Special Issue on Learning and Computations in Games and Economics
Guest Editors:Mehryar Mohri, Google Research and Courant Institute, firstname.lastname@example.orgVianney Perchet, Crest – ENSAE, Criteo AI Lab, email@example.com
Learning from past interactions between self-interested parties is increasingly critical in game-theoretical contexts. A key example of such interactions is the billions of online advertisement auctions run everyday, whose outcome determines the revenue of modern search engine and popular online sites. Learning from these interactions to bid or using the data collected to determine selling strategies therefore become crucial tasks.
These strategic interactions between multiple agents learning their behaviors go beyond single-item auctions and extend to more general mechanisms, including matchings, contracts, bargains, and many others with perhaps the final objective of designing new markets, as illustrated by the new 5G networks. More generally, learning also plays a critical role simply due to repetitions in generic games.
This raises both new theoretical and practical questions relating traditional game-theoretical concerns to learning and approximation algorithms in computer science and economics. This includes questions such as: How many samples or how many repetitions of a game are needed to determine a “good’’ strategy or to find “good’’ policies in those games? How are the solutions derived affected by the number of players or other aspects of the games? Are there stable and robust solutions? What can be shown about the long-term behavior when all sides are learning simultaneously and/or competitively?
These and many other related questions are among those we seek to tackle in this special issue of Dynamic Games and Applications. We invite submissions with strong contributions including, but not limited to the following areas:
Sample complexities and statistical guarantees for game-theoretic solutions.Approximation performances and efficient algorithms.Price of anarchy and variants.Regret and learning procedures in games.Counter-Factual Regret minimization for practical games.Online vs. batch algorithms.Mean field limits in games.(Deep) learning of combinatorial or complex auctions and mechanism.Incentives and efficiency in mechanisms (auction, matchings, etc.).Design of strategies in reinforcement Learning and stochastic games.
Submission timetable: Submissions will be processed as soon as they are received. The final deadline is October, 15 2021. We encourage early submissions. The accepted papers will appear online in advance of the production of the full special issue.
Click to:- download the Call for Papers- see the submission guidelines – submit your manuscript