Deep learning is an important machine learning method, which is widely deployed in both cloud end and mobile end. We developed the world's first deep learning processor chip Cambricon. Currently, Cambricon chips and IPs have been widely deployed in tens of millions of cellphone and servers. In this talk, we will disucss the technical challenges of deep learning processor and the corresponding solutions.
Yunji Chen was born in Nanchang, China, in 1983. He graduated from Special Class for Gifted Young, University of Science and Technology of China in 2002, and got his ph.d degree from Institute of Computing Technology (ICT), Chinese Academy of Sciences in 2007, both in Computer Science. Currently, he is a full professor at ICT. He leads the Intelligent Processor Research Center of ICT to develop deep learning processors. Yunji Chen has authored or coauthored 1 book and over 100 papers on various venues including ISCA, HPCA, MICRO, ASPLOS, ISSCC, Hot Chips, IJCAI, and AAAI. He served as the general co-chair of ASPLOS'17, and TPC members of DAC, ISCA, HPCA, MICRO, and ASPLOS. He was listed in MIT Global TR-35 (2015) for his contribution in deep learning processor. He was reported as "leader" and "poineer" of deep learning processor by the Science magazine.