Files and links (1)
SubmittedarXiv.org - Non-exclusive license to distribute, Open
Conference proceeding
Adversarial Robustness vs. Model Compression, or Both?
2019 IEEE/CVF International Conference on Computer Vision (ICCV), v 2019-, pp 111-120
Oct 2019
Featured in Collection : UN Sustainable Development Goals @ Drexel
Abstract
It is well known that deep neural networks (DNNs) are vulnerable to adversarial attacks, which are implemented by adding crafted perturbations onto benign examples. Min-max robust optimization based adversarial training can provide a notion of security against adversarial attacks. However, adversarial robustness requires a significantly larger capacity of the network than that for the natural training with only benign examples. This paper proposes a framework of concurrent adversarial training and weight pruning that enables model compression while still preserving the adversarial robustness and essentially tackles the dilemma of adversarial training. Furthermore, this work studies two hypotheses about weight pruning in the conventional setting and finds that weight pruning is essential for reducing the network model size in the adversarial setting; training a small model from scratch even with inherited initialization from the large model cannot achieve neither adversarial robustness nor high standard accuracy. Code is available at https://github.com/yeshaokai/Robustness-Aware-Pruning-ADMM.
Metrics
Details
- Title
- Adversarial Robustness vs. Model Compression, or Both?
- Creators
- Shaokai Ye - Tsinghua UniversityKaidi Xu - Northeastern UniversitySijia Liu - Xi'an Jiaotong UniversityHao Cheng - Tsinghua UniversityJan-Henrik Lambrechts - University of California, Los AngelesHuan Zhang - Group SenseAojun Zhou - Tsinghua UniversityKaisheng Ma - Northeastern UniversityYanzhi Wang - Northeastern UniversityXue Lin - Northeastern University
- Publication Details
- 2019 IEEE/CVF International Conference on Computer Vision (ICCV), v 2019-, pp 111-120
- Publisher
- IEEE
- Resource Type
- Conference proceeding
- Language
- English
- Academic Unit
- Computer Science (Computing)
- Web of Science ID
- WOS:000531438100012
- Scopus ID
- 2-s2.0-85081899910
- Other Identifier
- 991021871362004721
UN Sustainable Development Goals (SDGs)
This publication has contributed to the advancement of the following goals:
InCites Highlights
Data related to this publication, from InCites Benchmarking & Analytics tool:
- Collaboration types
- Industry collaboration
- Domestic collaboration
- International collaboration
- Web of Science research areas
- Computer Science, Artificial Intelligence
- Engineering, Electrical & Electronic