IMPROVING ROBUSTNESS TO OUT-OF-DISTRIBUTION DATA BY FREQUENCY-BASED AUGMENTATION

被引:6
|
作者
Mukai, Koki [1 ]
Kumano, Soichiro [1 ]
Yamasaki, Toshihiko [1 ]
机构
[1] Univ Tokyo, Tokyo, Japan
关键词
neural network; out-of-distribution; frequency; data augmentation;
D O I
10.1109/ICIP46576.2022.9897504
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Although Convolutional Neural Networks (CNNs) have high accuracy in image recognition, they are vulnerable to adversarial examples and out-of-distribution data, and the difference from human recognition has been pointed out. In order to improve the robustness against out-of-distribution data, we present a frequency-based data augmentation technique that replaces the frequency components with other images of the same class. When the training data are CIFAR10 and the out-of-distribution data are SVHN, the Area Under Receiver Operating Characteristic (AUROC) curve of the model trained with the proposed method increases from 89.22% to 98.15%, and further increased to 98.59% when combined with another data augmentation method. Furthermore, we experimentally demonstrate that the robust model for out-of-distribution data uses a lot of high-frequency components of the image.
引用
收藏
页码:3116 / 3120
页数:5
相关论文
共 50 条
  • [21] Improving out-of-distribution detection by enforcing confidence margin
    Tamang, Lakpa
    Bouadjenek, Mohamed Reda
    Dazeley, Richard
    Aryal, Sunil
    KNOWLEDGE AND INFORMATION SYSTEMS, 2025,
  • [22] Targeted Data Augmentation for Improving Model Robustness
    Mikolajczyk-Barela, Agnieszka
    Ferlin, Maria
    Grochowski, Michal
    INTERNATIONAL JOURNAL OF APPLIED MATHEMATICS AND COMPUTER SCIENCE, 2025, 35 (01) : 143 - 155
  • [23] Understanding and Improving Feature Learning for Out-of-Distribution Generalization
    Chen, Yongqiang
    Huang, Wei
    Zhou, Kaiwen
    Bian, Yatao
    Han, Bo
    Cheng, James
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 36 (NEURIPS 2023), 2023,
  • [24] LEARNING WITH OUT-OF-DISTRIBUTION DATA FOR AUDIO CLASSIFICATION
    Iqbal, Turab
    Cao, Yin
    Kong, Qiuqiang
    Plumbley, Mark D.
    Wang, Wenwu
    2020 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, 2020, : 636 - 640
  • [25] Understanding Out-of-distribution:A Perspective of Data Dynamics
    Adila, Dyah
    Kang, Dongyeop
    WORKSHOP AT NEURIPS 2021, VOL 163, 2021, 163 : 1 - 8
  • [26] Revisiting Out-of-distribution Robustness in NLP: Benchmark, Analysis, and LLMs Evaluations
    Yuan, Lifan
    Chen, Yangyi
    Cui, Ganqu
    Gao, Hongcheng
    Zou, Fangyuan
    Cheng, Xingyi
    Ji, Heng
    Liu, Zhiyuan
    Sun, Maosong
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 36 (NEURIPS 2023), 2023,
  • [27] PAC-Based Formal Verification for Out-of-Distribution Data Detection
    Prashant, Mohit
    Easwaran, Arvind
    2022 6TH INTERNATIONAL CONFERENCE ON SYSTEM RELIABILITY AND SAFETY, ICSRS, 2022, : 300 - 309
  • [28] Out-of-Distribution Detection in Hand Gesture Recognition Using Image Augmentation
    Lee, Hyeonji
    Yu, Yeonguk
    Lee, Kyoobin
    INTELLIGENT AUTONOMOUS SYSTEMS 18, VOL 1, IAS18-2023, 2024, 795 : 595 - 605
  • [29] Improving Out-of-Distribution Generalization by Adversarial Training with Structured Priors
    Wang, Qixun
    Wang, Yifei
    Zhu, Hong
    Wang, Yisen
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 35 (NEURIPS 2022), 2022,
  • [30] Improving Normalizing Flows with the Approximate Mass for Out-of-Distribution Detection
    Chali, Samy
    Kucher, Inna
    Duranton, Marc
    Klein, Jacques-Olivier
    2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS, CVPRW, 2023, : 750 - 758