Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{200428,
        author = {Abhishek Sharma and Neha and Urvashee and Monika and Deepak Sharma},
        title = {Adversarial Attacks on Neural Networks: A Comprehensive Survey, Analysis, and Experimental Test Cases},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {12},
        pages = {1016-1025},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=200428},
        abstract = {Neural networks have achieved remarkable performance across diverse machine learning tasks, yet they remain vulnerable to adversarial attacks—carefully crafted perturbations that cause misclassification or erroneous outputs. This paper presents a comprehensive survey of adversarial attacks on deep neural networks, covering theoretical foundations, attack taxonomies, benchmark datasets, experimental test cases, and state-of-the-art defenses. We analyze major attack paradigms including the Fast Gradient Sign Method (FGSM), Projected Gradient Descent (PGD), Carlini-Wagner (C&W) attacks, DeepFool, and physical-world attacks. We further present structured test case scenarios across image classification, natural language processing, and audio domains. Our analysis provides insights for researchers aiming to develop robust, trustworthy AI systems.},
        keywords = {Adversarial Examples, Neural Networks, FGSM, PGD, Robustness, Deep Learning, Machine Learning Security, Perturbation, Defense Mechanisms},
        month = {May},
        }