A Novel Convolutional Neural Network Architecture with a Continuous Symmetry

被引：0
|
作者：
Liu, Yao [1 ]
Shao, Hang [2 ]
Bai, Bing [1 ]
机构：
[1] Tsinghua Univ, Beijing, Peoples R China
[2] Zhejiang Future Technol Inst, Jiaxing, Peoples R China
来源：
ARTIFICIAL INTELLIGENCE, CICAI 2023, PT II | 2024年 / 14474卷
关键词：
Convolutional Neural Networks; Partial Differential Equations; Continuous Symmetry;
D O I：
10.1007/978-981-99-9119-8_28
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 introduces a new Convolutional Neural Network (ConvNet) architecture inspired by a class of partial differential equations (PDEs) called quasi-linear hyperbolic systems. With comparable performance on image classification task, it allows for the modification of the weights via a continuous group of symmetry. This is a significant shift from traditional models where the architecture and weights are essentially fixed. We wish to promote the (internal) symmetry as a new desirable property for a neural network, and to draw attention to the PDE perspective in analyzing and interpreting ConvNets in the broader Deep Learning community.
引用
收藏
页码：310 / 321
页数：12
相关论文
共 50 条
  • [31] Convolutional Neural Network Architecture for Recovering Watermark Synchronization
    Kim, Wook-Hyung
    Kang, Jihyeon
    Mun, Seung-Min
    Hou, Jong-Uk
    SENSORS, 2020, 20 (18) : 1 - 18
  • [32] A Reconfigurable Pipelined Architecture for Convolutional Neural Network Acceleration
    Xue, Chengbo
    Cao, Shan
    Jiang, Rongkun
    Yang, Hao
    2018 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS (ISCAS), 2018,
  • [33] Continuous Speech Recognition based on Convolutional Neural Network
    Zhang, Qing-qing
    Liu, Yong
    Pan, Jie-lin
    Yan, Yong-hong
    SEVENTH INTERNATIONAL CONFERENCE ON DIGITAL IMAGE PROCESSING (ICDIP 2015), 2015, 9631
  • [34] A novel convolutional neural network architecture of multispectral remote sensing images for automatic material classification
    Lin, Chuen-Horng
    Wang, Ting-You
    SIGNAL PROCESSING-IMAGE COMMUNICATION, 2021, 97
  • [35] A novel one-dimensional convolutional neural network architecture for chemical process fault diagnosis
    Niu, Xin
    Yang, Xia
    CANADIAN JOURNAL OF CHEMICAL ENGINEERING, 2022, 100 (02): : 302 - 316
  • [36] A Novel Architecture to verify Offline Hand-written Signatures using Convolutional Neural Network
    Alkaabi, Sultan
    Yussof, Salman
    Almulla, Sameera
    Al-Khateeb, Haider
    AAlAbdulsalam, Abdulrahman
    2019 INTERNATIONAL CONFERENCE ON INNOVATION AND INTELLIGENCE FOR INFORMATICS, COMPUTING, AND TECHNOLOGIES (3ICT), 2019,
  • [37] A Novel Convolutional Neural Network Architecture for Image Super-Resolution Based on Channels Combination
    Liu, Cun
    Li, Yuanxiang
    Luo, Jianhua
    Zhou, Yongjun
    2017 20TH INTERNATIONAL CONFERENCE ON INFORMATION FUSION (FUSION), 2017, : 1084 - 1091
  • [38] A novel granular neural network architecture
    Dick, S.
    Tappenden, A.
    Badke, Curtis
    Olarewaju, O.
    NAFIPS 2007 - 2007 ANNUAL MEETING OF THE NORTH AMERICAN FUZZY INFORMATION PROCESSING SOCIETY, 2007, : 42 - +
  • [39] A novel training algorithm for convolutional neural network
    Anuse, Alwin
    Vyas, Vibha
    COMPLEX & INTELLIGENT SYSTEMS, 2016, 2 (03) : 221 - 234
  • [40] A Novel Convolutional Neural Network for Statutes Recommendation
    Li, Chuanyi
    Ye, Jingjing
    Ge, Jidong
    Kong, Li
    Hu, Haiyang
    Luo, Bin
    PRICAI 2018: TRENDS IN ARTIFICIAL INTELLIGENCE, PT I, 2018, 11012 : 851 - 863
12345