.100519
摘要
Artificial synaptic devices that can emulate the functions of biological synapses and offer the advantages of brain-like computation are garnering considerable interest, as they have the potential to innovatively circumvent the von Neumann bottleneck at the device level. The foundation of research in artificial synaptic devices is the development of functional materials facilitating multilevel conductance state switching and storage. Gallium nitride (GaN), a third-generation semiconductor material with exceptional properties, has been used to implement artificial synaptic devices for neuromorphic computing. GaN-based artificial synapses have been demonstrated to enable a wide range of neuromorphic characteristics, exhibiting high photovoltaic efficiency, stability, low power consumption, and UV responsiveness. The intrinsic properties of GaN itself and in combination with other materials offer many possibilities for emulating synaptic function, which gives GaN materials the prospect of becoming the cornerstone of neuromorphic computing systems. Herein, to more fully comprehend the potential of GaN-based artificial synaptic devices in future neuromorphic systems, this paper reviews recent advancements in GaN synthesis and GaN-based artificial synapses driven by different working mechanisms. It also analyzes the prospective applications of GaN-based artificial synapses and the challenges likely to be encountered in their subsequent evolution. This review may be a significant turning point for developing neuromorphic computing using GaN materials. • This review highlights the recent advancements in GaN-based synaptic devices. • Innovative approaches to GaN synthesis are summarized. • GaN-based synaptic devices operating on different mechanisms are thoroughly discussed. • GaN-based neuromorphic applications are presented.
科研通智能强力驱动
Strongly Powered by AbleSci AI