【专家简介】：潘光明，新加坡南洋理工大学教授，博士生导师。2005年博士毕业于中国科学技术大学统计金融系；之后在新加坡国立大学、台湾中山大学、荷兰埃因霍温科技大学做博士后和学术交流工作；自2008年以来，在新加坡南洋理工大学工作；2013年遴选为国际统计学会会员(Elected Member of International Statistical Institute)。研究领域包括计量经济理论、高维统计、随机矩阵、多元统计等。主持新加坡国家基金项目5项，已在《Annals of Statistics》、《Journal of the Royal Statistical Society Series B》、《Journal of the American Statistical Association》、《Annals of Probability》、《Annals of Applied Probability》、《Bernoulli》、《IEEE Transactions on Signal Processing》、《IEEE Transactions on Information Theory》等顶级统计学杂志上发表60余篇学术论文，担任《Random Matrices: Theory and Applications》杂志编委。
【报告摘要】：Kernel spectral clustering with a single bandwidth can be inadequate for data exhibiting multiple characteristic pairwise-distance scales, a problem particularly prevalent in the high-dimensional regime. We address this issue through a multi-kernel formulation that aggregates kernels with different bandwidths.We develop a rigorous theoretical analysis of the resulting method under a general high-dimensional, multi-scale mixture model with heterogeneous cluster centers and covariance geometries. Under suitable eigen-gap and cluster-separation conditions, we show that approximate K-means applied to the multi kernel spectral embedding achieves exact recovery with high probability.
【报告时间】：2026年07月10日（周五）10:30-11:30
【报告地点】：崇真楼110
