Aviva Research Fellow
Applied Mathematics and Theoretical Physics
Clustering, Active Learning, Image Analysis
My research interests mainly lie in the field of unsupervised learning. Generally speaking, it is about learning from and modelling on data that do not have any labelling information. Within the field, I am particularly interested in clustering and active learning.
Clustering is the art of grouping a collection of unlabelled objects into a number of clusters, such that objects in the same group are more similar to each other as compared to objects in different groups. Part of my research is about designing new clustering methodologies with appropriate similarity measures and grouping criteria depending on the data sources.
However, without external labelling information, it is difficult to measure the effectiveness of the clustering model and usefulness of the clustering results. At the same time, it can be monetary expensive and labour intensive to acquire labels. This leads to another part of my research, which focuses on the design of new active learning strategies that determine the labels of which data objects to query, not only for the purpose of model validation but also for further improving the model performance effectively and efficiently.