On the convergence analysis of Yau-Yau algorithm
组织者
演讲者
孙泽钜
时间
2024年03月20日 14:30 至 15:00
地点
理科楼A-304
摘要
In this talk, we will discuss the convergence analysis of Yau-Yau algorithm from a probabilistic perspective. In its original form, the convergence analysis of Yau-Yau algorithm is based on the assumptions for particular trajectories with certain properties. Here, we would like to generalize it by showing that the convergence results also hold in the sense of mathematical expectation with assumptions on the regularity of the system coefficients only. This generalization result is also important for the analysis of the recently developed framework where the algorithm is conducted by deep learning.