KSKirubalakshmi Sekar

Kirubalakshmi Sekar

*This is an algorithmically generated author record based on publicly available information.

Kirubalakshmi Sekar has a total of 4 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • A Systematic Study on Adaptive and Explainable AI-Based Intrusion Detection Systems for Modern IoT Environments

    by Kirubalakshmi Sekar, Deemed To Be, Santhi Balachandran, Keerthana Sivamayilvelan, Vimala Devi Parthasarathy

    The increase of connected devices in the Internet of Things (IoT) and Industrial Internet of Things (IIoT) has led to a surge in the number of potential attacks. Traditional methods of intrusion detection can hardly be effective against zero-day attacks. AI-based solutions like Deep Learning (DL) and Machine Learning (ML) provide adaptable and intelligent detection of threats. The paper discusses the present implementation of AI-based IDS systems and the problems they still face in the real world. Deep learning methods, such as LSTMs (Long Short-Term Memory), CNNs (Convolutional Neural Networks), DNNs (Deep Neural Networks), and Deep Belief Networks, can achieve an accuracy for recognition of 99%. It discusses the various kinds of attacks and how it affects the real world applications. On the other hand, these systems have not become mainstream yet as they produce many false alarms and require heavy experimentation. According to this survey, the next step of innovation in IDS should be toward the creation of lightweight, interpretable models to ensure reliability and scalability.