Abstract
Deep neural networks (DNNs) have become the technology of choice for realizing a variety of complex tasks. However, as highlighted by many recent studies, even an imperceptible perturbation to a correctly classified input can lead to misclassification by a DNN. This renders DNNs vulnerable to strategic input manipulations by attackers, and also over-sensitive to environmental noise. To mitigate this phenomenon, practitioners apply joint classification by an ensemble of DNNs. By aggregating the classification outputs of different individual DNNs for the same input, ensemble-based classification reduces the risk of misclassifications due to the specific realization of the stochastic training process of any single DNN. However, the effectiveness of a DNN ensemble is highly dependent on its members not simultaneously erring on many different inputs. In this case study, we harness recent advances in DNN verification to devise a methodology for identifying ensemble compositions that are less prone to simultaneous errors, even when the input is adversarially perturbed - resulting in more robustly-accurate ensemble-based classification. Our proposed framework uses a DNN verifier as a backend, and includes heuristics that help reduce the high complexity of directly verifying ensembles. More broadly, our work puts forth a novel universal objective for formal verification that can potentially improve the robustness of real-world, deep-learning-based systems across a variety of application domains.
Original language | English |
---|---|
Title of host publication | Proceedings of the 22nd Conference on Formal Methods in Computer-Aided Design, FMCAD 2022 |
Editors | Alberto Griggio, Neha Rungta |
Publisher | Institute of Electrical and Electronics Engineers Inc. |
Pages | 27-37 |
Number of pages | 11 |
ISBN (Electronic) | 9783854480532 |
DOIs | |
State | Published - 2022 |
Event | 22nd International Conference on Formal Methods in Computer-Aided Design, FMCAD 2022 - Trento, Italy Duration: 17 Oct 2022 → 21 Oct 2022 |
Publication series
Name | Proceedings of the 22nd Conference on Formal Methods in Computer-Aided Design, FMCAD 2022 |
---|
Conference
Conference | 22nd International Conference on Formal Methods in Computer-Aided Design, FMCAD 2022 |
---|---|
Country/Territory | Italy |
City | Trento |
Period | 17/10/22 → 21/10/22 |
Bibliographical note
Publisher Copyright:© 2022 FMCAD Association and authors.