---
id: 20260603-T0-14
title: "多智能体博弈新方案：延迟奖励归因"
title_en: "Delayed Reward Attribution for Multi-Agent Games"
url: https://ai.daily.yangsir.net/daily/20260603-T0-14
issue_date: 2026-06-03
publish_date: 2026-06-02T04:00:00.000Z
category: research
source_name: "arXiv cs.AI"
source_url: https://arxiv.org/abs/2606.00017
---

# 多智能体博弈新方案：延迟奖励归因

研究者提出TIGER方案，通过图证据路由减少多模态生成幻觉。该方法在MindGames Arena泛化赛道中表现优异，解决多智能体博弈中的未来依赖问题。测试显示事实准确率提升35%，已开源代码库。

## English Version

**Delayed Reward Attribution for Multi-Agent Games**

TIGER introduces graph-based evidence routing to reduce hallucinations in multimodal generation. It excels in multi-agent strategic games by handling future dependencies and rule violations. Shows 35% fact accuracy improvement with open-source code.

---

**来源**：[arXiv cs.AI](https://arxiv.org/abs/2606.00017)

**详情页**：https://ai.daily.yangsir.net/daily/20260603-T0-14

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*