Prof. Xiaotie Deng is a Chair Professor of the Center on Frontiers of Computing Studies (CFCS) at Peking University, Director of Blockchain Committee at China Society for Industrial and Applied Mathematics (CSIAM), and Director of the Center for Multi-agent Research at Institute for Artificial Intelligence, Peking University. Before joining Peking University, he worked at Shanghai Jiaotong University, University of Liverpool, City University of Hong Kong, and York University. He received his Bachelor's degree from Tsinghua University in 1982, Master's degree from Chinese Academy of Sciences in 1984, and Ph.D. from Stanford University in 1989.
Prof. Deng's main research interests are algorithmic game theory, blockchain, internet economy, online algorithms and parallel computing. His recent research focuses on equilibrium computation, multi-agent games, internet economics, and game theory in blockchain.
Prof. Deng is a fellow of the Association of Computing Machinery (ACM) for his contributions to the interface of algorithms and game theory (2008), and the Institute of Electrical and Electronics Engineers (IEEE) for his contributions to computing in partial information and interactive environments (2019). He was elected as foreign member of Academia Europaea in 2020, CSIAM fellow in 2021. In 2022, he won ACM SIGecom Test of Time Award together with Xi Chen, Constantinos Daskalakis, Paul Goldberg, Christos Papadimitriou and Shanghua Teng.
Abstract: The widespread use of intelligent algorithms in games and economic activities has been creating new challenges of manipulated dynamic information to optimization, coordination and compromise in decision-making in the multiagent world. Could the wide applicable bounded rationality play a renewed role to set in order the world dynamic in operations and cognitions?
In this direction, we discuss some recent efforts in approximation, commitment manipulation, and insightful equilibrium for the settings of multi-agent learning stochastic games, extensive games and selfish blockchain mining.
We hope this renewed effort would be of help in the hybrid environment of human and multi agent games.
Prof. Hervé Moulin FRSE FBA (born 1950 in Paris) is a French mathematician who is the Donald J. Robertson Chair of Economics at the Adam Smith Business School at the University of Glasgow. He is known for his research contributions in mathematical economics, in particular in the fields of mechanism design, social choice, game theory and fair division. He has written five books and over 100 peer-reviewed articles.
Prof. Moulin was the George A. Peterkin Professor of Economics at Rice University (from 1999 to 2013):,the James B. Duke Professor of Economics at Duke University (from 1989 to 1999), the University Distinguished Professor at Virginia Tech (from 1987 to 1989), and Academic Supervisor at Higher School of Economics in St. Petersburg, Russia (from 2015 to 2022). He is a fellow of the Econometric Society since 1983, and the president of the Game Theory Society for the term 2016 - 2018. He also served as president of the Society for Social Choice and Welfare for the period of 1998 to 1999. He became a Fellow of the Royal Society of Edinburgh in 2015.
Prof. Moulin's research has been supported in part by seven grants from the US National Science Foundation. He collaborates as an adviser with the fair division website Spliddit, created by Ariel Procaccia.
Abstract: We revisit the trade-off between the liberal (Coasian) approach to fair division by decentralised person-to-person negotiations, apt to discover unforeseen compromises and promote cooperation, and the application of a deterministic, normatively appealing division rule completely eliminating unscripted outcomes.
At the ex ante stage where individual characteristics (preferences, efforts, demands, endowments, rights, etc..) are still private, our benevolent regulator stakes out the worst and best welfare a participant can achieve as a function only of their own characteristics: minimising the range from best to worst case limits the pitfalls of face to face bargaining. If the unsupervised agents fail to reach an agreement within the posted range, the regulator steps in to implement an outcome that does.
Discovering the full menu of worst and best cases welfare functions is a hard mathematical question, even in simple problems of resource allocations. We give full solutions for several iconic fair division problems with transferable utilities, including the allocation of indivible goods or costly chores and the exploitation of a commons. The corresponding menus of multi-valued bargaining ranges contain many familiar single-valued rules but also suggest new ones.
Prof. Hartline’s research introduces design and analysis methodologies from computer science to understand and improve outcomes of economic and legal systems. Optimal behavior and outcomes in complex environments are complex and, therefore, should not be expected; instead, the theory of approximation can show that simple and natural behaviors are approximately optimal in complex environments. This approach is applied to auction theory and mechanism design in his graduate textbook Mechanism Design and Approximation which is under preparation.
Prof. Hartline received his Ph.D. in 2003 from the University of Washington under the supervision of Anna Karlin. He was a postdoctoral fellow at Carnegie Mellon University under the supervision of Avrim Blum; and subsequently a researcher at Microsoft Research in Silicon Valley. He joined Northwestern University in 2008 where he is a professor of computer science. He was on sabbatical at Harvard University in the Economics Department during the 2014 calendar year and visiting Microsoft Research, New England for the Spring of 2015. He is currently on sabbatical at Stanford University in the Economics Department.
Prof. Hartline is the director of Northwestern’s Online Markets Lab, he was a founding codirector of the Institute for Data, Econometrics, Algorithms, and Learning from 2019-2022, and is a cofounder of virtual conference organizing platform Virtual Chair.
Abstract: Scoring rules are everywhere. Any decision problem where an agent has beliefs about an unknown state and takes an action and realizes payoffs according to the action and the realized state is a scoring rule. Behavioral subjects in experiments or evaluated and rewarded according to scoring rules. Machine learning algorithms are trained and evaluated according to scoring rules.
This talk will introduce an objective for optimizing proper scoring rules. The objective is to maximize the increase in payoff of a forecaster who exerts a binary level of effort to refine a posterior belief from a prior belief. In this framework we characterize optimal scoring rules in simple settings, give efficient algorithms for computing optimal scoring rules in complex settings, and identify simple scoring rules that are approximately optimal. In comparison, standard scoring rules in theory and practice -- for example the quadratic rule, scoring rules for the expectation, and scoring rules for multiple tasks that are averages of single-task scoring rules -- can be very far from optimal.
Prof. Inbal Talgam-Cohen is a computer science researcher specializing in the fields of algorithms, economics, and operations research. She holds an MSc from the prestigious Weizmann Institute, advised by Uriel Feige. In addition to her computer science background, Prof. Talgam-Cohen also has a law LLB degree and served as a judicial clerk to Supreme Court Justice Hon. Esther Hayut. Having completed her PhD at Stanford University under the guidance of Tim Roughgarden, Prof. Talgam-Cohen worked as a Post Doc and a senior research assistant under the supervision of Prof. Michal Feldman in her AGT Lab project. Prof. Talgam-Cohen has also joined internships at Microsoft Research and Yahoo! Labs during her summer breaks. In September 2018, Prof. Talgam-Cohen joined the esteemed Technion as an assistant professor of computer science. Her dedication and contributions have been recognized with several prestigious awards, including the Best Doctoral Dissertation from ACM SIGecom, the Stanford Interdisciplinary Graduate Fellowship (SIGF), and the esteemed Best Student Paper Award at ACM EC.
Abstract: The computational research of contract design is an exciting new frontier of AGT. I demonstrate the potential of the computational approach to shed new light on contract design through two projects, on approximation and on learning.
First, an analysis of approximation guarantees of simple contracts under the Bayesian framework provides a new justification for the prevalence of linear contracts. We consider a hidden-action principal-agent model, in which actions require different amounts of effort, and the agent’s cost per unit-of-effort is private. We show that linear contracts are near-optimal whenever there is sufficient uncertainty in the principal-agent setting.
Based on joint works with Tal Alon, Paul Duetting & Yingkai Li (EC'23), and Eden Saig and Nir Rosenfeld (NeurIPS'23 Spotlight).