SafeML: Safety Monitoring of Machine Learning Classifiers Through Statistical Difference Measures

Aslansefat, Koorosh; Sorokos, Ioannis; Whiting, Declan; Tavakoli Kolagari, Ramin; Papadopoulos, Yiannis

Abstract

Ensuring safety and explainability of machine learning (ML) is a topic of increasing relevance as data-driven applications venture into safety-critical application domains, traditionally committed to high safety standards that are not satisfied with an exclusive testing approach of otherwise inaccessible black-box systems. Especially the interaction between safety and security is a central challenge, as security violations can lead to compromised safety. The contribution of this paper to addressing both safety and security within a single concept of protection applicable during the operation of ML systems is active monitoring of the behavior and the operational context of the data-driven system based on distance measures of the Empirical Cumulative Distribution Function (ECDF). We investigate abstract datasets (XOR, Spiral, Circle) and current security-specific datasets for intrusion detection (CICIDS2017) of simulated network traffic, using distributional shift detection measures including the Kolmogorov-Smirnov, Kuiper, Anderson-Darling, Wasserstein and mixed Wasserstein-Anderson-Darling measures. Our preliminary findings indicate that there is a meaningful correlation between ML decisions and the ECDF-based distances measures of the input features. Thus, they can provide a confidence level that can be used for a) analyzing the applicability of the ML system in a given field (safety/security) and b) analyzing if the field data was maliciously manipulated. (Our preliminary code and results are available at https://github.com/ISorokos/SafeML.)

Citation

Aslansefat, K., Sorokos, I., Whiting, D., Tavakoli Kolagari, R., & Papadopoulos, Y. SafeML: Safety Monitoring of Machine Learning Classifiers Through Statistical Difference Measures. Presented at IMBSA: International Symposium on Model-Based Safety and Assessment, Lisbon

• Presentation Conference Type: Conference Paper (published)
• Conference Name: IMBSA: International Symposium on Model-Based Safety and Assessment
• Acceptance Date: Mar 1, 2020
• Online Publication Date: Sep 4, 2020
• Publication Date: 2020
• Deposit Date: Feb 17, 2021
• Publicly Available Date: Jul 1, 2021
• Journal: Lecture Notes in Computer Science
• Print ISSN: 0302-9743
• Publisher: Springer Verlag
• Peer Reviewed: Peer Reviewed
• Volume: 12297
• Pages: 197-211
• ISBN: 9783030589196
• DOI: https://doi.org/10.1007/978-3-030-58920-2_13
• Keywords: Safety; SafeML; Machine Learning; Deep Learning; Artificial Intelligence; Statistical difference; Domain adaptation
• Public URL: https://hull-repository.worktribe.com/output/3579760