Final Project of the Computer Science Engineering Degree
Done by: Esteban Alvarado.
This project consists of advanced phishing detection using the BERT masked language model.
Phishing has become one of the most used and dangerous cyber-attacks today. Every year there are million-dollar losses in companies due to the continuous difficulty of detecting this type of attacks at the time. However, the emerging development of the field of artificial intelligence (AI) has allowed the introduction of various effective tools to combat this problem, including language models. This project proposes to train the BERT masked language model (MLM) for advanced phishing detection, through a process known as Finetuning. Performance metrics show that this method provides favorable results, and, furthermore, its detection capacity is not limited only to URLs, but also HTML code and text or natural language. It is expected that this final trained model can accurately prevent future harmful phishing attacks that injure people and organizations.
For more information, visit the following links were the finetuned model and the dataset used in this project are provided: