Year: 2025
Type: conference-paper
Abstract: Recent progress in large language model (LLM)-based multi-agent collaboration highlights the power of structured communication in enabling collective intelligence.However, existing methods largely rely on static or graphbased inter-agent topologies, lacking the potential adaptability and flexibility in communication.In this work, we propose a new framework that rethinks multi-agent coordination through a sequential structure rather than a graph structure, offering a significantly larger topology space for multi-agent communication.Our method focuses on two key directions: (1) Next-Agent Prediction, which selects the most suitable agent role at each step, and (2) Next-Context Selection (NCS), which enables each agent to selectively access relevant information from any previous step.Together, these components construct task-adaptive communication pipelines that support both role flexibility and global information flow.Extensive evaluations across multiple benchmarks demonstrate that our approach achieves superior performance while substantially reducing communication overhead.
Authors: Song Wang, Zhen Tan, Zihan Chen, Shuang Zhou, Tianlong Chen, Jundong Li
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Awards: SCC-IRG Track 1: Community-Responsive Electrified and Adaptive Transit Ecosystem (CREATE): Planning, Operations, and Management, Collaborative Research: III: Small: Graph-Oriented Usable Interpretation, CAREER: Toward A Knowledge-Guided Framework for Personalized Decision Making, Collaborative Research: SAI-R: Dynamical Coupling of Physical and Social Infrastructures: Evaluating the Impacts of Social Capital on Access to Safe Well Water, SaTC: CORE: Small: Empowering Network Attack Detection with Complex Graph Modeling, III: Small: Collaborative Research: Demystifying Deep Learning on Graphs: From Basic Operations to Applications
OpenAlex ID: W4416036612
