A game theoretic approach for reputation mechanisms

A game theoretic approach for reputation mechanisms

Antimo Barbato
DEI PhD Student

DEI - Aula 3B
November 3rd, 2011
at 2.00 p.m.

Abstract:


Reputation mechanisms are currently used in a wide range of applications with the final goal of obtaining information about the quality of products or services based on feedbacks submitted by users. This information allows future clients to make better decisions on whether to buy or not that product/service. Unfortunately, the final outcome of these mechanisms may not reflect the real quality of a product/service since they are susceptible to fraud and deception. This problem can be solved by rewarding clients for submitting feedbacks based on a payment mechanism that incentives users to report their opinion truthfully. In this work we use a game theoretic approach for studying reputation mechanisms, addressing the design of incentive-compatible payments (i.e. honest reporting is a Nash Equilibrium) for non binary quality level scenarios. Numerical results are also provided and discussed.

Research area:
Artificial intelligence, robotics and computer vision