We propose a flexible statistical decision-theoretical framework for social choice to design and analyze user-specific social choice mechanisms by the unified approach from statistics, economics, and computation. We focus on Bayesian estimators within the framework to obtain new mechanisms, analyze their social choice axiomatic properties, prove an impossibility theorem, and propose efficient MCMC methods to compute them.

This talk is based on the following papers:

Hossein Azari Soufiani, David C. Parkes, and Lirong Xia. A Statistical Decision-Theoretic Framework for Social Choice. NIPS-14. Full oral presentation.

David Hughes, Kevin Hwang, and Lirong Xia. Computing Optimal Bayesian Decisions for Rank Aggregation via MCMC Sampling. UAI-15.

Lirong Xia. Bayesian Estimators as Voting Rules. UAI-16. Oral presentation.