Date: 30/07/2021, 10-11am, Friday
Discriminative Representation for Topological Data Analysis
Topological data analysis (TDA) extracts the topological features that are complementary to statistical quantities, which has been found in many applications in computer vision. In TDA, the persistence diagram (PD) is a favored tool for describing topological features, such as the shape of data. However, PD cannot be fitted into machine learning models, because PD is not in the vector space. How to extract discriminative features from PD remains largely unexplored. This presentation will introduce two recent advances: polynomial representation and Hilbert representation for PD, with the aim of extracting the discriminative topological features. Finally, potential applications in the field of computer vision, and biomedical science will be discussed.
Qian Li is a Postdoc Research Fellow at the School of Computer Science, University of Technology Sydney (UTS). She received her Ph.D from the Chinese Academy of Science, and Master by Research degree from the University of Luxembourg.
Her general research interests lie primarily in causal machine learning, topological data analysis, and optimization algorithms. Her recent interests lie in causal machine learning and exploring causal reasoning insights to tackle challenging problems in machine learning, such as robustness and explainability. Besides, her research focuses on utilizing advanced mathematical theory (such as algebraic geometry) to alleviate the challenges of separability and stability in topological data analysis. Her papers have been published in the top-tier (CORE A*/A) conferences and journals in the field of machine learning and computer vision such as CVPR, AAAI, CIKM, WWW, PAKDD, etc., and IEEE Transaction of Multimedia, Journal of Neurocomputing, Journal of Knowledge and Information Systems, etc.