Colloquia / Seminars / Industry Talks

Similarity measures have been used in a wide range of applications, e.g., machine learning, data visualization, image retrieval, and pattern matching. However, computing similarity measures is very time-consuming for different applications, especially for handling a large-scale dataset. As an example, using the kernel density estimation (based on one type of similarity measures, called kernel functions) for heatmap generation is very time-consuming, which takes at least two trillion operations on a 19201080 screen. In this talk, I will first illustrate how to utilize the bound functions of similarity measures to improve the efficiency in different applications. Then, I will discuss some advanced and tight bound functions for some famous similarity measures, e.g., kernel functions, and earth mover’s distance. After that, I will mention how to utilize our techniques to efficiently solve the real-life problem, i.e., interactive visualization of COVID-19 cases. Lastly, I will also illustrate some possible research directions for the future work.

Dr. Tsz Nam Chan is currently a research associate in The University of Hong Kong (HKU). Before that, he received the PhD degree in computing and the BEng degree in electronic and information engineering from The Hong Kong Polytechnic University in 2019 and 2014 respectively. His research mainly focuses on the efficiency issues for kernel methods, similarity search and spatial data analysis, especially for big data settings. He has published several research papers in top-tier conferences and journals in both data engineering and data mining area, including ACM SIGMOD, IEEE ICDE and IEEE TKDE. He also serves as the program committee members and reviewers of several prestigious conferences and journals, including IJCAI, WISE, IEEE TKDE and IEEE TC.