Enhancing machine learning robustness against adversarial attacks through cryptographic techniques
Dankan Gowda V, Pullela SVVSR Kumar, Praveen Damacharla, Manoj R. Tarambale, Prasanna Kumar Lakineni, Kalavakolanu Sripathi · Journal of Information and Optimization Sciences · 2025
This paper focuses with the issue of adversarial attacks on machine learning models, and also provides a possible solution to improve the model’s robustness through the use of cryptographic solutions. Thus, extending the virtues of cryptography to DL models, the proposed method seeks to shield them from adversarial alterations that might result in wrong classifications. The paper also describes the algorithmic parts of the proposed methodology as well as the analysis of the complexity of the present work. The efficiency of the proposed approach has been proven during the experiments; thus, model security is enhanced without compromising the performance significantly. Based on the findings of the work, the use of cryptographic approaches can be recommended as a promising avenue toward enhancing the security of machine learning.