The recent publications in this research area highlight significant advancements in algorithmic fairness, resource allocation, and online learning. A common theme across these studies is the development of mechanisms and algorithms that ensure fairness, efficiency, and truthfulness in various settings, from healthcare to recommendation systems. Innovations include the introduction of new allocation rules for cooperative games, improvements in the approximation ratios for equitable allocations, and novel approaches to handling private contexts in bandit games. Additionally, there's a notable focus on optimizing resource allocation under complex constraints and enhancing online learning algorithms to achieve better regret bounds and gradient equilibrium.

Noteworthy Papers

• Truthful mechanisms for linear bandit games with private contexts: Introduces a mechanism ensuring truthful reporting in contextual bandit games, achieving an $O(\ln T)$ frequentist regret.
• Approximately EFX and PO Allocations for Bivalued Chores: Improves the approximation ratio for EFX and PO allocations in bi-valued chores to $(2-1/k)$.
• A New Value for Cooperative Games on Intersection-Closed Systems: Proposes the uniform-dividend value (UD-value) for cooperative games, offering a unique allocation rule for intersection-closed systems.
• Finite-Horizon Single-Pull Restless Bandits: Introduces a novel variant of RMABs with a single-pull constraint and proposes an efficient index policy for scarce resource allocation.
• Resource Allocation under the Latin Square Constraint: Addresses the challenge of allocating items under the Latin square constraint, providing approximation algorithms and FPT algorithms for maximizing social welfare.
• Equitable Allocations of Mixtures of Goods and Chores: Explores the complexity of achieving EQ1 allocations in settings with mixtures of goods and chores, offering efficient algorithms for specific valuation types.
• Improved Regret Bounds for Online Fair Division with Bandit Learning: Presents a UCB algorithm that achieves $\tilde{O}(\sqrt{T})$ regret while maintaining proportionality in online fair division.
• Gradient Equilibrium in Online Learning: Theory and Applications: Introduces the concept of gradient equilibrium in online learning, demonstrating its applicability in debiasing predictions and calibrating quantiles under distribution shift.