Complexity of Representations in Deep Learning

被引:1
|
作者
Ho, Tin Kam [1 ]
机构
[1] IBM Watson Hlth, Yorktown Hts, NY USA
关键词
Keywords-representation learning; deep learning; classification; data complexity;
D O I
10.1109/ICPR56361.2022.9956594
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Deep neural networks use multiple layers of functions to map an object represented by an input vector progressively to different representations, and with sufficient training, eventually to a single score for each class that is the output of the final decision function. Ideally, in this output space, the objects of different classes achieve maximum separation. Motivated by the need to better understand the inner working of a deep neural network, we analyze the effectiveness of the learned representations in separating the classes from a data complexity perspective. Using a simple complexity measure, a popular benchmarking task, and a well-known architecture design, we show how the data complexity evolves through the network, how it changes during training, and how it is impacted by the network design and the availability of training samples. We discuss the implications of the observations and the potentials for further studies.
引用
收藏
页码:2657 / 2663
页数:7
相关论文
共 50 条
  • [41] Images of chemical structures as molecular representations for deep learning
    Wilkinson, Matthew R.
    Martinez-Hernandez, Uriel
    Wilson, Chick C.
    Castro-Dominguez, Bernardo
    [J]. JOURNAL OF MATERIALS RESEARCH, 2022, 37 (14) : 2293 - 2303
  • [42] Deep learning from crystallographic representations of periodic systems
    Maffettone, Phillip
    Cooper, Andrew
    [J]. ABSTRACTS OF PAPERS OF THE AMERICAN CHEMICAL SOCIETY, 2019, 258
  • [43] Learning Deep Representations for Ground-to-Aerial Geolocalization
    Lin, Tsung-Yi
    Cui, Yin
    Belongie, Serge
    Hays, James
    [J]. 2015 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2015, : 5007 - 5015
  • [44] Visual Cortex and Deep Networks - Learning Invariant Representations
    Clarke, Alasdair
    [J]. PERCEPTION, 2018, 47 (03) : 355 - 356
  • [45] Learning social representations with deep autoencoder for recommender system
    Yiteng Pan
    Fazhi He
    Haiping Yu
    [J]. World Wide Web, 2020, 23 : 2259 - 2279
  • [46] Hippocampal representations for deep learning on Alzheimer’s disease
    Ignacio Sarasua
    Sebastian Pölsterl
    Christian Wachinger
    [J]. Scientific Reports, 12
  • [47] Joint Unsupervised Learning of Deep Representations and Image Clusters
    Yang, Jianwei
    Parikh, Devi
    Batra, Dhruv
    [J]. 2016 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2016, : 5147 - 5156
  • [48] Learning Deep Network Representations with Adversarially Regularized Autoencoders
    Yu, Wenchao
    Zheng, Cheng
    Cheng, Wei
    Aggarwal, Charu C.
    Song, Dongjin
    Zong, Bo
    Chen, Haifeng
    Wang, Wei
    [J]. KDD'18: PROCEEDINGS OF THE 24TH ACM SIGKDD INTERNATIONAL CONFERENCE ON KNOWLEDGE DISCOVERY & DATA MINING, 2018, : 2663 - 2671
  • [49] DISL: Deep Isomorphic Substructure Learning for network representations
    Cui, Shicheng
    Li, Tao
    Chen, Shu-Ching
    Shyu, Mei-Ling
    Li, Qianmu
    Zhang, Hong
    [J]. Knowledge-Based Systems, 2021, 189
  • [50] Hippocampal representations for deep learning on Alzheimer's disease
    Sarasua, Ignacio
    Polsterl, Sebastian
    Wachinger, Christian
    [J]. SCIENTIFIC REPORTS, 2022, 12 (01)