报告题目：Metric learning with Lipschitz continuous functions
Metric learning enables classification algorithms to automatically learn a suitable distance metric from data, such that semantically similar instances are pulled together while dissimilar instances are pushed away. A learned metric can significantly improve the performance of distance-based classifiers (e.g. kNN). In this talk, I will briefly present some of our recent research efforts on metric learning with Lipschitz continuous functions, including the methodology, theoretical foundation and optimisation formulation of each work. A brief introduction to University College London (UCL) will also be given.
Dr Jing-Hao Xue received a BEng degree in telecommunication and information systems in 1993 and a DrEng degree in signal and information processing in 1998, both from Tsinghua University. He received an MSc degree in medical imaging and an MSc degree in statistics, both from Katholieke Universiteit Leuven in 2004, and a PhD degree in statistics from the University of Glasgow in 2008. He is an Associate Professor in the Department of Statistical Science at University College London and a Turing Fellow in the Alan Turing Institute. His research interests include statistical machine learning, high-dimensional data analysis, statistical pattern recognition and image analysis.