Cryptography Skill Trees: Key Topic Identification and Knowledge Graph Modelling in the Field of Cryptography Within a Cyber Security Context

Cybersecurity is a rapidly growing field and is now an inseparable part of modern organizations. To fulfill the rise in demand for skilled cybersecurity practitioners, cybersecurity education and training have become essential. This paper proposes the creation of a cybersecurity-related knowledge graph for the purpose of cybersecurity education. Due to limitations in time and resources, the knowledge graph created will cover the field of cryptography, a subfield of cybersecurity. Sources pertaining to cryptography are scraped from the web, pre-processed, and separated into documents based on topic. A label is decided for each document, using a combination of keyword extraction via KeyBERT and Gemini Pro. The cosine similarity between the documents are calculated and normalized to create a similarity matrix. Each keyword in the similarity matrix is grouped into clusters based on a certain threshold. The keyword clusters are then represented in the form of a knowledge graph. The resulting graph is well within our expectation, with keywords within the same cluster being close together and with general topics populating a central position. Ultimately, with enough time and resources, this research aims to cover the entirety of the cybersecurity field and contribute to the creation of an adaptive learning system for cyber security.