arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2210.05955stat.MLcs.LG
收藏

从离散观测中学习齐次线性ODE系统的可识别性与渐近性

Identifiability and Asymptotics in Learning Homogeneous Linear ODE Systems from Discrete Observations

University of Melbourne(墨尔本大学);School of Mathematics and Statistics(数学与统计学院);The University of Sydney(悉尼大学);School of Computer Science(计算机科学学院);Carnegie Mellon University(卡内基梅隆大学);Mohamed bin Zayed University of Artificial Intelligence(穆罕默德·本·扎耶德人工智能大学)

Yuanyuan Wang, Wei Huang, Mingming Gong, Xi Geng, Tongliang Liu, Kun Zhang, Dacheng Tao

AI总结:

本文研究了从离散观测学习齐次线性ODE系统的理论,推导了可识别性充分条件,证明了NLS估计的一致性与渐近正态性,并提出了参数置信集构建与因果结构推断方法。

摘要:

常微分方程(ODE)近期在机器学习中备受关注。然而,诸如可识别性与统计估计的渐近性等理论问题仍不清晰。本文从单条轨迹中等间隔无误差观测序列中,推导出齐次线性ODE系统可识别性的充分条件。当观测受测量噪声干扰时,我们证明在温和条件下,基于非线性最小二乘(NLS)方法的参数估计具有一致性与渐近正态性,收敛率为 $n^{-1/2}$。基于渐近正态性,我们为未知系统参数构建置信集,并提出推断ODE系统因果结构的新方法,即推断系统变量间是否存在因果联系。此外,我们将结果扩展至退化观测,包括聚合与时间缩放观测。据我们所知,这是学习线性ODE系统可识别性与渐近性的首项系统研究。我们还构建了不同系统维度的仿真以阐明所建立的理论结果。

英文摘要:

Ordinary Differential Equations (ODEs) have recently gained a lot of attention in machine learning. However, the theoretical aspects, e.g., identifiability and asymptotic properties of statistical estimation are still obscure. This paper derives a sufficient condition for the identifiability of homogeneous linear ODE systems from a sequence of equally-spaced error-free observations sampled from a single trajectory. When observations are disturbed by measurement noise, we prove that under mild conditions, the parameter estimator based on the Nonlinear Least Squares (NLS) method is consistent and asymptotic normal with $n^{-1/2}$ convergence rate. Based on the asymptotic normality property, we construct confidence sets for the unknown system parameters and propose a new method to infer the causal structure of the ODE system, i.e., inferring whether there is a causal link between system variables. Furthermore, we extend the results to degraded observations, including aggregated and time-scaled ones. To the best of our knowledge, our work is the first systematic study of the identifiability and asymptotic properties in learning linear ODE systems. We also construct simulations with various system dimensions to illustrate the established theoretical results.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑