Attackability Characterization of Adversarial Evasion Attack on Discrete Data
KAUST DepartmentComputer Science
Computer Science Program
Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
Machine Intelligence & kNowledge Engineering Lab
KAUST Grant NumberFCC/1/1976-19-01
Online Publication Date2020-08-20
Print Publication Date2020-08-23
Permanent link to this recordhttp://hdl.handle.net/10754/664815
MetadataShow full item record
AbstractEvasion attack on discrete data is a challenging, while practically interesting research topic. It is intrinsically an NP-hard combinatorial optimization problem. Characterizing the conditions guaranteeing the solvability of an evasion attack task thus becomes the key to understand the adversarial threat. Our study is inspired by the weak submodularity theory. We characterize the attackability of a targeted classifier on discrete data in evasion attack by bridging the attackability measurement and the regularity of the targeted classifier. Based on our attackability analysis, we propose a computationally efficient orthogonal matching pursuit-guided attack method for evasion attack on discrete data. It provides provably attack efficiency and performances. Substantial experimental results on real-world datasets validate the proposed attackability conditions and the effectiveness of the proposed attack method.
CitationWang, Y., Han, Y., Bao, H., Shen, Y., Ma, F., Li, J., & Zhang, X. (2020). Attackability Characterization of Adversarial Evasion Attack on Discrete Data. Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. doi:10.1145/3394486.3403194
SponsorsOur research in this publication was supported by funding from King Abdullah University of Science and Technology (KAUST), under award number FCC/1/1976-19-01 and KAUST AI Initiative, and NSFC No. 61828302.