Repairing Deep Neural Networks: Fix Patterns and Challenges

被引：81
|
作者：
Islam, Md Johirul [1 ]
Pan, Rangeet [1 ]
Giang Nguyen [1 ]
Rajan, Hridesh [1 ]
机构：
[1] Iowa State Univ, Dept Comp Sci, 226 Atanasoff Hall, Ames, IA 50011 USA
来源：
2020 ACM/IEEE 42ND INTERNATIONAL CONFERENCE ON SOFTWARE ENGINEERING (ICSE 2020) | 2020年
关键词：
deep neural networks; bugs; bug fix; bug fix patterns;
D O I：
10.1145/3377811.3380378
中图分类号：
TP31 [计算机软件];
学科分类号：
081202 ; 0835 ;
摘要：
Significant interest in applying Deep Neural Network (DNN) has fueled the need to support engineering of software that uses DNNs. Repairing software that uses DNNs is one such unmistakable SE need where automated tools could be beneficial; however, we do not fully understand challenges to repairing and patterns that are utilized when manually repairing DNNs. What challenges should automated repair tools address? What are the repair patterns whose automation could help developers? Which repair patterns should be assigned a higher priority for building automated bug repair tools? This work presents a comprehensive study of bug fix patterns to address these questions. We have studied 415 repairs from Stack Overflow and 555 repairs from GitHub for five popular deep learning libraries Caffe, Keras, Tensorflow, Theano, and Torch to understand challenges in repairs and bug repair patterns. Our key findings reveal that DNN bug fix patterns are distinctive compared to traditional bug fix patterns; the most common bug fix patterns are fixing data dimension and neural network connectivity; DNN bug fixes have the potential to introduce adversarial vulnerabilities; DNN bug fixes frequently introduce new bugs; and DNN bug localization, reuse of trained model, and coping with frequent releases are major challenges faced by developers when fixing bugs. We also contribute a benchmark of 667 DNN (bug, repair) instances.
引用
收藏
页码：1135 / 1146
页数：12
相关论文
共 50 条
  • [21] Identifying Recurring Patterns with Deep Neural Networks for Natural Image Denoising
    Xia, Zhihao
    Chakrabarti, Ayan
    2020 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV), 2020, : 2415 - 2423
  • [22] PatterNet: Explore and Exploit Filter Patterns for Efficient Deep Neural Networks
    Khaleghi, Behnam
    Mallappa, Uday
    Yaldiz, Duygu
    Yang, Haichao
    Shah, Monil
    Kang, Jaeyoung
    Rosing, Tajana
    PROCEEDINGS OF THE 59TH ACM/IEEE DESIGN AUTOMATION CONFERENCE, DAC 2022, 2022, : 223 - 228
  • [23] Temporally Adaptive Common Spatial Patterns with Deep Convolutional Neural Networks
    Mousavi, Mahta
    de Sa, Virginia R.
    2019 41ST ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY (EMBC), 2019, : 4533 - 4536
  • [24] Current and Future Patterns of Global Wildfire Based on Deep Neural Networks
    Zhang, Guoli
    Wang, Ming
    Yang, Baolin
    Liu, Kai
    EARTHS FUTURE, 2024, 12 (02)
  • [25] Repairing Misclassifications in Neural Networks Using Limited Data
    Henriksen, Patrick
    Leofante, Francesco
    Lomuscio, Alessio
    37TH ANNUAL ACM SYMPOSIUM ON APPLIED COMPUTING, 2022, : 1031 - 1038
  • [26] Repairing Neural Networks by Leaving the Right Past Behind
    Tanno, Ryutaro
    Pradier, Melanie F.
    Nori, Aditya
    Li, Yingzhen
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 35 (NEURIPS 2022), 2022,
  • [27] Algorithm of scattered data repairing based on neural networks
    Du Libin
    Sun Jichang
    Hou Guangli
    Liu Yan
    ICEMI 2007: PROCEEDINGS OF 2007 8TH INTERNATIONAL CONFERENCE ON ELECTRONIC MEASUREMENT & INSTRUMENTS, VOL III, 2007, : 495 - +
  • [28] Challenges in the Path Toward a Scalable Silicon Photonics Implementation of Deep Neural Networks
    Burgos, Carlos Mauricio Villegas
    Vamivakas, Nickolas
    IEEE JOURNAL OF QUANTUM ELECTRONICS, 2019, 55 (05)
  • [29] Deep convolutional neural networks for annotating gene expression patterns in the mouse brain
    Zeng, Tao
    Li, Rongjian
    Mukkamala, Ravi
    Ye, Jieping
    Ji, Shuiwang
    BMC BIOINFORMATICS, 2015, 16
  • [30] Study of Training Patterns for Employing Deep Neural Networks in Optical Communication Systems
    Chuang, Chun-Yen
    Liu, Li-Chun
    Wei, Chia-Chien
    Liu, Jun-Jie
    Henrickson, Lindor
    Wang, Chih-Lin
    Chen, Young-Kai
    Chen, Jyehong
    2018 EUROPEAN CONFERENCE ON OPTICAL COMMUNICATION (ECOC), 2018,
12345