Photo of Lei Deng

Artificial intelligence & robotics

Lei Deng

Main designer of Tianjic Chip, the world's first heterogeneous fusion brain-inspired chip

Year Honored
2019

Region
China

In August 2019, the Tianjic Chip achieved by Shi Luping's team from the Center for Brain Inspired Computing Research (CBICR) of Tsinghua University was published in Nature as a cover story. It was the first hybrid neuromorphic chip in the world, bridging computer science and neuroscience to develop artificial general intelligence. Its processing ability was successfully verified in the application scenario of autonomous bicycles, and it was also the first time that the research of the domestic team in the field of AI and chips appeared on the cover of Nature.

As the first author of this thesis, Deng Lei is the first Ph.D in the field of brain-inspired computing in China. It can be said that he is the first group of young scientists to see the future of brain-inspired computing.

Deng Lei studied in the Precision Instrument Department of the University of Science and Technology of China, focusing on robots and optoelectronic systems. After completing his undergraduate studies in 2012, the concept of "brain-inspired computing" in an academic environment aroused his interest. The lecture that moved him was given by Shi Luping, the tutor of his future doctorate.

At that time, "brain-inspired computing" did not even have related entries in domestic search engines. However, Deng Lei believes that it will be a combination of machine learning and neuromorphic computing technology that will make this field more essential and profound, which was a huge attraction to him: "I prefer to reveal the secrets of the general intelligence of the human brain rather than choosing one of two technical directions alone."

Therefore, when Shi Luping joined Tsinghua in 2013 and started to organize the CBICR, Deng Lei also became a member of the team and started his research in brain-inspired computing.

After the birth of Tianjic, the team also developed the first-generation brain-inspired computing software tool chain, which can support automatic mapping and compilation from the neural network programming platform to Tianjic. At present, the team has started research on the next-generation chip, which is expected to be completed early next year.

Recalling the constructing of the center from zero, Deng Lei said with a smile: "I feel quite lonely without the seniors." But with the follow-up research on Tianjic continuing in this direction, he also believes that more and more people will join this extremely challenging direction in the future, opening up more possibilities.