Tsinghua University News

Tsinghua University hosted Nature Conference on Neuromorphic Computing

Tsinghua University hosted Nature Conference on Neuromorphic Computing


On October 28th, the Nature Conference on neuromorphic computing hosted by Tsinghua University and Beijing Innovation Center for Future Chips was opened in Beijing. Nearly 40 top scientists from world-renowned universities, research institutes and enterprises were invited to participate in the conference. There were also more than 400 professional participants who attended the conference. The three-day conference had 38 keynote lectures, three high-end mini-courses, one group discussion, and more than 50 posters. In addition, achievements on brain-inspired computing were exhibited during the conference to showcase vividly the useful applications of a brain-inspired computing system to participants.

It was pointed out in the opening speech by You Zheng, an academic from the Chinese Academy of Engineering and Vice President of Tsinghua University, that neuromorphic computing is a typical interdisciplinary field covering a number of disciplines such as materials science, physics, electronic engineering, computer science, and neuroscience. The aims of this international conference were to build an interdisciplinary open communication platform for researchers in the field of neuromorphic computing, and to stimulate the spark of innovation and promote cooperation in this field.

The themes of reports on the conference included information-processing mechanisms and the strategic advantages of neuromorphic computing, new materials and device performance control for neuromorphic computing, emerging neuromorphic computing devices, the array and architecture of memristor, a circuit system for neuromorphic computing, a system-on-chip for embedded artificial intelligence, architecture planning for neuromorphic computing, collaboration on algorithms and hardware, and other related topics,  while comparing the advantages and disadvantages of different implementation paths of neuromorphic computing. Opportunities and challenges to the development of neuromorphic computing were also suggested.

Compared with the traditional computing pattern based on von Neumann Architecture, neuromorphic computing has advantages such as low energy consumption and high computational efficiency. Being expected to support further breakthroughs in artificial intelligence technology, it has received positive attention from both academic and industrial circles.