Electrical Engineering
      and Computer Sciences

Electrical Engineering and Computer Sciences

COLLEGE OF ENGINEERING

UC Berkeley

   

2009 Research Summary

Balancing Public Reputation-based Incentives with Anonymity

View Current Project Information

Anant Sahai and Parvathinathan Venkitasubramaniam1

The Internet revolution opens the possibility of new forms of "social production" of goods and services. Not only does it reduce transaction costs, but it offers the possibility of new kinds of incentives to promote contributions from various people. In a sense, this is a distributed control problem in which the "plant" consists of humans. Many current business/service models are based on a kind of cherry-picking or hero-amplification: there are billions of humans and natural variability among them makes some of them more willing to contribute. The goal of cherry-picking is to encourage enough of these special individuals to contribute in an endevour that uses the Internet to amplify their effect so that millions more can benefit. The challenge is to see if we can also use the Internet to encourage social production among non-heros.

This project studies the issues motivated by a practical context of peer review. The review cycle for papers takes way too long in many disciplines. The problem is that while authors want to have their own papers reviewed fast, they are often unwilling to review the papers of others in a timely manner. This paper explores what would be required to incentivize fast reviews using a public reputation system that exploits the fact that the referees are drawn from the same pool as paper authors.

The challenge is that the identity of referees should remain as anonymous as possible. We are exploring a model in which authors have an incentive to commit to reviewing papers and are rewarded for meeting this commitment in a way that prioritizes their own papers for reviews. This ensures stability (bounded reviewing delays) for all fair contributers while freeloaders face a potentially unstable system. However, a naive implementation of reputations leaks information that would allow authors to infer the likely identities of their referees. We explore the introduction of a distortion to the observed public reputation process that can be used to enhance anonymity while preserving the incentives for timely refereeing.

Instead of "control with communication constraints" we get the problem of "incentives with anonymity constraints."

1UC Berkeley