arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2411.01720cs.GTcs.LG
收藏

无遗憾学习下福利最大化的障碍

Barriers to Welfare Maximization with No-Regret Learning

Yale University(耶鲁大学);Strategy Robot, Inc.(策略机器人公司);Strategic Machine, Inc.(战略机器公司);Optimized Markets, Inc.(优化市场公司);Carnegie Mellon University(卡内基梅隆大学)

Ioannis Anagnostides, Alkis Kalavasis, Tuomas Sandholm

AI总结:

本文针对无遗憾学习达到逼近最优社会福利的粗相关均衡所需迭代次数,首次给出两人一般和博弈的计算下界,通过扩展经典归约并利用最大团不可近似性与植入团猜想,证明即使在集中式计算中也无法在多项式时间内获得非平凡稀疏性。

摘要:

在线学习与博弈论交叉领域的一项著名结果保证,无遗憾玩家的重复互动会导致粗相关均衡(CCE)——一种自然的博弈论解概念。尽管这一基础问题历史悠久且近年来受到极大关注,一个基本问题仍然悬而未决:无遗憾玩家需要多少次迭代才能逼近均衡?在本文中,我们在两人(一般和)博弈中,在CCE达到逼近最优社会福利(或其他自然目标)的约束下,首次为该问题建立了计算下界。从技术角度看,我们的方法围绕证明计算近最优的$T$-稀疏CCE(即$T$个乘积分布的混合)的下界展开,从而即使在集中式计算模型中也限定了无遗憾学习的迭代复杂度。我们的证明通过扩展Gilboa和Zemel [1989]关于最优纳什均衡到稀疏(近似)CCE的经典归约来进行。特别地,我们表明最大团问题的不可近似性排除了在多项式时间内获得任何非平凡稀疏性的可能性。此外,我们通过植入团猜想将我们的硬度结果加强到适用于低精度情形。

英文摘要:

A celebrated result in the interface of online learning and game theory guarantees that the repeated interaction of no-regret players leads to a coarse correlated equilibrium (CCE) -- a natural game-theoretic solution concept. Despite the rich history of this foundational problem and the tremendous interest it has received in recent years, a basic question still remains open: how many iterations are needed for no-regret players to approximate an equilibrium? In this paper, we establish the first computational lower bounds for that problem in two-player (general-sum) games under the constraint that the CCE reached approximates the optimal social welfare (or some other natural objective). From a technical standpoint, our approach revolves around proving lower bounds for computing a near-optimal $T$-sparse CCE -- a mixture of $T$ product distributions, thereby circumscribing the iteration complexity of no-regret learning even in the centralized model of computation. Our proof proceeds by extending a classical reduction of Gilboa and Zemel [1989] for optimal Nash to sparse (approximate) CCE. In particular, we show that the inapproximability of maximum clique precludes attaining any non-trivial sparsity in polynomial time. Moreover, we strengthen our hardness results to apply in the low-precision regime as well via the planted clique conjecture.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑