Attack Detection from Internet of Things using TPE based Self-Attention based Bidirectional Long-Short Term Memory

In the realm of cybersecurity, artificial intelligence (AI) emerges as a potent tool for combating cyber threats effectively. By harnessing AI capabilities, cybersecurity teams streamline operations, accelerate threat detection and response, and enhance precision in defensive strategies. The increasing complexity of cyber threats necessitates the development of sophisticated intrusion detection systems (IDSs) This study contributes to the existing cybersecurity knowledge by introducing an advanced Intrusion Detection Scheme tailored for heterogeneous Industrial Internet of Things (IIoT) networks. Leveraging Deep Transfer Learning (DTL), our research enhances IDS capabilities specifically suited for IIoT environments. We employ Self-Attention based Bidirectional Long-Short Term Memory (SA-BiLSTM) to detect attacks, complemented by the Tree-Structured Parzen Estimator (TPE) for parameter optimization in BiLSTM. Through rigorous evaluation, our architecture consistently achieves an outstanding accuracy rate of approximately 99% in detecting a diverse range of cyberattacks. Notably, our system effectively counters 14 different types of cyber threats, affirming its efficacy in safeguarding IIoT networks. These results underscore the significance of ongoing advancements in creating robust, adaptable, and responsive intrusion detection systems tailored to the intricate landscape of IIoT networks.