Given twice, 2016
Elicitation for Preferences Single Peaked on Trees
How few questions reveal everyone's ranking when preferences are single-peaked on a tree?
An intermediate research talk, for graduate students and up.
Abstract
As published on my old research pageI gave this talk when I was visiting Duke University during the summer of 2016. It was based on joint work with Palash Dey. The talk focused on the problem of preference elicitation, where the goal is to understand the preferences of agents (which we model by total orders) by querying them about their pairwise preferences.
Source: web.archive.org.