Operator inference with roll outs for learning reduced models from scarce and low-quality data

被引:9
|
作者
Uy, Wayne Isaac Tan [1 ]
Hartmann, Dirk [2 ]
Peherstorfer, Benjamin [1 ]
机构
[1] NYU, Courant Inst Math Sci, 251 Mercer St, New York, NY 10012 USA
[2] Siemens Ind Software GmbH, Munich, Germany
基金
美国国家科学基金会;
关键词
Data-driven modeling; Scientific machine learning; Model reduction; Scarce and noisy data; EIGENSYSTEM REALIZATION-ALGORITHM; ORDER REDUCTION; DYNAMICS; TIME; APPROXIMATION; INTERPOLATION; IDENTIFICATION; DECOMPOSITION; FRAMEWORK; SYSTEMS;
D O I
10.1016/j.camwa.2023.06.012
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
Data-driven modeling has become a key building block in computational science and engineering. However, data that are available in science and engineering are typically scarce, often polluted with noise and affected by measurement errors and other perturbations, which makes learning the dynamics of systems challenging. In this work, we propose to combine data-driven modeling via operator inference with the dynamic training via roll outs of neural ordinary differential equations. Operator inference with roll outs inherits interpretability, scalability, and structure preservation of traditional operator inference while leveraging the dynamic training via roll outs over multiple time steps to increase stability and robustness for learning from low-quality and noisy data. Numerical experiments with data describing shallow water waves and surface quasi-geostrophic dynamics demonstrate that operator inference with roll outs provides predictive models from training trajectories even if data are sampled sparsely in time and polluted with noise of up to 10%.
引用
收藏
页码:224 / 239
页数:16
相关论文
共 50 条
  • [21] The Impacts of Low-Quality Training Data on Information Extraction from Clinical Reports
    Marcheggiani, Diego
    Sebastiani, Fabrizio
    ERCIM NEWS, 2018, (112): : 45 - 46
  • [22] Hamiltonian operator inference: Physics-preserving learning of reduced-order models for canonical Hamiltonian systems
    Sharma, Harsh
    Wang, Zhu
    Kramer, Boris
    PHYSICA D-NONLINEAR PHENOMENA, 2022, 431
  • [23] Gradient preserving Operator Inference: Data-driven reduced-order models for equations with gradient structure
    Geng, Yuwei
    Singh, Jasdeep
    Ju, Lili
    Kramer, Boris
    Wang, Zhu
    COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING, 2024, 427
  • [24] Robust and Lightweight Deep Learning Model for Industrial Fault Diagnosis in Low-Quality and Noisy Data
    Shin, Jaegwang
    Lee, Suan
    ELECTRONICS, 2023, 12 (02)
  • [25] A Hopfield Neural Network Based Algorithm for Haplotype Assembly from Low-quality Data
    Chen, Xiao
    Peng, Qinke
    Han, Libin
    Wang, Xiao
    PROCEEDINGS OF THE 2014 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN), 2014, : 1328 - 1333
  • [26] Intelligent fault diagnosis of rolling bearings with low-quality data: A feature significance and diversity learning method
    Chen, Jiayu
    Lin, Cuiyin
    Yao, Boqing
    Yang, Lechang
    Ge, Hongjuan
    RELIABILITY ENGINEERING & SYSTEM SAFETY, 2023, 237
  • [27] OmniFuse: A general modality fusion framework for multi-modality learning on low-quality medical data
    Wu, Yixuan
    Chen, Jintai
    Hu, Lianting
    Xu, Hongxia
    Liang, Huiying
    Wu, Jian
    INFORMATION FUSION, 2025, 117
  • [28] NONINTRUSIVE REDUCED-ORDER MODELS FOR PARAMETRIC PARTIAL DIFFERENTIAL EQUATIONS VIA DATA-DRIVEN OPERATOR INFERENCE
    Mcquarrie, Shane A.
    Khodabakhshi, Parisa
    Willcox, Karen E.
    SIAM JOURNAL ON SCIENTIFIC COMPUTING, 2023, 45 (04): : A1917 - A1946
  • [29] Low-Quality Sensor Data-Based Semi-Supervised Learning for Medical Image Segmentation
    Li, Hengfan
    Xu, Xuanbo
    Liu, Ziheng
    Xia, Qingfeng
    Xia, Min
    SENSORS, 2024, 24 (23)
  • [30] Two-Stage Enhancement Scheme for Low-Quality Fingerprint Images by Learning From the Images
    Yang, Jucheng
    Xiong, Naixue
    Vasilakos, Athanasios V.
    IEEE TRANSACTIONS ON HUMAN-MACHINE SYSTEMS, 2013, 43 (02) : 235 - 248