---
id: 20260722-T0-06
title: "Shapley Context Pruning：用博弈论为RAG系统做上下文重排序与剪枝"
title_en: "Shapley Context Pruning: Game Theory Approach for RAG Reranking and Trimming"
url: https://ai.daily.yangsir.net/daily/20260722-T0-06
issue_date: 2026-07-22
publish_date: 2026-07-21T04:00:00.000Z
category: research
source_name: "arXiv cs.AI"
source_url: https://arxiv.org/abs/2607.16209
---

# Shapley Context Pruning：用博弈论为RAG系统做上下文重排序与剪枝

针对RAG系统中上下文冗余与效率问题，新研究提出Shapley Context Pruning方法。该方法将上下文重排序与剪枝问题建模为合作博弈，利用Shapley值评估每个上下文片段对最终生成的贡献，从而去除低价值内容。相比传统纯基于词法重排序的方法，新框架更具可解释性与统一性。

## English Version

**Shapley Context Pruning: Game Theory Approach for RAG Reranking and Trimming**

To improve efficiency in RAG systems, researchers propose Shapley Context Pruning, a framework that models context reranking and pruning as a cooperative game. It uses Shapley values to evaluate each text segment's contribution to the final output, discarding low-value content. This provides an interpretable and unified alternative to traditional lexical reranking.

---

**来源**：[arXiv cs.AI](https://arxiv.org/abs/2607.16209)

**详情页**：https://ai.daily.yangsir.net/daily/20260722-T0-06

---

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