Year: 2026
Type: book-chapter
Abstract: The rising complexity and the magnitude of cyber threats have now played their part in shifting the paradigm from traditional rule-based cybersecurity to intelligent and adaptive artificial intelligence (AI)–based security solutions. This chapter explores emerging AI trends that are disrupting the domain of cybersecurity: supervised and unsupervised machine learning, various deep learning architectures, such as convolutional neural networks, recurrent neural networks, and transformers, and forward-looking approaches, such as reinforcement learning, transfer learning, and zero-shot learning. These technologies provide advanced threat detection, real-time adaptive defense, and better anomaly detection—even for attacks that no one has previously seen. Supervised models provide great accuracy under known threat scenarios, but unsupervised methods and ensemble techniques perform very well for generic anomaly detection that does not need massive amounts of labeled data. Deep learning methods allow further automation of feature extraction and pattern recognition in high-dimensional data; reinforcement learning fosters self-evolving cybersecurity systems, whereas transfer learning is about fast adaptation across domains with limited retraining. Still largely theoretical, zero-shot learning holds a lot of promise in preemptively scouting new avenues of cyber threats. This chapter discusses these methods comparatively, emphasizing their unique advantages and applied uses, but also some of their challenges concerning model interpretability, lack of data, and adversarial robustness. The entire conversation will be made comprehensive with visuals, case applications, and performance insights, which will prepare the reader to fully grasp how AI is indeed changing the face of cybersecurity into a proactive and intelligent area.
Source: Wiley eBooks
Authors: Utpal Ghosh, Shrabanti Kundu
Institution: Sarojini Naidu College for Women
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W7220719503