KAUST Ph.D. students David Evangelista and Xianjin Yang won best paper awards at international conferences this summer for their work in mean-field game theory. Evangelista's paper focused on solutions for stationary mean-field games with congestion, while Yang's paper developed numerical methods for homogenization problems. The awards were presented at the 18th International Symposium on Dynamic Games and Applications in France and the 12th American Institute of Mathematical Sciences (AIMS) Conference in Taiwan. Why it matters: The recognition highlights KAUST's strength in applied mathematics and computational science, specifically in the emerging field of mean-field games with applications across various domains.
This paper introduces rational counterfactuals, a method for identifying counterfactuals that maximize the attainment of a desired consequent. The approach aims to identify the antecedent that leads to a specific outcome for rational decision-making. The theory is applied to identify variable values that contribute to peace, such as Allies, Contingency, Distance, Major Power, Capability, Democracy, and Economic Interdependency. Why it matters: The research provides a framework for analyzing and promoting conditions conducive to peace using counterfactual reasoning.
This paper introduces DaringFed, a novel dynamic Bayesian persuasion pricing mechanism for online federated learning (OFL) that addresses the challenge of two-sided incomplete information (TII) regarding resources. It formulates the interaction between the server and clients as a dynamic signaling and pricing allocation problem within a Bayesian persuasion game, demonstrating the existence of a unique Bayesian persuasion Nash equilibrium. Evaluations on real and synthetic datasets demonstrate that DaringFed optimizes accuracy and convergence speed and improves the server's utility.
Munther Dahleh from MIT gave a talk on information design under uncertainty, focusing on the challenges of creating an information marketplace. The talk addressed the externality faced by firms when information is allocated to competitors, and considered two models for this externality. The presentation included mechanisms for both models and highlighted the impact of competition on the revenue collected by the seller. Why it matters: The research advances understanding of information markets and mechanism design, relevant to the growing data economy in the GCC region.
This paper addresses exploration in reinforcement learning (RL) in unknown environments with sparse rewards, focusing on maximum entropy exploration. It introduces a game-theoretic algorithm for visitation entropy maximization with improved sample complexity of O(H^3S^2A/ε^2). For trajectory entropy, the paper presents an algorithm with O(poly(S, A, H)/ε) complexity, showing the statistical advantage of regularized MDPs for exploration. Why it matters: The research offers new techniques to reduce the sample complexity of RL, potentially enhancing the efficiency of AI agents in complex environments.
The paper introduces MIRAGE, a framework for evaluating LLMs' ability to simulate human behaviors in murder mystery games. MIRAGE uses four methods: TII, CIC, ICI and SCI to assess the LLMs' role-playing proficiency. Experiments show that even GPT-4 struggles with the complexities of the MIRAGE framework.
KAUST Professor Peter Markowich discusses the role of mathematics in football, describing a match as a random process with a drift. The randomness stems from player conditions, referee decisions, weather, and more, while the drift represents the higher probability of the better team winning. He notes that the complexity arising from 11 players on each side increases the randomness compared to sports like tennis. Why it matters: This perspective highlights the interplay of chance and skill in sports, offering a mathematical lens for understanding game dynamics.
Dr. Yali Du from King's College London will give a presentation on learning to cooperate in multi-agent systems. Her research focuses on enabling cooperative and responsible behavior in machines using reinforcement learning and foundation models. She will discuss enhancing collaboration within social contexts, fostering human-AI coordination, and achieving scalable alignment. Why it matters: This highlights the growing importance of research into multi-agent systems and human-AI interaction, crucial for developing AI that integrates effectively and ethically into society.