跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+197新增论文
+151其中 AI 相关
+16已有论文修订
+91新增 AI 导读

arXiv 最近完成扫描:2026-11-21 12:19 · 全站最近入库:2026-11-21 12:19各信源进度 →

统计口径与数据新鲜度

统计 2026-11-21 03:18 — 2026-11-22 03:18(北京时间) · 点击数字查看对应简报;AI 相关包含在新增论文内。

最近完成检查:2026-11-21 12:19最近发现新增版本:2026-11-21 12:19库内论文最新更新时间:2026-11-20 02:59

每分钟刷新已保存的数据。空检查不代表采集异常;新收录可能包含历史论文。AI 导读按新增版本计数。

29,158 已收录论文24,181 AI 相关8,992 有修订112 HF 社区发现

浏览论文库

时间范围跟随排序:版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常,查看状态 →

找到 1 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-21 12:17(北京时间)
新论文 · v1arXiv:2611.23614

Expanding and Evaluating the Applicability of Safe Pareto Improvements

Nathaniel Sauerberg, Caspar Oesterheld

原文摘要

Even when players can make credible joint commitments in games, disagreement about the game's default outcome (e.g., about equilibrium selection) can prevent them from agreeing on any particular payoff profile. We study safe Pareto improvements (SPIs): commitments that sidestep this obstacle by inducing a new game that is isomorphic to the original (after iteratively removing strictly dominated actions) but entrywise Pareto better. We do so by constructing token games: cheap-talk games rendered meaningful by commitments to take payoff-relevant actions (in prior work, strategies in the original game) as a function of the token outcome. Since these token games preserve the original game's strategic structure, players should expect them to be played isomorphically and hence agree that they Pareto improve on the original. In this work, we study token game SPIs under three additional commitment abilities: (1) burning utility, (2) transferring utility between players, and (3) transferring utility with an outside party conditional on the token outcome (e.g., via insurance-like contracts). We characterize when SPIs exist in each case and give efficient algorithms for optimizing over them. Finally, we propose evaluating SPI frameworks by how often they enable SPIs in randomly generated games. In such experiments, our methods substantially outperform prior work.

cs.GT