1. Fast approximate spectral clustering
    -- A framework for fast computation of spectral clustering based on the notion of "continuity".

2. Cluster Forests
    -- Randomized feature-pursuit based cluster ensemble algorithm (unsupervised Random Forests for clustering).

3. Public health and medical imaging
    -- Issues related to public health and algorithms for the scoring of tissue images for cancer study.

4. Social network mining
    -- Mining and understanding of user behaviors in social networks.

5. Education in Statistics and Data Science
    -- Various aspects including curriculum design and tool development.

6. Cost-bounded learning
    -- Machine learning algorithms that incorporate costs.

7. Remote sensing and environmental science
    -- Applications, methods, and theory for the analysis of remote sensing images and various issues in environment and ecology.

8. Distributed information sharing, learning and inference
    -- Efficient algorithms for information sharing, learning and inference over data distributed in multiple sites.

9. Random projection forests
    -- A versatile tool for data mining, statistical inference and machine learning. It could be used to learn various properties in the data. Example applications include large scale k-nearest neighbor search and representation learning etc.