Kernel methods for estimating latent variable models

12noon lunch and lecture ¿Perspectives in Machine Learning¿3:30-4:30pm Seminar reception followingResearchI am heading the Machine Learning Group at Georgia Institute of Technology. I conduct research in the development of machine learning methodology, and the applications of machine learning to interdisciplinary problems. I am fascinated by the prospect of intelligent systems which can learn from massive volumes of complex, uncertain and high dimensional data, and reveal trends and patterns too subtle for humans to detect. Problems I have addressed are as diverse as finding disease markers from thousands of candidate genes, modeling and predicting nonlinear spatial-temporal dynamics in sensor time-series, and estimating and analyzing social and biological networks. Data arising from these applications are often characterized by complex statistical features (eg. multi-modality, skewness), and long-range and hierarchical dependencies among the involved variables.