Who Deserves the Reward? SHARP: Shapley Credit-based Optimization for Multi-Agent System

Yanming Li, Xuelin Zhang, Wenjie Lu, Ziye Tang, Maodong Wu, Haotian Luo, Tongtong Wu, Zijie Peng, Hongze Mi, Yibo Feng, Naiqiang Tan, Chao Huang, Hong Chen, Li Shen · 2026

Read Who Deserves the Reward? SHARP: Shapley Credit-based Optimization for Multi-Agent System by Yanming Li, Xuelin Zhang, Wenjie Lu, Ziye Tang, Maodong Wu, Haotian Luo, Tongtong Wu, Zijie Peng, Hongze Mi, Yibo Feng, Naiqiang Tan, Chao Huang, Hong Chen, Li Shen (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-08335.txt.

Description

Teams of AI agents that use tools can split hard jobs, but training them is tricky because credit is hard to assign when the team wins or fails. Yanming Li and thirteen co-authors introduce SHARP, a framework that shares credit fairly using Shapley values, a classic tool for dividing rewards among contributors. Their 2026 arXiv paper targets multi-agent systems built on large language models.

SHARP blends three signals: a team-wide accuracy reward, a per-agent marginal-credit reward, and a bonus for efficient tool use, with advantages normalized across trajectory groups to steady training. Across real-world benchmarks it beats recent baselines by wide margins, improving matches by about 24 percent over single-agent setups and 14 percent over other multi-agent methods. Readers get a principled fix for one of the field's most stubborn training problems.

Catalog record

CREATORYanming Li, Xuelin Zhang, Wenjie Lu, Ziye Tang, Maodong Wu, Haotian Luo, Tongtong Wu, Zijie Peng, Hongze Mi, Yibo Feng, Naiqiang Tan, Chao Huang, Hong Chen, Li Shen
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE71 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-08335

Browse the full library · About the Tower of Babel