Victoria Wilkes is an M.Sc. student interested in multi-agent decision problems on social and economic networks and other topics at the intersection of economics and computer science. She completed her B.A. in Economics at the University of Cambridge, examining the effect of clustering on competition and diffusion on social networks for her undergraduate dissertation.
Amirali Salehi-Abari is a Ph.D. student interested in decision making on social and economic networks. In particular, he is working on designing algorithms and mechanisms in social choice (group decision) contexts that support effective decision making when various forms of network externalities are present. He is currently investigating various forms of network externalities that are derived from or induced by relationships present in social networks. He is also developing models of social networks in which agent preferences are correlated with those of their neighbors/friends, and corresponding algorithms for and analyses of network formation, inference and learning.
Joel Oren Joel Oren is a Ph.D student (co-supervied by Allan Borodin) interested in computational social choice, with a focus on the combinatorial optimization and the game theoretic aspects of social choice. His past and present projects include: the design of algorithms for multi-winner social choice and slate optimization, efficient methods for preference elicitation, and the study of influence diffusion processes in social networks.
Joanna Drummond: Joanna Drummond is a Ph.D. student interested in decision theory, applied machine learning, and preference elicitation. She is currently applying decision-theoretic techniques to investigate stable matching problems with partial preference information and support effective preference elicitation. In her M.Sc. work, Joanna developed exact and approximate algorithms to compute optimal regret minimizing matches given partial information. She completed her B.Sc. in Computer Science and Mathematics at the University of Pittsburgh. Supervised by Diane Litman, her undergraduate research consisted of applying supervised machine learning techniques to detect users' affect in spoken tutorial dialogues. This work was completed with the ITSPOKE group.
Andrew Perrault: Andrew Perrault is a Ph.D. student interested in mechanism design and other problems at the intersection of economics and computer science. He is interested in optimization and mechanism design in smart grid settings, with a focus on coordinating the activities and electricity purchasing behavior of large consumer coalitions to benefit the His M.Sc. research dealt with optimal coordination, and supporting pricing schemes, of the activities of generators and consumers in micro-grids. As an undergraduate, he was associated with the Institute for Computational Sustainability at Cornell and there contributed to solving the problem of generating spatially-balanced Latin squares of arbitrary dimension. In the Personal Robotics Lab, he participated in a project that used machine learning to select grasping locations for a robotic manipulator that take into account future placing plans. He is a founder of theschoolfund.org. He enjoys rock climbing and sailing.
Tyler Lu is a Ph.D. student interested in decision-theoretic and
machine learning-based approaches to social choice and preference
aggregation. He is currently applying ideas from statistical
learning and preference elicitation to cope with incomplete
preferences when making collective decisions, with an eye towards
analyzing the social-choice theoretic properties of such models.
He is also interested in applying these methodologies to recommender
systems and matching markets. For his master's degree at the University of
Waterloo, he developed
theoretical models and analyses of learning with unlabeled data.
His other interests include Bayesian learning and MCMC, statistical
learning theory, and discrete algorithms.
Tyler is also the Co-founder of an exciting new start-up, Granata Decision Systems.
Xin Sui is a Ph.D. student interested in game theory and mechanism design. In particular, he is working on designing mechanisms that incentivize self-interested agents to truthful reveal their preference to achieve a social desirable outcomes. His recent research has focused on mechanism design and optimization schemes for single-peaked and spatial preferences (e.g., as in facility location problems), and on the empirical analysis of voting data to discover the extent of single-peaked or spatial preferences. He has also worked on the design of incremental, partial revelation mechanisms that allow explicit tradeoffs to be made between outcome quality, communication, and privacy, and the application of these to problems ranging from auction design to facility location.
Laurent Charlin (co-supervised by Rich Zemel) completed his Ph.D. in 2014, and developed learning methods (including active learning methods) for matching problems with partial user preferences, as well as the application of such methods to recommender systems. He is the developer of the widely used Toronto Paper Matching System as well. His M.Sc. focused on the automated discovery of abstractions in partially observable Markov decision processes (POMDPs). His other interests include statistical learning theory, and problems at the intersection of economics and machine learning. He is currently a postdoc at Princeton working with David Blei.
Kevin Regan completed his Ph.D. in 2013, developing novel preference elicitation techniques for sequential decision making. In particular, he has developed models for the elicitation of reward functions for Markov decision processes (MDPs) and robust optimization techniques from MDPs with imprecisely specified reward functions. He has applied this work in a variety of domains including assistive technology, autonomic computing, and web site design. His main interests lie in the intersection of preference elicitation, reasoning under uncertainty and mechanism design. He has also done work with multi-agent systems, reputation and trust models, and electronic markets. He is currently working at Google in Cambridge, MA.
Darius Braziunas completed his Ph.D. in 2012 on new, effective ways of eliciting, representing and reasoning with user preferences and utility functions. As part of his research, he developed the UTPref Recommendation System that maintains an explicit (but incomplete) multiattribute utility model of user preferences, and uses minimax regret to guide decision-theoretically sound elicitation of preferences and recommendation of options. Other areas of interest include sequential decision making models (MDPs and POMDPs), recommendation systems, collaborative filtering, and discrete optimization. Darius was a Research Scientist at Thoora, Inc. (Toronto) for a time during his Ph.D. studies, and is now a Senior Research Scientist at Kobo, Inc. (Toronto).
Bowen Hui completed her Ph.D. in 2011, in which she developed a decision-theoertic framework for providing automated help and interface customization in software enviornments in a way that is sensitive to the preferences and abilities of different users. This research lies at the intersection of AI and HCI, involving preference elicitation, probabilistic inference, POMDPs, and other modeling techniques. She is also interested in preference elicitation, multi-agent communication, psychological effects of preference elicitation and negotiation, among other topics. She is currently an Instructor in Computer Science at the University of British Columbia, Okanagan.
Paolo Viappiani is an CNRS Researcher at Laboratoire d'informatique de Paris 6 at Université Pierre et Marie Curie. His research interests include recommender systems, preference elicitation, mathematical models for social networks, interactive optimization and machine learning. He was a postdoc here from 2008-2010, where much of his work focused on open-ended preference elicitation, set-based recommendation, and handling noisy responses in preference elicitation models.
Jesse Hoey is Assistant Professor at the University of Waterloo. Jesse's interests include decision making under uncertainty (POMDPs), machine learning, computer vision, among other things. He has devoted a lot of attention to assistive technologies. He was a postdoc here from 2004-2006, working jointly in Computer Science and the Toronto Rehabilitation Institute. He was a lecturer at the University of Dundee from 2006-2010 before returning to Ontario.
Scott Sanner is Senior Researcher in the Statistical Machine Learning Group at NICTA and an adjunct professor at the Australian National University. He is interested in a variety of areas of AI, including decision-making under uncertainty, machine learning, reinforcement learning, knowledge representation and reasoning, and information retrieval. He completed his PhD, "First-order decision-theoretic planning in structured relational environments," in 2008.
Georgios Chalkiadakis is an Assistant Professor in the Department of Electronic and Computer Engineering at the Technical University of Crete. He is interested in various aspects of multiagent systems and decision making under uncertainty, especially coalition formation and reinforcement learning. He completed his PhD, "A Bayesian Approach to Multiagent Reinforcement Learning and Coalition Formation under Uncertainty," in 2007.
Nathanael Hyafil is Head of Risk Methodologies at GDF SUEZ Energy Management & Trading in Paris. His Ph.D. work dealt with on computational aspects of mechanism design and game theory, in particular, on the design of mechanisms requiring only partial type revelation while preserving incentive properties and reasonable decision quality. He completed his PhD, "Mechanism Design with Partial Revelation", in 2007.
Michael Pavlin is an Assistant Professor in Operations & Decision Sciences, at Wilfred Laurier University's School of Business and Economics. He is interested in computational issues that arise in economic systems. His recent work focused on ascending auctions in the presence of externalities and is motivated by applications to wireless communication networks. He completed his MSc, "Ascending Auction for Markets with Externalities and Applications to Routing in Wireless Networks," in 2006.
Pascal Poupart is an Associate Professor of Computer Science at the University of Waterloo. Pascal's interests include decision making under uncertainty (including MDPs and POMDPs), machine learning, applications to health informatics and dialog systems. He completed his PhD, "Exploiting Structure to Efficiently Solve Large Scale Partially Observable Markov Decision Processes," in 2005.
Bob Price is Research Scientist at Xerox PARC (following a postdoc at The University of Alberta with the Alberta Ingenuity Center for Machine Learning. His interests include reinforcement learning (including imitation, teaching and knowledge transfer), machine learning, Markov decision processes, and web personalization. He completed his PhD, "Accelerating Reinforcement Learning through Imitation," in 2003.
Richard Dearden completed his Ph.D. in Fall 2000 on Markov decision processes as a model for decision theoretic planning and reinforcement learning. He developed methods for the learning and use of Bayesian methods for estimating value functions and models in reinforcement learning and developing exploration methods based on uncertainty in these estimates. He has been deeply involved in much of the work on structured policy construction algorithms for MDPs as well. He is currently a Senior Lecturer at the University of Birmingham School of Computer Science.
Alexander Kress completed his M.Sc. degree in 2004. His thesis title was "An Incremental Elicitation Approach to Limited-Precision Auctions."
Darius Braziunas finished his M.Sc. degree in 2003. His thesis title was "Stochastic Local Search for POMDP controllers".
Tianhan Wang finished an M.Sc. degree in 2003. His thesis title was "Preference Elicitation using the Minimax Regret Decision Criterion."
I'll augment this list with my students from my time at UBC sometime soon.