Mechanisms and Machine Science Print ISSN: 2211-0984 | Electronic ISSN: 2211-0992 Volume 15 : 1, 2023 291 Defending Against Adversarial Attacks in Text-based Fake News Detection: A Machine Learning Survey 1 Pushpendra Kumar The NorthCap University, Gurugram, Haryana, India Shaveta Arora The NorthCap University, Gurugram, Haryana, India shavetaarora@ncuindia.edu Shraddha Arora The NorthCap University, Gurugram, Haryana, India shraddhaarora@ncuindia.edu Abstract Textual data-based fake news detection has a significant challenge in the form of Adversarial attacks. There can be issues like fact distortion, subject-object exchange, and fact confounding due to adversarial attacks in the area of text-based fake news detection. Therefore, it is crucial to outline the weaknesses of different computational models used in text-based fake news detection and highlight areas where the models may be vulnerable to manipulation. More effective defenses can be developed by understanding the nature of adversarial attacks. This study reviews and summarizes various possible adversarial attacks and strategies. Further, this study outlines various strategies to safeguard against these attacks in text-based fake news detection. Adversarial attacks can erode trust in machine learning models, particularly in applications such as fake news detection where the stakes are high. Studying adversarial attacks in text-based false news detection is extremely useful for enhancing the robustness and efficacy of the computational models used in machine learning, developing more effective defenses against attacks, enhancing model interpretability, and enhancing trust in machine learning more broadly. Keywords Fake News Detection, Adversarial Attacks, Machine Learning, NLP. 1. Introduction Fake news pertains to false or misleading information presented as if it were true and disseminated through various media channels, like social media, news websites, and traditional media outlets. Fake news can take many forms, including hoaxes, rumors, conspiracy theories, propaganda, and satire. 1 Address Author Correspondence to Pushpendra Kumar at pushpendra22csd010@ncuindia.edu Accepted: 27 April 2023 / Published online: 15 May 2023 © The Author(s), 2023. Paper ID, 100155.