报告题目：Asymptotic minimax risk of stochastic block model for community extraction
报 告 人：刘秉辉 教授
所在单位：东北师范大学
报告时间：2022年8月17日 星期三 14:00-15:00
报告地点：腾讯会议 ID：726-117-696
校内联系人：程建华 chengjh@jlu.edu.cn
报告摘要：Most existing community discovery methods focus on partitioning all nodes of the network into communities. However, many real networks contain background nodes that do not belong to any communities. In such a situation, typical methods tend to artificially split the background nodes and group them together with communities with relatively stronger connection, hence lead to distorted results. To avoid this, some community extraction methods have been developed to achieve community discovery with background nodes. There are two limitations of the existing methods for community extraction: first, they have difficulties in handling large-scale networks due to high computational complexity; second, some theoretical results, such as the minimax risk, of these community extraction models need to be further investigated. Motivated by the situation, we establish the asymptotic minimax risk of the stochastic block model for community extraction. We also propose a refinement algorithm with polynomial complexity to achieve fast computation for community extraction. Further, we illustrate that the community structure estimated by the proposed algorithm obtain the established asymptotic minimax risk. We illustrate the advantages and feasibility of the proposed algorithm via extensive simulated networks and a political blog network.
报告人简介：刘秉辉，东北师范大学，教授、博导，统计系主任；研究方向为应用统计、机器学习和网络数据分析；在统计学、计算机&人工智能、计量经济学领域期刊发表SCI论文二十余篇，部分成果发表在：统计学顶级期刊Journal of the American Statistical Association、Annals of Statistics，计算机&人工智能顶级期刊Artificial Intelligence、Journal of Machine Learning Research，计量经济学顶级期刊Journal of Econometrics；以及领域重要期刊Journal of Business & Economic Statistics、Annals of Applied Statistics等；入选国家天元数学东北中心优秀青年学者奖励计划、吉林省拔尖创新人才；主持国家自然科学基金面上项目2项，参与国家自然科学基金重点项目、科技部重点研发计划项目。
