ADVERSARIAL ATTACKS AND DEFENSES IN MACHINE LEARNING

Authors

  • Dr. Ayesha Khan Dr. Ayesha Khan Department of Computer Science,National University of Computer and Emerging Sciences (FAST-NUCES), Karachi, Pakistan.

DOI:

https://doi.org/10.71465/csb5

Keywords:

Adversarial attacks, machine learning, adversarial defense, robustness, cybersecurity, deep learning, AI security

Abstract

Adversarial attacks pose significant threats to the robustness and reliability of machine learning (ML) systems by subtly manipulating input data to produce erroneous outputs. These vulnerabilities are of paramount concern in sensitive domains such as healthcare, autonomous driving, and cybersecurity. This article provides a comprehensive analysis of adversarial attacks, categorizing them based on their strategies and objectives. Furthermore, it discusses state-of-the-art defense mechanisms and their effectiveness in mitigating these threats. The study concludes by identifying open challenges and future research directions in adversarial ML to enhance system security and robustness.

image

Downloads

Published

2018-06-30