Learning Theory and Methods of Spiking Neural Networks

Assistant Professor Zhaofei Yu
Peking University, China

Spiking Neural Networks (SNNs) are considered as the third generation of artificial neural networks, which incorporate temporal dynamics in addition to neuron and synapse states. Compared to the previous two generations of artificial neural networks, SNNs possess features such as high biological plausibility, low power consumption, and efficiency. This report aims to introduce the basic principles of spiking neural networks, analyze the current research status and development trends, and present the latest advancements in the study of learning theory and methods of spiking neural networks, based on the recent research work conducted by our team.

Biosketch

Zhaofei Yu is currently an Assistant Professor at Peking University, where he leads the Spike Vision Lab. He received the PhD degree from Tsinghua University in 2017. His research interests are neuromorphic computation, computational neuroscience, and computer vision. He has published more than 60 papers in Nature Biomedical Engineering, Science Advance, Cell Patterns, PloS Computational Biology, IEEE Transactions on Pattern Analysis and Machine Intelligence and NeurIPS/ICML/ICLR/CVPR/ICCV/ECCV/AAAI/IJCAI. He has served as an editor of Frontiers in Neuroscience and an area chair of ICML/ACMMM.