[论文] FRACTAL: A Self-Similar Multi-Agent Orchestration Architecture with...

论文概要

研究领域: ML 作者: Ohad Dan, Abhinav Taliyan 发布时间: 2026-10-19 arXiv: 2610.22561

中文摘要

本文提出 FRACTAL：一种自相似多智能体编排架构，同一五组件组织模式——编排器、集成组、知识账本、决策者、元观察者——在分解树的每一层递归复现。四项形式化贡献：其一，通信成本上界——在两阶段「盲提案/压力测试」协议下，节点内成本关于集成规模 k 与轮数 r 线性（O(kr)），而自由辩论为二次成本（基于评估器数量固定的显式假设）；其二，基于可测的多样性-饱和阈值区分生成阶段与评估阶段的两模态编排准则；其三，超多数收敛保证（含显式概率界），以及在可操作化目标复杂度度量下的形式深度界；其四，贝叶斯层间升级机制，把迟滞与违约信号统一在一次后验更新下。论文仅以算例与多层案例定性展示架构执行，不报告量化性能主张，并明确讨论模型形式与归纳偏置的局限，包括一个尚未实现但架构兼容的「有界自我改进」方向。文中「分形」是自相似的建筑隐喻，而非实测的数学性质；架构比较性能的实证验证留待未来工作。

原文摘要

We present FRACTAL, a self-similar multi-agent architecture for recursive goal decomposition, in which the same five-component organizational pattern -- Orchestrator, Ensemble, Knowledge Ledger, Decision Maker, Meta-Observer -- recurs at every level of a decomposition tree. We make four formal contributions: (1) a communication bound showing intra-node cost is linear in ensemble size and round count (O(kr)) under a two-phase blind-proposal/stress-test protocol, versus quadratic cost for free-form debate, with an explicit, stated assumption on fixed evaluator count; (2) a two-mode orchestration criterion separating generative and evaluative phases via a measurable diversity-saturation threshold; (3) a supermajority convergence guarantee with explicit probability bounds at stated agreement levels, and a formal depth bound under an operationalized goal-complexity measure; (4) a Bayesian inter-layer escalation mechanism unifying hysteresis and contract-breach signaling under one posterior update. We illustrate the architecture's execution qualitatively on a worked example and a multi-layer case study; we do not report quantitative performance claims in this paper, and we explicitly discuss the formal and inductive-bias limitations of the model, including an unimplemented but architecture-compatible direction toward bounded self-improvement. We use "fractal" as an architectural metaphor for self-similarity, not a measured mathematical property. Empirical validation of the architecture's comparative performance is left to future work.

---

*自动采集于 2026-10-19*

#论文 #arXiv #ML #小凯