Recently, the 34th International Joint Conference on Artificial Intelligence (IJCAI) announced the acceptance results of the paper. The paper Wave-driven Graph Neural Networks with Energy Dynamics for Over-mood Mitigation written by Zhao Qin's research group from the School of Information and Computer Science was accepted by the conference. Wu Peihan, a master's student of SHNU, and Qi Hongda, a young teacher, were co-first authors, Associate Professor Zhao Qin was the corresponding author, and Shanghai Normal University was the sole signing unit.
A wave-driven graph neural network framework has been proposed, which redefines feature propagation through wave equations. Unlike diffusion, the wave equation combines second-order dynamics, balancing smoothing and oscillatory behavior to preserve high-frequency components while ensuring effective information flow. To enhance the stability and convergence of the discretization of the wave equation in the graph structure, an energy mechanism inspired by kinetic and potential energy dynamics was introduced to stabilize propagation by balancing time evolution and structural alignment.
IJCAI is the oldest academic conference in the field of artificial intelligence and one of the three flagship conferences in the international field of artificial intelligence. It is known as the Olympics of the artificial intelligence world and represents the highest level and development direction of international artificial intelligence research. Its paper acceptance rate is the lowest among top computer conferences, only 19.3% this year. This is the first time that SHNU has published a paper on IJCAI.
