摘要
基于第216期"双清论坛"关于人工智能芯片发展的主题报告和分组讨论,本文着重介绍了面向神经形态计算的智能芯片与器件技术的国内外主要研究进展及存在的挑战,包括人工神经网络加速器、基于传统CMOS的神经形态智能芯片、新型神经形态器件技术及基于新器件的神经形态芯片等内容,探讨了神经形态智能芯片面临的关键问题和未来发展趋势。
Recapitulating from the discussion upon development of artificial intelligent chip in the 216 th Shuangqing Forum,the recent advances and scientific challenges in intelligent chipand device technology for neuromorphic computing are analyzed in this paper,including artificial neural network accelerator,neuromorphic chip based on conventional CMOS technology,emerging neuromorphic devices and neuromorphic c hips.The key challenges and future development trends for neuromorphic chip are outlined.
作者
王宗巍
杨玉超
蔡一茂
朱涛
丛杨
王志衡
黄如
Wang Zongwei;Yang Yuchao;Cai Yimao;Zhu Tao;Cong Yang;Wang Zhiheng;Huang Ru(Institute of Microelectronics»Peking University,Beijing 100871;Academy for Artificial Intelligence,Peking University,Beijing 100871;College of Optoelectronic Engineering,Chongqing University,Chongqing 400044;Shenyang Institute of Automation y Chinese Academy of Science,Shenyang 110016;Department of Information Sciences,National Natural Science Foundation of China,Beijing 100085)
出处
《中国科学基金》
CSCD
北大核心
2019年第6期656-662,共7页
Bulletin of National Natural Science Foundation of China
关键词
人工智能
类脑计算
神经形态芯片
神经形态器件
artificial intelligence
brain-inspired computing
neuromorphic chip
neuromorphicdevice
