Research

My main current topics of interest are:

• machine learning
• theory and algorithms
• natural language processing
• speech processing
• string algorithms
• computational biology

The following are some selected topics with a list of publications:

• domain adaptation and sample bias correction
• ranking
• learning kernels
• transductive inference
• rational kernels
• shortest-distance algorithms
• speech recognition
• weighted automata and transducers

Some recent talks:

• Discrepancy and Adaptation, NIPS 2011 Workshop, Sierra Nevada, Spain, December 2011.
• Domain adaptation in regression, ALT 2011, Espoo, Finland, October 2011.
• Tutorial: Learning Kernels, ICML 2011, Bellevue, Washington, July 2011.
• Learning Kernels Can Help Performance, Center for Information and Systems Emgineering (CISE), Boston University, Boston, MA, April 2011.
• Learning Bounds for Importance Weighting, NIPS 2010 poster, Vancouver, Canada, December 2010.
• Domain Adaptation Theory and Algorithms, Polytechnic Institute, Brooklyn, NY, December 2010.
• Learning in a Non-Ideal World, PRIML (Penn Research in Machine Learning), University of Pennsylvania, PA, October 2010.
• Generalization Bounds for Learning Kernels, ICML 2010, Haifa, Israel, June 2010.
• Matrix Approximation for Large-Scale Learning, MMDS 2010, Stanford University, Stanford, CA, June 2010.
• Learning kernels, NIPS workshop, Vancouver, Canada, December 2009.
• Learning languages and rational kernels, NIPS workshop, Vancouver, Canada, December 2009.
• Learning with expert advice (tutorial), ML Seminar, Courant Institute, New York, NY, December 2009.
• Learning with imperfect data, Google MTV & UC Berkeley, CA, April 2009.