Search papers, labs, and topics across Lattice.
This paper introduces Gradient-Based Connections (GBC), a novel method for optimizing multi-agent systems (MAS) by providing fine-grained credit assignment through gradient-based connection weights. By modeling MAS as a computational graph and propagating task-specific loss signals, GBC allows for precise identification of error sources and targeted optimization of prompts. Experiments demonstrate that GBC significantly enhances multi-agent performance on benchmarks like MultiWOZ and 蟿-bench, outperforming both single-agent and existing multi-agent baselines, with improved attribution quality correlating with optimization success.
Fine-grained credit assignment in multi-agent systems can dramatically boost performance, revealing error sources with unprecedented precision.
Multi-agent systems (MAS) built on large language models (LLMs) provide a promising framework for solving complex tasks through role specialization and structured interaction. However, their performance is often limited by miscoordination and, more fundamentally, the lack of fine-grained credit assignment across agents. Existing approaches typically rely on coarse-grained feedback, making it difficult to identify which agents or interaction steps are responsible for errors. We propose Gradient-Based Connections (GBC), an approach for fine-grained attribution and optimization of multi-agent systems. GBC models a MAS as a computational graph and introduces gradient-based connection weights to quantify the influence of each agent's output on downstream agents at the token level. By constructing an attribution graph and propagating task-specific loss signals backward, our method enables precise identification of error sources and targeted prompt optimization. We further develop AgentChord, an efficient implementation that leverages prefix-based gradient computation. Experiments on MultiWOZ and 蟿-bench show that GBC improves multi-agent performance and outperforms strong single-agent and multi-agent baselines, and higher attribution quality is associated with greater optimization effectiveness. Code is available at: https://github.com/yxc-cyber/AgentChord.
