The DOI matches its Crossref record; 40 of the 40 references it cites were found in a registry or an index.
AI2: Safety and Robustness Certification of Neural Networks with Abstract Interpretation
Open accessAbstract
We present AI2, the first sound and scalable analyzer for deep neural networks. Based on overapproximation, AI2can automatically prove safety properties (e.g., robustness) of realistic neural networks (e.g., convolutional neural networks). The key insight behind AI2is to phrase reasoning about safety and robustness of neural networks in terms of classic abstract interpretation, enabling us to leverage decades of advances in that area. Concretely, we introduce abstract transformers that capture the behavior of fully connected and convolutional neural network layers with rectified linear unit activations (ReLU), as well as max pooling layers. This allows us to handle real-world neural networks, which are often built out of those types of layers. We present a complete implementation of AI2together with an extensive evaluation on 20 neural networks. Our results demonstrate that: (i) AI2is precise enough to prove useful specifications (e.g., robustness), (ii) AI2can be used to certify the effectiveness of state-of-the-art defenses for neural networks, (iii) AI2is significantly faster than existing analyzers based on symbolic analysis, which often take hours to verify simple fully connected networks, and (iv) AI2can handle deep convolutional networks, which are beyond the reach of existing methods.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155
Gradient-based learning applied to document recognition
Lecun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998) · Proceedings of the IEEE · doi:10.1109/5.726791 · Cited by 58,805
Learning Multiple Layers of Features from Tiny Images
Krizhevsky, A. (2024) · doi:10.57702/zp44cu3g · Cited by 23,395
Receptive fields, binocular interaction and functional architecture in the cat’s visual cortex
Hubel, D. H., & Wiesel, T. N. (1962) · The Journal of Physiology · doi:10.1113/jphysiol.1962.sp006837 · Cited by 13,680
Towards Evaluating the Robustness of Neural Networks
Carlini, N., & Wagner, D. (2017) · 2017 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp.2017.49 · Cited by 8,859
Intriguing properties of neural networks
Szegedy, C., Zaremba, W., Sutskever, I., Bruna, J., Erhan, D., Goodfellow, I., & Fergus, R. (2013) · arXiv · doi:10.48550/arxiv.1312.6199 · Cited by 8,143
Towards Deep Learning Models Resistant to Adversarial Attacks
Madry, A., Makelov, A., Schmidt, L., Tsipras, D., & Vladu, A. (2017) · arXiv · doi:10.48550/arxiv.1706.06083 · Cited by 6,602
Cousot, P., & Cousot, R. (1977) · Proceedings of the 4th ACM SIGACT-SIGPLAN symposium on Principles of programming languages - POPL '77 · doi:10.1145/512950.512973 · Cited by 6,279
Adversarial examples in the physical world
Kurakin, A., Goodfellow, I. J., & Bengio, S. (2018) · Artificial Intelligence Safety and Security · doi:10.1201/9781351251389-8 · Cited by 5,375
DeepFool: A Simple and Accurate Method to Fool Deep Neural Networks
Moosavi-Dezfooli, S.-M., Fawzi, A., & Frossard, P. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.282 · Cited by 5,317
On Assessing ML Model Robustness: A Methodological Framework (Academic Track)
Awadid, A., & Robert, B. (2025) · OASIcs : OpenAccess Series in Informatics · doi:10.4230/oasics.saia.2024.1 · Cited by 4,702
Deep neural networks are easily fooled: High confidence predictions for unrecognizable images
Nguyen, A., Yosinski, J., & Clune, J. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298640 · Cited by 3,258
End to End Learning for Self-Driving Cars
Bojarski, M., Del Testa, D., Dworakowski, D., Firner, B., Flepp, B., Goyal, P., Jackel, L. D., Monfort, M., Muller, U., Zhang, J., et al. (2016) · arXiv · doi:10.48550/arxiv.1604.07316 · Cited by 3,198
Distillation as a Defense to Adversarial Perturbations Against Deep Neural Networks
Papernot, N., McDaniel, P., Wu, X., Jha, S., & Swami, A. (2016) · 2016 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp.2016.41 · Cited by 2,982
Ensemble Adversarial Training: Attacks and Defenses
Tramèr, F., Kurakin, A., Papernot, N., Goodfellow, I., Boneh, D., & McDaniel, P. (2017) · arXiv · doi:10.48550/arxiv.1705.07204 · Cited by 2,176
Reluplex: An Efficient SMT Solver for Verifying Deep Neural Networks
Katz, G., Barrett, C., Dill, D. L., Julian, K., & Kochenderfer, M. J. (2017) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-63387-9_5 · Cited by 1,706
Automatic discovery of linear restraints among variables of a program
Cousot, P., & Halbwachs, N. (1978) · Proceedings of the 5th ACM SIGACT-SIGPLAN symposium on Principles of programming languages - POPL '78 · doi:10.1145/512760.512770 · Cited by 1,614
DeepXplore: Automated Whitebox Testing of Deep Learning Systems
Pei, K., Cao, Y., Yang, J., & Jana, S. (2017) · Proceedings of the 26th Symposium on Operating Systems Principles · doi:10.1145/3132747.3132785 · Cited by 1,217
Safety Verification of Deep Neural Networks
Huang, X., Kwiatkowska, M., Wang, S., & Wu, M. (2017) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-63387-9_1 · Cited by 818
Abstract Interpretation Frameworks
COUSOT, P., & COUSOT, R. (1992) · Journal of Logic and Computation · doi:10.1093/logcom/2.4.511 · Cited by 710
Towards Deep Neural Network Architectures Robust to Adversarial Examples
Gu, S., & Rigazio, L. (2014) · arXiv (Cornell University) · doi:10.48550/arxiv.1412.5068 · Cited by 622
Handwritten digit recognition: applications of neural network chips and automatic learning
Le Cun, Y., Jackel, L. D., Boser, B., Denker, J. S., Graf, H. P., Guyon, I., Henderson, D., Howard, R. E., & Hubbard, W. (1989) · IEEE Communications Magazine · doi:10.1109/35.41400 · Cited by 480
Apron: A Library of Numerical Abstract Domains for Static Analysis
Jeannet, B., & Miné, A. (2009) · Lecture Notes in Computer Science · doi:10.1007/978-3-642-02658-4_52 · Cited by 464
The Space of Transferable Adversarial Examples
Tramèr, F., Papernot, N., Goodfellow, I., Boneh, D., & McDaniel, P. D. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1704.03453 · Cited by 395
Droid-Sec: deep learning in android malware detection
Yuan, Z., Lu, Y., Wang, Z., & Xue, Y. (2014) · Proceedings of the 2014 ACM conference on SIGCOMM · doi:10.1145/2619239.2631434 · Cited by 371
Shaham, U., Yamada, Y., & Negahban, S. (2018) · Neurocomputing · doi:10.1016/j.neucom.2018.04.027 · Cited by 355
Adversarial Perturbations Against Deep Neural Networks for Malware Classification
Grosse, K., Papernot, N., Manoharan, P., Michael Backes, A., & McDaniel, P. D. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1606.04435 · Cited by 299
An Abstraction-Refinement Approach to Verification of Artificial Neural Networks
Pulina, L., & Tacchella, A. (2010) · Lecture Notes in Computer Science · doi:10.1007/978-3-642-14295-6_24 · Cited by 261
Learning with a Strong Adversary
Huang, R., Xu, B. H., Schuurmans, D., & Szepesvári, C. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1511.03034 · Cited by 240
Adversarial Example Defense: Ensembles of Weak Defenses are not Strong
He, W., Wei, J., Chen, X., Carlini, N., & Song, D. X. (2017) · Cited by 221
Robust Physical-World Attacks on Machine Learning Models.Preprint
Evtimov, I., Eykholt, K., Fernandes, E., Kohno, T., Li, B., Prakash, A., Rahmati, A., & Song, D. X. (2017) · arXiv (Cornell University) · Cited by 195
Exploring the space of adversarial images
Tabacof, P., & Valle, E. (2016) · 2016 International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2016.7727230 · Cited by 168
Chaudhuri, S., Gulwani, S., Lublinerman, R., & Navidpour, S. M. (2011) · ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering (ESEC/FSE) · doi:10.1145/2025113.2025131 · Cited by 136
Self-citationThe Zonotope Abstract Domain Taylor1+
Ghorbal, K., Goubault, É., & Putot, S. (2009) · Lecture notes in computer science · doi:10.1007/978-3-642-02658-4_47 · Cited by 116
Continuity analysis of programs
Chaudhuri, S., Gulwani, S., & Lublinerman, R. (2010) · ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages (POPL) · doi:10.1145/1706299.1706308 · Cited by 110
Self-citationHandwritten Digit Recognition: Applications of Neural Net Chips and Automatic Learning
LeCun, Y., Jackel, L. D., Boser, B. E., Denker, J. S., Graf, H. P., Guyon, I., Henderson, D., Howard, R. E., & Hubbard, W. (1990) · Springer eBooks · doi:10.1007/978-3-642-76153-9_35 · Cited by 105
Fast polyhedra abstract domain
Singh, G., Püschel, M., & Vechev, M. (2016) · ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages (POPL) · doi:10.1145/3009837.3009885 · Cited by 87
Self-citationStatic Analysis in Disjunctive Numerical Domains
Sankaranarayanan, S., Ivančić, F., Shlyakhter, I. A., & Gupta, A. (2006) · Lecture notes in computer science · doi:10.1007/11823230_2 · Cited by 86
Inferring Disjunctive Postconditions
Popeea, C., & Chin, W.-N. (2007) · Lecture notes in computer science · doi:10.1007/978-3-540-77505-8_26 · Cited by 63
Towards Verification of Artificial Neural Networks.
Scheibler, K., Winterer, L., Wimmer, R., & Becker, B. (2015) · MBMV · Cited by 49
Cited by
Certified Robustness to Adversarial Examples with Differential Privacy
Lecuyer, M., Atlidakis, V., Geambasu, R., Hsu, D., & Jana, S. (2019) · 2019 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp.2019.00044 · Cited by 750
An abstract domain for certifying neural networks
Singh, G., Gehr, T., Püschel, M., & Vechev, M. (2019) · Proceedings of the ACM on Programming Languages · doi:10.1145/3290354 · Cited by 609
Self-citationHuang, X., Kroening, D., Ruan, W., Sharp, J., Sun, Y., Thamo, E., Wu, M., & Yi, X. (2020) · Computer Science Review · doi:10.1016/j.cosrev.2020.100270 · Cited by 532
The Marabou Framework for Verification and Analysis of Deep Neural Networks
Katz, G., Huang, D. A., Ibeling, D., Julian, K., Lazarus, C., Lim, R., Shah, P., Thakoor, S., Wu, H., Zeljić, A., et al. (2019) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-25540-4_26 · Cited by 517
“Several recent papers have proposed verification schemes that rely on DNNlevel reasoning [5,23].”
BackgroundMethodThinking responsibly about responsible AI and ‘the dark side’ of AI
Mikalef, P., Conboy, K., Lundström, J. E., & Popovič, A. (2022) · European Journal of Information Systems · doi:10.1080/0960085x.2022.2026621 · Cited by 435
Metadata
- Authors
- Gehr, T., Mirman, M., Drachsler-Cohen, D., Tsankov, P., Chaudhuri, S., & Vechev, M.
- Venue
- 2018 IEEE Symposium on Security and Privacy (SP)
- Publisher
- IEEE
- Published
- 2018-05
- Pages
- 3-18
- ISSN
- 1081-6011 (ISSN-L)
- OpenAlex
- W2794609696
- Cited by
- 799
- References
- 40
- FWCI
- 108.77
- Type
- conference paper
- Language
- en
- Open access
- bronze
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.