研究人员开发了Onnes，一种用于诊断稀释制冷机故障的新型模拟器，稀释制冷机是量子计算的关键基础设施。这个基于物理的数字孪生整合了来自Bluefors日志的真实世界噪声指纹，并模拟了六种故障类别。当由多智能体LLM驱动时，Onnes在故障分类准确性方面表现出显著的提高，在少样本演示下达到0.990，并在实时模拟中有效检测了正在发生的故障。 AI
影响 这项研究展示了LLM在复杂科学基础设施中的一项新颖应用，有望提高量子计算运营的可靠性和效率。
排序理由 该集群描述了一篇详细介绍特定技术应用新模拟器的研究论文。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
arXiv:2607.05805v1 Announce Type: new Abstract: Dilution refrigerators are the enabling infrastructure of superconducting quantum computers, yet their fault diagnosis is still dominated by threshold alarms that report that something is wrong, not what. We present Onnes, a physics…
Dilution refrigerators are the enabling infrastructure of superconducting quantum computers, yet their fault diagnosis is still dominated by threshold alarms that report that something is wrong, not what. We present Onnes, a physics-grounded digital-twin simulator of a dilution r…