I am affiliated with the UH Data Mining and Machine Learning Group. My Advisor is Dr. Ricardo Vilalta.

I am broadly interested in Machine Learning and it's applications to the basic sciences (particularly physics and astronomy).

My current research activities focus on dealing with the labeled-unlabeled problem also known as the semi-supervised learning problem. In many real world applications (such as image analysis, drug discovery, credit scrore checking etc.) labeled data is often fairly expensive to obtain, however unlabeled data is abundantly available. The goal of Semi-suprevised learning research is to train superior learners by extracting information from the unlabeled data (in addition to the labeled data). I am presently working on a co-training style ensemble based semi-supervised learner.

On a more applcation based note, I am working on developing techniques for automated classification of martian landscape into constituent landforms. We are primarily interested in classifying various crater components. Here is a breif abstract describing our work. Also a poster describing the work so far can be found in the publications section. The project is a collaboration between the Lunar and Planetary Institute (Dr. Tomasz Stepinski) and the University of Houston. The project is funded by NASA's Applied Information Systems Department .