My professional interest is in developing analytics for decision support. Currently, my focus is analytics based on unstructured information. Specifically, I am working on document retrieval using syntactic and semantic constraints derived from natural language questions, query expansion using structured knowledge bases, and information extraction from documents using entity linking and text classification.
I am exploring how such analytics can help physicians make better clinical decisions. They are faced with the challenging task of considering increasing amounts of unstructured information, both from patients' electronic medical records as well as the growing body of medical knowledge contained in text books, clinical guidelines, and published research. This domain is characterized by changing information retrieval requirements that reduce the lifespan of pre-trained classifiers. Our interest is in developing teachable classifier systems that can learn from the human experts during the course of use.
My previous experience has been in the development of decision-support analytics based on structured information such as those obtained from enterprise databases. I have applied them in a variety of business areas such as semiconductor manufacturing, data centers, grid computing, and manufacturing supply chains.
I received a Ph.D. in Electrical and Computer Engineering from Vanderbilt University in 1995 where my research was on task planning for robots operating in environments not specifically designed for them.