通过动态优化减少大模型多智能体通信冗余，提升效率与安全性
SafeSieve: From Heuristics to Experience in Progressive Pruning for LLM-based Multi-Agent Communication
- 结合语义评估与反馈经验，渐进式剪枝智能体间通信
- 在多个基准上实现94.01%准确率，令牌消耗降低12.4%-27.8%
- 抗提示注入攻击，适用于资源受限的异构部署场景
基于大模型的多智能体系统具备强大协作能力，但常面临通信冗余和令牌开销过大的问题。现有方法多依赖预训练图神经网络或贪心算法，往往割裂任务前后的优化过程，缺乏统一策略。为此，我们提出SafeSieve，一种渐进式自适应多智能体剪枝算法，通过创新的双机制动态优化智能体间通信。该方法融合初始大模型语义评估与累积性能反馈，实现从启发式初始化到经验驱动优化的平滑过渡。不同于传统贪心Top-k剪枝，SafeSieve采用0-扩展聚类，保留结构连贯的智能体群组，同时剔除无效连接。在SVAMP、HumanEval等基准上的实验表明，SafeSieve平均准确率达94.01%，令牌使用量降低12.4%-27.8%。结果还显示其对提示注入攻击具有鲁棒性（平均准确率仅下降1.23%）。在异构环境下，部署成本降低13.3%的同时保持性能。这些成果验证了SafeSieve作为高效、无需GPU、可扩展的实际多智能体框架的潜力。
原文摘要 · Abstract (English)
LLM-based multi-agent systems exhibit strong collaborative capabilities but often suffer from redundant communication and excessive token overhead. Existing methods typically enhance efficiency through pretrained GNNs or greedy algorithms, but often isolate pre- and post-task optimization, lacking a unified strategy. To this end, we present SafeSieve, a progressive and adaptive multi-agent pruning algorithm that dynamically refines the inter-agent communication through a novel dual-mechanism. SafeSieve integrates initial LLM-based semantic evaluation with accumulated performance feedback, enabling a smooth transition from heuristic initialization to experience-driven refinement. Unlike existing greedy Top-k pruning methods, SafeSieve employs 0-extension clustering to preserve structurally coherent agent groups while eliminating ineffective links. Experiments across benchmarks (SVAMP, HumanEval, etc.) showcase that SafeSieve achieves 94.01% average accuracy while reducing token usage by 12.4%-27.8%. Results further demonstrate robustness under prompt injection attacks (1.23% average accuracy drop). In heterogeneous settings, SafeSieve reduces deployment costs by 13.3% while maintaining performance. These results establish SafeSieve as an efficient, GPU-free, and scalable framework for practical multi-agent systems. Our code can be found here: https://github.com/csgen/SafeSieve
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。