With the complexity of network attacks, traditional intrusion detection technologies have a high false positive rate and false negative rate, low accuracy rate, and poor generalization ability in Physical and Wireless Networks (PWN). Thus, there is a need to introduce novel intrusion detection technologies to improve detection performance. In recent years, Deep Learning (DL) and Natural Language Processing (NLP) have led to vast progress in the field of pattern recognition, machine learning, and anomaly detection. DL technology has excellent performance in processing complex large-scale data, and NLP technology enables computers to analyze, understand and process natural language. For example, information extraction technology can extract attack knowledge units from heterogeneous intrusion detection data, knowledge fusion technology can realize semantic representation of network attacks through multi-source heterogeneous knowledge fusion, and knowledge inference technology can mine potential attack semantics from intrusion knowledge base. Thus, they can bring new ideas for processing multi-source heterogeneous intrusion data, and effectively improve the detection performance and the generalization ability.

Various challenges have emerged with the development of network techniques and the evolution of advanced attack methods. For example, few attack samples, lack of prior knowledge, semantic representation and comprehension of attack and background knowledge, and heterogeneity of multi-source network data expose the limitations of traditional intrusion detection methods. Traditional methods are insufficient to adequately address existing detection problems, and then they hinder the application of DL and NLP in intrusion detection paradigms. Thus, the effective deployment and efficient execution of intrusion detection models based on DL and NLP in PWN has become the focus of attention in academia and industry.

This Special Issue invites participation from academic and industry researchers working on network security problems in PWN, including technologies, theories, and architectures in DL and NLP. This Special Issue will provide an opportunity to discuss and express views on the current trends, challenges, and state-of-the-art solutions. We welcome original research and review articles on this topic.

Potential topics include but are not limited to the following:

• Representation of network infrastructure and services in PWN
• Representation and semantic comprehension of Cyber Threat Intelligence based on NLP in PWN
• Representation of network confrontation capability in PWN
• Generation and comprehension of attack semantics in PWN
• Generation and comprehension of communication semantics for multi-type network protocols
• Generation and comprehension of entity behavior semantics in PWN
• Construction of security knowledge base in PWN
• Few-shot learning for intrusion detection in PWN
• Intrusion detection framework based on DL in PWN
• Processing for multi-source heterogeneous data in PWN
• NLP technologies for intrusion detection in PWN
• Explainable AI for intrusion detection in PWN