Lirong Xia is Assistant Professor at the Computer Science Department of the Rensselaer Polytechnic Institute (RPI). His research interests are mainly focused on Artificial intelligence, Algorithm design, Social choice (voting) theory, Game theory and Mechanism design. He will be a General Co-Chair at AMMA 2015, the Third Conference on Auctions, Market Mechanisms and Their Applications. We talked with him about the upcoming conference and about some important issues it will examine.
What is the main contribution AMMA can bring to the field of market mechanisms?
I think the most distinctive and exciting feature of AMMA is that it highlights real-world applications of market mechanisms. In general, the AMMA series provides a forum to promote more multidisciplinary interactions and interplays between economists and computer scientists, and between theoreticians and applied researchers.
AMMA 2015 will examine some theoretical and empirical questions like: ‘Is there a need for new mechanisms for specific applications?’ What is your personal opinion about this question?
I believe the answer is yes because there is no “optimum” mechanism that works well for all applications. In social choice, this has been shown by many remarkable impossibility results such as Arrow’s theorem and Gibbard-Satterthwaite theorem. For particular applications, some mechanisms such as VCG and Gale-Shapley might be more promising. Still, existing mechanisms often face many challenges such as high computational complexity, lack of information, complications in preference representation and elicitation, and unpredictable human behavior, etc., especially in new applications.
How to address these challenges has recently ignited enthusiasm in computer science. For example, multi-agent resource allocation studies how to distribute resources among autonomous agents. Building upon ideas and mechanisms in market design and auction theory, researchers have invented new mechanisms as well as new principles to evaluate them. Some of recent progresses have been summarized in the Handbook of Computational Social Choice, which will hopefully be published later this year.
By bringing together theoreticians and applied researchers, we are hoping to better understand and formulate these challenges for researchers to refine the theory and design application-specific mechanisms to improve efficiency and effectiveness.
AMMA 2015 will also focus on crowdsourcing, a process that is getting a foothold in today’s market. In your opinion, what role will it have in the future?
It is a very exciting direction and we expect to see more interactions between market mechanisms and crowdsourcing. For one thing, mechanisms in market design and auctions may improve efficiency of task allocation and information aggregation in crowdsourcing. For another, crowdsourcing platforms may also serve as testbed for new mechanisms and supporting assumptions on preference structure, human behavior, etc.