Research Synopsis

My principal research interests lie in the development of machine learning and statistical methodology; especially for solving problems involving automated learning, reasoning, and decision-making in high-dimensional, multimodal, and dynamic possible worlds in social and biological systems.


Currently the following themes are studied in my group:


o   Foundations of Statistical Learning, including theory and algorithms for:


1)    Learning time/space varying models and evolving structures

2)    High-dimensional inference under structured sparsity, complex conditions for variables, and complex tasks (e.g., multi-attributes, multi-task, missing value, etc.)

3)    Nonparametric Bayesian inference, Posterior regularization

4)    Latent space inference (e.g., topic modeling, sparse coding, deep learning)

5)    RKHS embedding, nonparametric inference, and spectral methods for graphical models

6)    Distributed and online algorithms for optimization, approximate inference, and sampling on Tara-scale data.


o   Large-scale Information & Intelligent System:


1)    Development of general-purpose backend system, programming interface, and infrastructure for large-scale machine learning in distributed computing environment

2)    Multi-view latent space models, topics models, and sparse coding for image/text/relational data mining

3)    Evolving structure, stable metrics, and prediction for dynamic social networks, goal-driven network design and optimization

4)    Web-scale image understanding, search, prediction, and storyline synthesis

5)    User modeling, personalization, temporal analysis, and computational advertising

6)    Information visualization, indexing and storage, web/mobile app development


o   Computational Biology:


1)    Understanding genome-microenvironment interactions in cancer and embryogenesis via joint analysis of genomic, proteomic, and pathway signaling data

2)    Genetic analysis of population variation, demography and evolution

3)    Statistical inference of genome-transcriptome-phenome association in complex diseases

4)    Personalized diagnosis and treatment of spectrum diseases via next generation sequencing and computational "omic" analysis

5)    Biological image and text mining


Last updated 08/1/2013