Detecting Hierarchical Changes in Latent Variable Models

被引:2
|
作者
Fukushima, Shintaro [1 ]
Yamanishi, Kenji [1 ]
机构
[1] Univ Tokyo, Grad Sch Informat Sci & Technol, Tokyo, Japan
关键词
Change detection; Latent variable model; Hierarchical change detection; Data stream; Minimum description length principle;
D O I
10.1109/ICDM50108.2020.00120
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
This paper addresses the issue of detecting hierarchical changes in latent variable models (HCDL) from data streams. There are three different levels of changes for latent variable models: 1) the first level is the change in data distribution for fixed latent variables, 2) the second one is that in the distribution over latent variables, and 3) the third one is that in the number of latent variables. It is important to detect these changes because we can analyze the causes of changes by identifying which level a change comes from (change interpretability). This paper proposes an information-theoretic framework for detecting changes of the three levels in a hierarchical way. The key idea to realize it is to employ the MDL (minimum description length) change statistics for measuring the degree of change, in combination with DNML (decomposed normalized maximum likelihood) code-length calculation. We give a theoretical basis for making reliable alarms for changes. Focusing on stochastic block models, we employ synthetic and benchmark datasets to empirically demonstrate the effectiveness of our framework in terms of change interpretability as well as change detection.
引用
收藏
页码:1028 / 1033
页数:6
相关论文
共 50 条
  • [1] USE OF LATENT VARIABLE MODELS FOR DETECTING DISCRIMINATION IN SALARIES
    MCFATTER, RM
    PSYCHOLOGICAL BULLETIN, 1987, 101 (01) : 120 - 125
  • [2] Understanding Masked Autoencoders via Hierarchical Latent Variable Models
    Kong, Lingjing
    Ma, Martin Q.
    Chen, Guangyi
    Xing, Eric P.
    Chi, Yuejie
    Morency, Louis-Philippe
    Zhang, Kun
    2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR, 2023, : 7918 - 7928
  • [3] Hierarchical clustering with discrete latent variable models and the integrated classification likelihood
    Come, Etienne
    Jouvin, Nicolas
    Latouche, Pierre
    Bouveyron, Charles
    ADVANCES IN DATA ANALYSIS AND CLASSIFICATION, 2021, 15 (04) : 957 - 986
  • [4] Hierarchical clustering with discrete latent variable models and the integrated classification likelihood
    Etienne Côme
    Nicolas Jouvin
    Pierre Latouche
    Charles Bouveyron
    Advances in Data Analysis and Classification, 2021, 15 : 957 - 986
  • [5] Latent variable models
    Bishop, CM
    LEARNING IN GRAPHICAL MODELS, 1998, 89 : 371 - 403
  • [6] Latent variable and latent structure models
    von Eye, A
    CONTEMPORARY PSYCHOLOGY-APA REVIEW OF BOOKS, 2004, 49 (02): : 204 - 204
  • [7] Latent variable and latent structure models
    Bunting, B
    BRITISH JOURNAL OF MATHEMATICAL & STATISTICAL PSYCHOLOGY, 2003, 56 : 184 - 185
  • [8] Latent variable models are network models
    Molenaar, Peter C. M.
    BEHAVIORAL AND BRAIN SCIENCES, 2010, 33 (2-3) : 166 - +
  • [9] Latent variable and latent structure models.
    Glaser, DN
    STRUCTURAL EQUATION MODELING-A MULTIDISCIPLINARY JOURNAL, 2003, 10 (01) : 165 - 174
  • [10] FrictionSegNet: Simultaneous Semantic Segmentation and Friction Estimation Using Hierarchical Latent Variable Models
    Otoofi, Mohammad
    Laine, Leo
    Henderson, Leon
    Midgley, William J. B.
    Justham, Laura
    Fleming, James
    IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, 2024, 25 (12) : 19785 - 19795