Understanding the Information Propagation Effects of Communication Topologies in LLM-based Multi-Agent Systems
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866909627033059328 |
|---|---|
| author | Shen, Xu Liu, Yixin Dai, Yiwei Wang, Yili Miao, Rui Tan, Yue Pan, Shirui Wang, Xin |
| author_facet | Shen, Xu Liu, Yixin Dai, Yiwei Wang, Yili Miao, Rui Tan, Yue Pan, Shirui Wang, Xin |
| contents | The communication topology in large language model-based multi-agent systems fundamentally governs inter-agent collaboration patterns, critically shaping both the efficiency and effectiveness of collective decision-making. While recent studies for communication topology automated design tend to construct sparse structures for efficiency, they often overlook why and when sparse and dense topologies help or hinder collaboration. In this paper, we present a causal framework to analyze how agent outputs, whether correct or erroneous, propagate under topologies with varying sparsity. Our empirical studies reveal that moderately sparse topologies, which effectively suppress error propagation while preserving beneficial information diffusion, typically achieve optimal task performance. Guided by this insight, we propose a novel topology design approach, EIB-leanrner, that balances error suppression and beneficial information propagation by fusing connectivity patterns from both dense and sparse graphs. Extensive experiments show the superior effectiveness, communication cost, and robustness of EIB-leanrner. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_23352 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Understanding the Information Propagation Effects of Communication Topologies in LLM-based Multi-Agent Systems Shen, Xu Liu, Yixin Dai, Yiwei Wang, Yili Miao, Rui Tan, Yue Pan, Shirui Wang, Xin Multiagent Systems Artificial Intelligence The communication topology in large language model-based multi-agent systems fundamentally governs inter-agent collaboration patterns, critically shaping both the efficiency and effectiveness of collective decision-making. While recent studies for communication topology automated design tend to construct sparse structures for efficiency, they often overlook why and when sparse and dense topologies help or hinder collaboration. In this paper, we present a causal framework to analyze how agent outputs, whether correct or erroneous, propagate under topologies with varying sparsity. Our empirical studies reveal that moderately sparse topologies, which effectively suppress error propagation while preserving beneficial information diffusion, typically achieve optimal task performance. Guided by this insight, we propose a novel topology design approach, EIB-leanrner, that balances error suppression and beneficial information propagation by fusing connectivity patterns from both dense and sparse graphs. Extensive experiments show the superior effectiveness, communication cost, and robustness of EIB-leanrner. |
| title | Understanding the Information Propagation Effects of Communication Topologies in LLM-based Multi-Agent Systems |
| topic | Multiagent Systems Artificial Intelligence |
| url | https://arxiv.org/abs/2505.23352 |