GBC: Gradient-Based Connections for Optimizing Multi-Agent Systems

  • 类型：arxiv
  • 标识：2606.28187
  • 链接：https://arxiv.org/abs/2606.28187
  • 主分类：agent
  • 形态：method
  • 被引：0
  • 被引来源：Semantic Scholar + OpenAlex
  • S2被引：0
  • OpenAlex被引：0
  • 影响力被引：0
  • TLDR：Gradient-Based Connections (GBC) is proposed, an approach for fine-grained attribution and optimization of multi-agent systems that improves multi-agent performance and outperforms strong single-agent and multi-agent baselines and higher attribution quality is associated with greater optimization effectiveness.
  • OpenAlex ID：W7166526419
  • OpenAlex DOI：10.48550/arxiv.2606.28187
  • DOI：10.48550/arxiv.2606.28187
  • DOI来源：OpenAlex
  • 开放获取：green
  • 开放获取链接：https://doi.org/10.48550/arxiv.2606.28187
  • OpenAlex更新：2026-07-19
  • 待LLM分类：否
  • 标题中文：GBC：用于多智能体系统优化的基于梯度的连接
  • TLDR中文：提出 Gradient-Based Connections（GBC）——一种面向多智能体系统的细粒度归因与优化方法，可提升多智能体性能，超越强力的单智能体与多智能体基线；且归因质量越高，优化效果越显著。
  • 来源文件：
  • /inbox/tom/_candidates/2026-06-29-agent-rag-longcontext-candidates.json
  • [S2 enrich]
  • [OpenAlex backfill]