报告人：黄翔宇 博士后（清华大学）

时间：2025年08月29日 15:00-

地点：理科楼LA103

摘要：We consider $L^p$-energy minimization dynamics on finite connected graphs. Given initial opinion profile on each vertex, at each step, the dynamics choose uniformly random vertex, and update its value by minimizing the $L^p$-energy. The case $p = 2$ is an asynchronous version of the dynamics introduced by deGroot in 1974 as a model for non-Bayesian social learning. We are interested in the case for general $p>1$. One of the case $p=\infty$ is also called the Lipschitz learning. In this talk, I will show the number of steps to reduce oscillation of values in dynamics for $p>1$, spercifically for $p=\infty$.

邀请人：数学研究中心

欢迎广大师生积极参与！