CNN-Cert: An Efficient Framework for Certifying Robustness of Convolutional Neural Networks
Verifying robustness of neural network classifiers has attracted great interests and attention due to the success of deep neural networks and their unexpected vulnerability to adversarial perturbations. Although finding minimum adversarial distortion of neural networks (with ReLU activations) has be...
| Main Authors: | , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Association for the Advancement of Artificial Intelligence (AAAI)
2021
|
| Online Access: | https://hdl.handle.net/1721.1/129951 |
| _version_ | 1826207886205779968 |
|---|---|
| author | Boopathy, Akhilan Weng, Tsui-Wei Chen, Pin-Yu Liu, Sijia Daniel, Luca |
| author2 | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science |
| author_facet | Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Boopathy, Akhilan Weng, Tsui-Wei Chen, Pin-Yu Liu, Sijia Daniel, Luca |
| author_sort | Boopathy, Akhilan |
| collection | MIT |
| description | Verifying robustness of neural network classifiers has attracted great interests and attention due to the success of deep neural networks and their unexpected vulnerability to adversarial perturbations. Although finding minimum adversarial distortion of neural networks (with ReLU activations) has been shown to be an NP-complete problem, obtaining a non-trivial lower bound of minimum distortion as a provable robustness guarantee is possible. However, most previous works only focused on simple fully-connected layers (multilayer perceptrons) and were limited to ReLU activations. This motivates us to propose a general and efficient framework, CNN-Cert, that is capable of certifying robustness on general convolutional neural networks. Our framework is general - we can handle various architectures including convolutional layers, max-pooling layers, batch normalization layer, residual blocks, as well as general activation functions; our approach is efficient - by exploiting the special structure of convolutional layers, we achieve up to 17 and 11 times of speed-up compared to the state-of-the-art certification algorithms (e.g. Fast-Lin, CROWN) and 366 times of speed-up compared to the dual-LP approach while our algorithm obtains similar or even better verification bounds. In addition, CNN-Cert generalizes state-of-the-art algorithms e.g. Fast-Lin and CROWN. We demonstrate by extensive experiments that our method outperforms state-of-the-art lower-bound-based certification algorithms in terms of both bound quality and speed. |
| first_indexed | 2024-09-23T13:56:30Z |
| format | Article |
| id | mit-1721.1/129951 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T13:56:30Z |
| publishDate | 2021 |
| publisher | Association for the Advancement of Artificial Intelligence (AAAI) |
| record_format | dspace |
| spelling | mit-1721.1/1299512022-09-28T17:16:50Z CNN-Cert: An Efficient Framework for Certifying Robustness of Convolutional Neural Networks Boopathy, Akhilan Weng, Tsui-Wei Chen, Pin-Yu Liu, Sijia Daniel, Luca Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science MIT-IBM Watson AI Lab Verifying robustness of neural network classifiers has attracted great interests and attention due to the success of deep neural networks and their unexpected vulnerability to adversarial perturbations. Although finding minimum adversarial distortion of neural networks (with ReLU activations) has been shown to be an NP-complete problem, obtaining a non-trivial lower bound of minimum distortion as a provable robustness guarantee is possible. However, most previous works only focused on simple fully-connected layers (multilayer perceptrons) and were limited to ReLU activations. This motivates us to propose a general and efficient framework, CNN-Cert, that is capable of certifying robustness on general convolutional neural networks. Our framework is general - we can handle various architectures including convolutional layers, max-pooling layers, batch normalization layer, residual blocks, as well as general activation functions; our approach is efficient - by exploiting the special structure of convolutional layers, we achieve up to 17 and 11 times of speed-up compared to the state-of-the-art certification algorithms (e.g. Fast-Lin, CROWN) and 366 times of speed-up compared to the dual-LP approach while our algorithm obtains similar or even better verification bounds. In addition, CNN-Cert generalizes state-of-the-art algorithms e.g. Fast-Lin and CROWN. We demonstrate by extensive experiments that our method outperforms state-of-the-art lower-bound-based certification algorithms in terms of both bound quality and speed. 2021-02-22T17:54:22Z 2021-02-22T17:54:22Z 2019-01 2018-11 2020-12-07T16:09:19Z Article http://purl.org/eprint/type/ConferencePaper 2159-5399 2374-3468 https://hdl.handle.net/1721.1/129951 Boopathy, Akhilan et al. “CNN-Cert: An Efficient Framework for Certifying Robustness of Convolutional Neural Networks.” Paper in the Proceedings of the AAAI Conference on Artificial Intelligence, 33, 1 Thirty-Third AAAI Conference on Artificial Intelligence (AAAI-19), Honolulu, Hawaii, January 27–February 1, 2019, AAAI: 33013240 © 2019 The Author(s) en 10.1609/AAAI.V33I01.33013240 Proceedings of the AAAI Conference on Artificial Intelligence Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf Association for the Advancement of Artificial Intelligence (AAAI) arXiv |
| spellingShingle | Boopathy, Akhilan Weng, Tsui-Wei Chen, Pin-Yu Liu, Sijia Daniel, Luca CNN-Cert: An Efficient Framework for Certifying Robustness of Convolutional Neural Networks |
| title | CNN-Cert: An Efficient Framework for Certifying Robustness of Convolutional Neural Networks |
| title_full | CNN-Cert: An Efficient Framework for Certifying Robustness of Convolutional Neural Networks |
| title_fullStr | CNN-Cert: An Efficient Framework for Certifying Robustness of Convolutional Neural Networks |
| title_full_unstemmed | CNN-Cert: An Efficient Framework for Certifying Robustness of Convolutional Neural Networks |
| title_short | CNN-Cert: An Efficient Framework for Certifying Robustness of Convolutional Neural Networks |
| title_sort | cnn cert an efficient framework for certifying robustness of convolutional neural networks |
| url | https://hdl.handle.net/1721.1/129951 |
| work_keys_str_mv | AT boopathyakhilan cnncertanefficientframeworkforcertifyingrobustnessofconvolutionalneuralnetworks AT wengtsuiwei cnncertanefficientframeworkforcertifyingrobustnessofconvolutionalneuralnetworks AT chenpinyu cnncertanefficientframeworkforcertifyingrobustnessofconvolutionalneuralnetworks AT liusijia cnncertanefficientframeworkforcertifyingrobustnessofconvolutionalneuralnetworks AT danielluca cnncertanefficientframeworkforcertifyingrobustnessofconvolutionalneuralnetworks |