AI-BASED INTRUSION DETECTION IN IOT SYSTEMS USING BERT

The rapid growth of Internet of Things (IoT) devices has significantly increased the vulnerability of modern networks to cyber threats. Conventional Intrusion Detection Systems (IDS) are often inadequate in handling dynamic and complex attack patterns. This paper presents a comprehensive review of recent advancements in applying Natural Language Processing (NLP) techniques, particularly transformer-based Large Language Models such as BERT, GPT-2, and BART, for intrusion detection. By representing network traffic as sequential data similar to natural language, these models are capable of learning deep contextual relationships and temporal patterns. Various architectures, including lightweight models for IoT environments, predictive security frameworks, and meta-learning approaches for zero- day attack detection, are analyzed. The study highlights that LLM-based approaches outperform traditional machine learning and deep learning methods in terms of accuracy, adaptability, and robustness.