Learning Interpretable Representations with Informative Entanglements

被引:0
|
作者
Beyazit, Ege [1 ]
Tuncel, Doruk [2 ]
Yuan, Xu [1 ]
Tzeng, Nian-Feng [1 ]
Wu, Xindong [3 ]
机构
[1] Univ Louisiana Lafayette, Lafayette, LA 70504 USA
[2] Johannes Kepler Univ Linz, Linz, Austria
[3] Mininglamp Acad Sci, Beijing, Peoples R China
来源
PROCEEDINGS OF THE TWENTY-NINTH INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE | 2020年
基金
美国国家科学基金会;
关键词
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
Learning interpretable representations in an unsupervised setting is an important yet a challenging task. Existing unsupervised interpretable methods focus on extracting independent salient features from data. However they miss out the fact that the entanglement of salient features may also be informative. Acknowledging these entanglements can improve the interpretability, resulting in extraction of higher quality and a wider variety of salient features. In this paper, we propose a new method to enable Generative Adversarial Networks (GANs) to discover salient features that may be entangled in an informative manner, instead of extracting only disentangled features. Specifically, we propose a regularizer to punish the disagreement between the extracted feature interactions and a given dependency structure while training. We model these interactions using a Bayesian network, estimate the maximum likelihood parameters and calculate a negative likelihood score to measure the disagreement. Upon qualitatively and quantitatively evaluating the proposed method using both synthetic and real-world datasets, we show that our proposed regularizer guides GANs to learn representations with disentanglement scores competing with the state-of-the-art, while extracting a wider variety of salient features.
引用
收藏
页码:1970 / 1976
页数:7
相关论文
共 50 条
  • [31] Learning biologically-interpretable latent representations for gene expression data Pathway Activity Score Learning Algorithm
    Karagiannaki, Ioulia
    Gourlia, Krystallia
    Lagani, Vincenzo
    Pantazis, Yannis
    Tsamardinos, Ioannis
    MACHINE LEARNING, 2023, 112 (11) : 4257 - 4287
  • [32] ELLA: Empowering LLMs for Interpretable, Accurate and Informative Legal Advice
    Hu, Yutong
    Luo, Kangcheng
    Feng, Yansong
    PROCEEDINGS OF THE 62ND ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, VOL 3: SYSTEM DEMONSTRATIONS, 2024, : 374 - 387
  • [33] An accurate and interpretable deep learning model for environmental properties prediction using hybrid molecular representations
    Zhang, Jun
    Wang, Qin
    Su, Yang
    Jin, Saimeng
    Ren, Jingzheng
    Eden, Mario
    Shen, Weifeng
    AICHE JOURNAL, 2022, 68 (06)
  • [34] Learning Interpretable and Transferable Representations via Wavelet-Constrained Transformer for Industrial Acoustic Diagnosis
    Ren, Jiaxin
    Hu, Chenye
    Shang, Zuogang
    Li, Yasong
    Zhao, Zhibin
    Yan, Ruqiang
    IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, 2025, 74
  • [35] Statistical criticality arises in most informative representations
    Cubero, Ryan John
    Jo, Junghyo
    Marsili, Matteo
    Roudi, Yasser
    Song, Juyong
    JOURNAL OF STATISTICAL MECHANICS-THEORY AND EXPERIMENT, 2019,
  • [36] Informative shape representations for human action recognition
    Wang, Liang
    Suter, David
    18TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION, VOL 2, PROCEEDINGS, 2006, : 1266 - +
  • [37] Where and What? Examining Interpretable Disentangled Representations
    Zhu, Xinqi
    Xu, Chang
    Tao, Dacheng
    2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021, 2021, : 5857 - 5866
  • [38] Towards a classification of entanglements of Galois representations attached to elliptic curves
    Daniels, Harris B.
    Lozano-Robledo, Alvaro
    Morrow, Jackson S.
    REVISTA MATEMATICA IBEROAMERICANA, 2023, 39 (03) : 803 - 844
  • [39] Interpretable Machine Learning
    Chen V.
    Li J.
    Kim J.S.
    Plumb G.
    Talwalkar A.
    Queue, 2021, 19 (06): : 28 - 56
  • [40] The school textbook and learning: Entanglements and disentanglements
    Manghi Haquin, Dominique
    REVISTA SIGNOS, 2015, 48 (88): : 270 - 276