CS Postdoc Profiles

Research Profile

I expect to receive my Ph.D. in Computer Science from the University of Maryland College Park in June 2010 as a member of the Laboratory for Computational Cultural Dynamics under the supervision of Dr. V.S. Subrahmanian. My thesis work focuses on stochastic reasoning about what kinds of actions are expected from agents in an adversarial setting. This work involved the development of Action Probabilistic Logic Programs (an extension to the foundational work on Probabilistic Logic Programs by Ng and Subrahmanian in 1992) and the treatment of a host of different problems related to the main goal, deriving both theoretical and empirical results in the process. These problems focused on: scalable computation of most probable models of probabilistic logic programs (a problem previously not addressed in the literature); abductive inference in probabilistic logic programs, in which we wish to find conditions under which certain models will be the most probable ones, or certain queries are entailed with given probabilities; a novel approach to represent fulfillment of promises between agents, with the objective of reasoning about the likelihood with which pending or future promises will be fulfilled (and to what degree); and computing optimal policies in multi-agent settings in which the reasoning agent maintains an action probabilistic logic program modeling the rest of the agents’ actions.

Apart from my thesis work, I have also been involved in work on reasoning under uncertainty in relational databases. The first project of this kind that I worked on tackled the problem of managing inconsistent information; as a result we proposed Inconsistency Management Policies, a novel approach to the problem which combines both identifying logical inconsistencies (i.e., violations of integrity constraints) and obtaining a measure of how large these inconsistencies are. The second of these projects involved extending seminal work on probabilistic schema mappings (by Dong et al., 2008) in order to answer aggregate queries in this setting. We proposed novel semantics for these queries, and obtained efficient algorithms, which we proved correct and evaluated empirically.

Even though my main interests lie in the areas of Knowledge Representation and Reasoning and Logic Programming, I have been interested in Databases for some time now. During my late undergraduate and early graduate careers I was a Teaching Assistant and then Head Teaching Assistant in Database Theory and Design; the projects described above have certainly helped consolidate my interest in the area, and I now enjoy working in the intersection of Databases and KR.

Finally, I believe that I satisfy the essential criteria to become a successful postdoctoral researcher, since I have: a research record with strong publications in reasoning under uncertainty, knowledge representation and reasoning, logic programming, and database theory; been actively involved in the development of prototype implementations (in Java) of all the work described above; collaborated with a number of researchers of different backgrounds in the context of interdisciplinary projects; and presented many papers and given talks on my work. I provide a more in-depth description of my qualifications and experience as well as the names and contact information of my references in my CV, which can be found on my web page.

Contact Information

E-Mail: EMAIL OBFUSCATED

Phone: +1 301 226 7877

Categories Posted To:

AI / Machine Learning / Robotics / Vision, Databases / Information Retrieval / Data Mining