用归一化流建模数据似然，支持高能物理实验结果的可靠传递。
Communicating Likelihoods with Normalising Flows
- 基于归一化流从样本中学习未分箱似然分布
- 通过联合分布的柯尔莫哥洛夫-斯米尔诺夫检验验证模型可靠性
- 适用于高能物理中实验与理论似然的可复现传递
我们提出一种基于机器学习的工作流程，从样本中建模未分箱的似然分布。相比现有方法，关键进步在于使用严格的统计检验（如联合分布的柯尔莫哥洛夫-斯米尔诺夫检验）验证所学似然的准确性。该方法确保了实验和现象学似然在后续分析中的可靠传递。我们在高能物理领域通过三个案例研究验证了其有效性。为促进广泛应用，我们提供了开源参考实现 nabu。
原文摘要 · Abstract (English)
We present a machine-learning-based workflow to model an unbinned likelihood from its samples. A key advancement over existing approaches is the validation of the learned likelihood using rigorous statistical tests of the joint distribution, such as the Kolmogorov-Smirnov test of the joint distribution. Our method enables the reliable communication of experimental and phenomenological likelihoods for subsequent analyses. We demonstrate its effectiveness through three case studies in high-energy physics. To support broader adoption, we provide an open-source reference implementation, nabu.
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。