Detection and Analysis of Advanced Phishing Attacks (DAAPA) MSc by Research

In this project our aim is to develop a novel, context-aware, domain-specific NLP framework that protects systems against real-world, low-volume Business Email Compromise and spear-phishing attacks. The objectives are: To identify and characterise the distinct linguistic and pragmatic and AI-generated stylistic features utilised in contemporary advanced phishing attacks across diverse corpora. To develop and curate a large-scale, high-fidelity dataset of malicious and benign communications. To design and implement the Domain-Adapted Transformer Model incorporating Hybrid Linguistic-Contextual and Anomaly-based features. To evaluate the model against traditional and general deep learning models using comprehensive de facto metrics. To develop an interpretability mechanism to attribute classification decisions to specific linguistic features, and so enhance model trustworthiness and operational utility.

IELTS 6.5 overall with 6.0 in reading and writing, 5.5 in listening and speaking

September 2026

6 July 2026

• Headington, United Kingdom