Incorporating Causal Graphical Prior Knowledge into Predictive Modeling via Simple Data Augmentation

被引:0
|
作者
Teshima, Takeshi [1 ,2 ]
Sugiyama, Masashi [1 ,2 ]
机构
[1] Univ Tokyo, Grad Sch Frontier Sci, Tokyo, Japan
[2] RIKEN, Wako, Saitama, Japan
来源
UNCERTAINTY IN ARTIFICIAL INTELLIGENCE, VOL 161 | 2021年 / 161卷
关键词
DISCOVERY; NETWORKS;
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Causal graphs (CGs) are compact representations of the knowledge of the data generating processes behind the data distributions. When a CG is available, e.g., from the domain knowledge, we can infer the conditional independence (CI) relations that should hold in the data distribution. However, it is not straightforward how to incorporate this knowledge into predictive modeling. In this work, we propose a model-agnostic data augmentation method that allows us to exploit the prior knowledge of the CI encoded in a CG for supervised machine learning. We theoretically justify the proposed method by providing an excess risk bound indicating that the proposed method suppresses overfitting by reducing the apparent complexity of the predictor hypothesis class. Using real-world data with CGs provided by domain experts, we experimentally show that the proposed method is effective in improving the prediction accuracy, especially in the small-data regime.
引用
收藏
页码:86 / 96
页数:11
相关论文
共 50 条
  • [41] Incorporating prior knowledge with simulation data to estimate PSF multipliers using Bayesian logistic regression
    Kim, Yochan
    Park, Jinkyun
    RELIABILITY ENGINEERING & SYSTEM SAFETY, 2019, 189 : 210 - 217
  • [42] Incorporating Prior Knowledge in Local Parametric Modeling for Frequency Response Measurements: Applied to Thermal/Mechanical Systems
    Evers, Enzo
    de Jager, Bram
    Oomen, Tom
    IEEE TRANSACTIONS ON CONTROL SYSTEMS TECHNOLOGY, 2022, 30 (01) : 142 - 152
  • [43] Incorporating prior shape knowledge via data-driven loss model to improve 3D liver segmentation in deep CNNs
    Mohagheghi, Saeed
    Foruzan, Amir Hossein
    INTERNATIONAL JOURNAL OF COMPUTER ASSISTED RADIOLOGY AND SURGERY, 2020, 15 (02) : 249 - 257
  • [44] Incorporating prior shape knowledge via data-driven loss model to improve 3D liver segmentation in deep CNNs
    Saeed Mohagheghi
    Amir Hossein Foruzan
    International Journal of Computer Assisted Radiology and Surgery, 2020, 15 : 249 - 257
  • [45] Incorporating Systems Structure in Data-Driven High-Dimensional Predictive Modeling
    Seabra dos Reis, Marco P.
    28TH EUROPEAN SYMPOSIUM ON COMPUTER AIDED PROCESS ENGINEERING, 2018, 43 : 1039 - 1044
  • [46] Calibrate: Frequency Estimation and Heavy Hitter Identification with Local Differential Privacy via Incorporating Prior Knowledge
    Jia, Jinyuan
    Gong, Neil Zhenqiang
    IEEE CONFERENCE ON COMPUTER COMMUNICATIONS (IEEE INFOCOM 2019), 2019, : 2008 - 2016
  • [47] Incorporating prior knowledge via volumetric deep residual network to optimize the reconstruction of sparsely sampled MRI
    Wu, Yan
    Ma, Yajun
    Capaldi, Dante Pietro
    Liu, Jing
    Zhao, Wei
    Du, Jiang
    Xing, Lei
    MAGNETIC RESONANCE IMAGING, 2020, 66 : 93 - 103
  • [48] Normalizing RNA-Sequencing Data by Modeling Hidden Covariates with Prior Knowledge
    Mostafavi, Sara
    Battle, Alexis
    Zhu, Xiaowei
    Urban, Alexander E.
    Levinson, Douglas
    Montgomery, Stephen B.
    Koller, Daphne
    PLOS ONE, 2013, 8 (07):
  • [49] Generic protein-ligand interaction scoring by integrating physical prior knowledge and data augmentation modelling
    Cao, Duanhua
    Chen, Geng
    Jiang, Jiaxin
    Yu, Jie
    Zhang, Runze
    Chen, Mingan
    Zhang, Wei
    Chen, Lifan
    Zhong, Feisheng
    Zhang, Yingying
    Lu, Chenghao
    Li, Xutong
    Luo, Xiaomin
    Zhang, Sulin
    Zheng, Mingyue
    NATURE MACHINE INTELLIGENCE, 2024, : 688 - +
  • [50] Contrastive Curriculum Learning for Sequential User Behavior Modeling via Data Augmentation
    Bian, Shuqing
    Zhao, Wayne Xin
    Zhou, Kun
    Cai, Jing
    He, Yancheng
    Yin, Cunxiang
    Wen, Ji-Rong
    PROCEEDINGS OF THE 30TH ACM INTERNATIONAL CONFERENCE ON INFORMATION & KNOWLEDGE MANAGEMENT, CIKM 2021, 2021, : 3737 - 3746