Tom DuBois(tdubois) I completed my Ph.D. in Computer Science from the University of Maryland with the submission of my dissertation (Using and Manipulating Probabilistic Connectivity in Social Networks) in April and graduated in May 2011.  Slides from my dissertation defense are available here. Before coming to Maryland, I completed by B.S. in Computer Science from Carnegie Mellon University and worked at Cadence Design Systems as a software engineer. Research Interests My main research interests here at UMD are the design and analysis of randomized and probabilistic algorithms for social networks or parallel and/or distributed systems. Problems where random or uncoordinated local choices produce global results are very interesting to me. Professor Aravind Srinivasan has been my Ph.D. advisor and I cannot say enough good things about him. He does great research, cares about his students, and and helps them connect with interesting problems. One of our projects (joint work with some at the Virginia Bioinformatics Institute) involves modeling the spread of disease over large social contact networks and optimizing the application of a constrained intervention (such as a fixed number of vaccinations) to contain it. My other major project (joint work with Professor Jennifer Golbeck) has been social network trust inference. My dissertation proposal is available here along with the accompanying presentation. Here is my CV.

Publications

Thomas DuBois, Stephen Eubank, Aravind Srinivasan, The Effect of Random Edge Removal of Network Degree Sequence, the Electronic Journal of Combinatorics, Volume 19 (2012). Available here.

Barrett C, Beckman R, Bisset K, et al., Optimizing Epidemic Protection for Socially Essential Workers, Proceedings of the 2012 SIGHIT International Health Informatics Symposium, January 2012. My presentation slides are here.

Thomas DuBois, Jennifer Golbeck, Aravind Srinivasan, Network Clustering Approximation Algorithm Using One Pass Black Box Sampling, available here.

Thomas DuBois, Jennifer Golbeck, Aravind Srinivasan, Predicting Trust and Distrust in Social Networks, winner of the Best Paper Award in the Third IEEE International Conference on Social Computing, October 2011. Here is my presentation.

Thomas DuBois, Jennifer Golbeck, John Klient, and Aravind Srinivasan, Improving Recommendation Accuracy by Clustering Social Networks with Trust, 3rd ACM Conference on Recommender Systems workshop: Recommender Systems and the Social Web. Here are the full version and the accompanying presentation.

Thomas DuBois, Jennifer Golbeck, and Aravind Srinivasan, Rigorous probabilistic trust-inference with applications to clustering, Proceedings of the 2009 International Conference on Web Intelligence. Here are the full version and the accompanying presentation.

Tom DuBois, Bryant Lee, Yi Wang, Marc Olano and Uzi Vishkin, XMT-GPU: A PRAM Architecture for Graphics Computation, Proceedings of ICPP-08: the 37th IACC International Conference on Parallel Processing (Portland, Oregon, September 8-12, 2008). Here is the presentation.

DuBois, T.M.; Rudnicky, A.I., "An open concept metric for assessing dialog system complexity,"  Automatic Speech Recognition and Understanding, 2001.  ASRU '01.  IEEE Workshop on, pp. 264-267, 2001

Teaching

I was the sole instructor for the Summer 2008 Algorithms course, CS 451.

In addition I have been a teaching assistant for the following courses at the University of Maryland (as well as several additional courses at Carnegie Mellon as an undergraduate): Cryptography, Algorithms (351, 451, and graduate level), Complexity Theory, and Operating Systems.