跳到正文
返回论文列表
cs.AI提交于 已译

内部共识,外部竞争:去中心化 LLM 团队用于周界防御

Consensus Within, Competition Between: Decentralized LLM Teams for Perimeter Defense

Khawaja Abaid Ullah · Ananya Acharya · Mohammad Javad Khojasteh

中文摘要

用大语言模型(LLM)协调多智能体团队时,一个根本性的架构问题是:任务分配应集中在一个 LLM 控制器,还是分散到自主 LLM 智能体,尤其在通信不可靠时。在一个刻意简化的周界防御模拟中,对抗由 LLM 控制的入侵者,比较了中央指挥官、基于拍卖的共识团队和独立智能体在五种网络条件下的表现。共识团队的突破率为 15%,指挥官为 32%,独立智能体为 29%;在不同网络条件中,只有有限的无线电范围改变了团队的突破率,且提高了指挥官的突破率。在共识团队中,让低成本 LLM 在投标计算中拥有更多权限,使突破率上升 38 个百分点。当空域中出现鸟类和民用无人机时,在输入文本状态中为每种添加一张照片。只有接近前沿的模型能将它们与入侵者区分开。伪装成现有标本鸟无人机的入侵者被低成本 LLM 和接近前沿的 LLM 都误认为鸟类,在接近前沿模型的决策中有 59-61% 未被追击(此前为 1-16%)。然而,在提示中加入一句预警后,这一比例降至 0-25%。发现,拥有专门外部冲突解决协议的 LLM 团队突破次数最少。代码:https://github.com/the-kas-lab/consensus-within-competition-between。

关键要点

  1. 01比较集中式指挥官、拍卖共识团队与独立智能体在周界防御中的突破率。
  2. 02共识团队突破率 15%,低于指挥官的 32% 和独立智能体的 29%。
  3. 03低成本 LLM 在共识团队中掌握更多投标权限,使突破率上升 38 个百分点。
  4. 04伪装成标本鸟无人机的入侵者被 LLM 误判为鸟类,接近前沿模型 59-61% 未追击。
  5. 05提示中加入一句预警可将误判未追击率降至 0-25%。

解读

尚无解读。

原始英文摘要

arXiv:2610.41044v1 Announce Type: cross Abstract: The use of large language models (LLMs) to coordinate multi-agent teams raises a fundamental architectural question: whether task allocation should be centralized in a single LLM controller or distributed across autonomous LLM agents, particularly when communication is unreliable. In a deliberately simplified perimeter-defense simulation against LLM-controlled intruders, we compared a central commander, an auction-based consensus team and independent agents under five network conditions. Breach rates were 15% for the consensus team, 32% for the commander and 29% for independent agents, and of the different network conditions, only limited radio range changed a team's breach rate, raising the commander's. In consensus team, giving a low-cost LLM more authority over the bid computation raised breaches by 38 percentage points. With birds and civilian drones in airspace, we added a photograph for each to the input text state. Only the near-frontier model treated them apart from intruders. Intruders disguised as existing taxidermy bird drones were mistook as birds by both low-cost and near-frontier LLMs, going unpursued in 59-61% of near-frontier decisions (1-16% before). However, a one-sentence forewarning in the prompt lowered it to 0-25%. We find that LLM team with a dedicated external conflict resolution protocol results in fewest breaches. Code: https://github.com/the-kas-lab/consensus-within-competition-between.

同方向论文 · cs.AI

查看全部 →

反向发现 · 这篇论文相关的工具与提示词

去交叉发现 →

关联工具 · 1

Google DeepMind89

Google DeepMind 发布 Gemini 4 Carbon 前沿模型

Google DeepMind 向开发者、企业和消费者开放 Gemini 4 Carbon,称其在网络安全与长程智能体工作流上达到前沿水平。

推荐理由:官方披露了 Carbon 在编码、网络安全与科研推理上的基准表现和定价,可据此判断它与 Argon 的分工。