Convolutional neural networks (CNNs) have achieved immense popularity in areas like computer vision, image processing, speech proccessing, and many others. Unfortunately, despite their excellent performance, they are prone to producing erroneous results — for example, minor perturbations to their inputs can result in severe classification errors. In this paper, we present the Cnn-Abs framework, which implements an abstraction-refinement based scheme for CNN verification. Specifically, Cnn-Abs simplifies the verification problem through the removal of convolutional connections in a way that soundly creates an over-approximation of the original problem; it then iteratively restores these connections if the resulting problem becomes too abstract. Cnn-Abs is designed to use existing verification engines as a backend, and our evaluation demonstrates that it can significantly boost the performance of a state-of-the-art DNN verification engine, reducing runtime by 15.7% on average.
|Original language||American English|
|Title of host publication||Automated Technology for Verification and Analysis - 20th International Symposium, ATVA 2022, Proceedings|
|Editors||Ahmed Bouajjani, Lukáš Holík, Zhilin Wu|
|Publisher||Springer Science and Business Media Deutschland GmbH|
|Number of pages||6|
|State||Published - 2022|
|Event||20th International Symposium on Automated Technology for Verification and Analysis, ATVA 2022 - Virtual, Online|
Duration: 25 Oct 2022 → 28 Oct 2022
|Name||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
|Conference||20th International Symposium on Automated Technology for Verification and Analysis, ATVA 2022|
|Period||25/10/22 → 28/10/22|
Bibliographical noteFunding Information:
Acknowledgements. This work was partially supported by the Semiconductor
This work was partially supported by the Semiconductor Research Corporation, the Binational Science Foundation (grant numbers 2017662 and 2020250), the Israel Science Foundation (683/18), and the National Science Foundation (1814369).
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.