At the University of Toronto I was co-advised by Allan Borodin and Kate Larson, and focused on computational economics and multi-agent systems problems. My research focused on stable matching with partial preference information. We applied decision theoretic techniques (which have successfully been applied for voting with partial preference information) to this bipartite graph matching problem. Interesting new problems arise because of this bipartite structure. We focused on the non-transferable utility setting, as this is the setting for many important real-world variants of this problem (e.g. the residency matching problem, school assignment problem, etc.). I received my Masters Spring 2013, with my Masters Thesis focusing on preference elicitation for the stable matching problem. My Masters was advised by Craig Boutilier. After my Masters, I began pursuing my PhD.
My undergraduate research at the University of Pittsburgh primarily focused on applying machine learning techniques to spoken educational data. I was advised by Diane Litman, working in the ITSPOKE Lab. I primarily worked on detecting students' affect while using an intelligent spoken physics tutor, using both prosodic and lexical features. I also investigated the effect of different kinds of training data on these models.
Additionally, as an undergraduate, I participated in the CRA-W's DREU program, where I traveled to the University of Southern California's Information Sciences Institute to work with Jihie Kim. Working with forum data from a computer science class at USC, I performed a statistical corpus study to analyze student posting behaviour, and applied machine learning techniques to automatically categorize student posts.