On October 11, 2025, our institute held a seminar entitled Robustness Enhancement Methods for Spiking Neural Networks in Room 511, West Section of Nanyong Building, Suzhou Campus of Nanjing University. The seminar was delivered by Dr. Ding Jianhao from Peking University and hosted by Assistant Professor Zhang Shaoqun of our institute.
Dr. Ding Jianhao presented his latest research findings in the field of spiking neural network (SNN) robustness enhancement. His report elaborated on robust training approaches from three perspectives: input signals, membrane potential variables, and encoding timing. He highlighted that, compared to traditional Artificial Neural Networks (ANNs), the new generation of SNNs inherently exhibits stronger resistance to perturbations. Dr. Ding also emphasized that future SNN research should focus on domains where SNNs can surpass traditional ANNs by leveraging their inherent advantages in event-driven processing and temporal sequence modeling.
During the Q&A session, participating students and faculty actively raised questions on topics such as event cameras, reparameterization methods, and theoretical analyses of robustness. Dr. Ding Jianhao addressed each query in detail, fostering a lively atmosphere and generating positive feedback.
The seminar concluded successfully. It was attended by Associate Professor Fu Yuxiang from the School of Integrated Circuits and Associate Professor Li Wenbin from the School of Artificial Intelligence and Science.
