Speaker: Jamie Tucker-Foltz, Harvard University

Time: 09:00 a.m., Aug 29, 2024, GMT+8

Venue: Online Talk: Zoom Meeting ID: 810 7822 5258, Passcode: 605986

Abstract: 

Social choice theory studies the problem of aggregating individual preferences into a single set of preferences for society as a whole. It is typically assumed that one has full access to complete descriptions of each individual's preferences. However, online platforms are emerging which seek to aggregate complex preferences over a vast space of alternatives, rendering it infeasible to learn any individual's preferences completely. Instead, preferences must be elicited by an algorithm using simple queries. In this talk, I will present work from two of my recent papers characterizing both the information-theoretic and computational limits of such algorithms:
https://arxiv.org/abs/2402.11104 - EC 2024 paper about ranked preferences
https://arxiv.org/abs/2211.15608 - AAAI 2023 paper about approval preferences
No prior knowledge of social choice theory will be assumed.

Source: Center on Frontiers of Computing Studies, PKU