Natural-Language Processing for Intrusion Detection

Intrusion-detection systems seek to electronically identify malicious traffic as it enters a defended network. Social engineering, a unique type of attack traffic, attempts to compromise a network or system´s security metrics by exploiting the human end user through natural language, based on common psychological flaws and deception. These attacks have been difficult to defend against in the past with IDSs because natural language is highly variable. Natural-language processing teaches computers the semantic meaning of natural-language text. Thus, an NLP system reads plain English (among other languages) and categorizes what it´s seen in terms of conceptual themes and ontological concept.