Abstract: Algorithms transform information into decisions. But in many contexts, information is not easily available, or is held by the same people who are affected by the decisions. How do we design provably robust algorithms that obtain good information and produce good decisions in these strategic contexts? This talk will survey recent work in this setting. Our first main problem will be eliciting and aggregating forecasts for group decisionmaking. Our second will be coordinating matchings on graphs, such as matching employers to employees.

Bio: Bo Waggoner is an Assistant Professor of Computer Science at the University of Colorado-Boulder, where he is part of the Theory and Algorithmic Economics groups. His work focuses on algorithms for eliciting information and making decisions in contexts with strategic behavior. Prior to CU Boulder, he was a postdoc at Microsoft Research and at Penn.