Publication Type

Conference Proceeding Article

Version

publishedVersion

Publication Date

9-2018

Abstract

Peer assessment is a major method for evaluating the performance of employee, accessing the contributions of individuals within a group, making social decisions and many other scenarios. The idea is to ask the individuals of the same group to assess the performance of the others. Scores or rankings are then determined based on these evaluations. However, peer assessment can be biased and manipulated, especially when there is a conflict of interests. In this paper, we consider the problem of eliciting the underlying ordering (i.e. ground truth) of n strategic agents with respect to their performances, e.g., quality of work, contributions, scores, etc. We first prove that there is no deterministic mechanism which obtains the underlying ordering in dominant-strategy implementation. Then, we propose a Two-Stage Mechanism in which truth-telling is the unique strict Nash equilibrium yielding the underlying ordering. Moreover, we prove that our two-stage mechanism is asymptotically optimal, since it only needs $n + 1$ queries and we prove an $\Omega(n)$ lower bound on query complexity for any mechanism. Finally, we conduct experiments on several scenarios to demonstrate that the proposed two-stage mechanism is robust.

Keywords

Mechanism design, Peer assessment, Nash equilibrium

Discipline

Databases and Information Systems

Research Areas

Data Science and Engineering; Intelligent Systems and Optimization

Publication

Proceedings of the 11th International Symposium on Algorithmic Game Theory (SAGT 2018), Beijing, China, September 11-14

First Page

176

Last Page

188

ISBN

9783319996592

Identifier

10.1007/978-3-319-99660-8_16

Publisher

Springer

City or Country

Beijing, China

Additional URL

https://doi.org/10.1007/978-3-319-99660-8_16

Download

Share

COinS