AI seminar: An incentive mechanism for promoting honesty in e-marketplaces

Speaker: Jie Zhang

In this talk, we consider the challenge of designing an electronic marketplace populated with buying and selling agents that learn to choose the best business partners, for their users. In particular, we present a novel mechanism that creates incentives for honesty in this electronic marketplace. In our mechanism, buyers model the trustworthiness of other buyers by using a novel method that combines both private and public reputation values. They then select the most trustworthy ones as their neighbors, used to ask advice about sellers. In contrast with other research on creating incentives in electronic marketplaces, we take advantage of the social network of buying agents in order to promote honesty. In particular, we have sellers modeling the reputation of buyers. Reputable buyers provide fair ratings of sellers, and are likely to be neighbors of many other buyers. Sellers will provide more attractive products to reputable buyers, in order to build their reputation. We discuss how a marketplace operating with our mechanism leads to better profit both for honest buyers and sellers. With honesty encouraged, our work promotes the acceptance of agent-oriented e-commerce by human users. We conclude this talk with a brief discussion of topics that we are currently investigating to deepen our development of the model.