Key points are not available for this paper at this time.
High Resolution Image Download MS PowerPoint Slide Conventional digital computing follows the von Neumann architecture, where memory and processing units are physically separated, resulting in limited throughput caused by data transfer bottlenecks, and high power consumption. In contrast, the human brain integrates computation and memory at synapses, enabling highly parallel and energy-efficient processing with only ∼20 W. Neuromorphic computing aims to replicate this efficiency by merging storage and computation within hardware. Oxide semiconductors (OSs) have emerged as promising candidates for synaptic transistors due to their favorable electrical, optical, and structural characteristics, as well as compatibility with low-temperature and large-area processing. This review outlines the fundamental principles of biological synapses and synaptic plasticity metrics, examines recent developments in OS-based synaptic transistors, and discusses prospects and challenges in implementing OS synaptic devices for neuromorphic hardware and artificial intelligence applications.
Lee et al. (Tue,) studied this question.