global-journal-of-computer-science-and-technology-d-neural-ai Global Journal of Computer Science and Technology - D: Neural & AI 0975-4350 0975-4172 Global Journals Publishing Group Incorporated 83126 Optimizing the Running Time of a Trigger Search Algorithm Based on the Principles of Formal Verification of Artificial Neural Networks Optimizing Formal Verification for Neural Triggers TonkikhAleksey StroevaEkaterina RUSSIA, Moscow State University 28 08 2024 24 D1 11 19

The article examines the problem of scalability of the algorithm for searching for a trigger in images, which is based on the operating principle of the Deep Poly formal verification algorithm. The existing implementation had a number of shortcomings. According to them, the requirements for the optimized version of the algorithm were formulated, which were brought to practical implementation. Achieved 4 times acceleration compared to the original implementation.

 formal verification machine learning trigger injection attacks backdoor attacks your Full Text

The article examines the problem of scalability of the algorithm for searching for a trigger in images, which is based on the operating principle of the Deep Poly formal verification algorithm. The existing implementation had a number of shortcomings. According to them, the requirements for the optimized version of the algorithm were formulated, which were brought to practical implementation. Achieved 4 times acceleration compared to the original implementation