Preparation of GaSb-based Artificial Visual Synapses with Ultra-low Energy Consumption

Description

Rapid development of artificial intelligence techniques ignites the emerging demand on accurate perception and understanding of optical signals from external environments via brain-like visual systems. Taking inspiration from the human brain, artificial neuromorphic systems that adopt an in-memory computing approach can address the energy and throughput inefficiency of conventional von Neumann computing architectures such that the data processing and memorizing units are no longer physically separated. In particular, owing to the dynamic changes of synaptic connection strength, the artificial synapses are capable to process data and identify patterns more robust, plastic, and fault tolerant, giving rise to the adaptivity to indeterministic information. All these exceptional characteristics of the neuromorphic computing architecture have made it of great interest for brain-inspired technological applications such as visual information processing, which involves enormous parallel data that are correlated. Here, we plan to use CVD technology to grow high-quality GaSb materials, based on optimized growth condition and device structures, we will realize artificial visual synapse devices aiming at sensing, storing and processing various types of infrared information in one device. Then artificial neuromorphic imaging systems with ultra-low energy consumption will be designed and constructed.