应我校邀请，Kai Ming Ting教授(FederationUniversity Australia)将于10月11日来我校做专题学术讲座，欢迎全校师生参加！报告的具体安排如下：
报告主题：Classification UnderStreaming Emerging New Classes: A Solution Using Completely-Random Trees
报告人：Kai Ming Ting 博士、教授
内容提要:This talk reports aninvestigation on an important problem in stream mining, i.e., classificationunder streaming emerging new classes or SENC. The SENC problem can bedecomposed into three subproblems: detecting emerging new classes, classifyingknown classes, and updating models to integrate each new class as part of knownclasses. The common approach is to treat it as a classification problem andsolve it using either a supervised learner or a semi-supervised learner. Wepropose an alternative approach by using unsupervised learning as the basis tosolve this problem. The proposed method employs completely-random trees whichhave been shown to work well in unsupervised learning and supervised learningindependently in the literature. The completely-random trees are used as asingle common core to solve all three subproblems in SENC: unsupervisedlearning, supervised learning, and model update on data streams. We show thatthe proposed unsupervised-learning-focused method often achieves significantlybetter outcomes than existing classification-focused methods.
报告人简介：After receiving his PhD from the University of Sydney,Kai Ming Ting had worked at the University of Waikato, Deakin University andMonash University. He joins Federation University Australia since 2014. He hadpreviously held visiting positions at Osaka University, Nanjing University, andChinese University of Hong Kong. His current research interests are in theareas of mass estimation, mass-based dissimilarity, anomaly detection, ensembleapproaches, data streams, data mining and machine learning in general. He hasserved as a program committee co-chair for the Twelfth Pacific-Asia Conferenceon Knowledge Discovery and Data Mining (PAKDD-2008). He was a member of theprogram committee for a number of international conferences including ACMSIGKDD International Conference on Knowledge Discovery and Data Mining, andInternational Conference on Machine Learning. He has received research fundingfrom Australian Research Council, US Air Force of Scientific Research(AFOSR/AOARD), Toyota InfoTechnology Center, and Australian Institute ofSports. Awards received include the Runner-up Best Paper Award in 2008 IEEEICDM (for Isolation Forest), and the Best Paper Award in 2006 PAKDD. He is thecreator of isolation techniques, mass estimation and mass-based dissimilarity.