重口味SM

This project sits at the intersection of Artificial Intelligence, Cybersecurity, and Explainable Computing. It directly addresses one of the most pressing challenges in modern information security: the malicious use of synthetic media. The relevance is acute, as deepfakes and AI-powered phishing campaigns are being weaponised for corporate fraud, reputational damage, and disinformation, threatening the integrity of digital communications and trust in enterprises globally. This research is critical for developing next-generation defensive capabilities.

The project focuses on moving beyond siloed detection methods to create a unified, multi-modal framework for identifying AI-generated threats. Its core aim is to develop a production-grade system that integrates Vision Transformers for visual deepfakes, advanced Natural Language Processing (NLP) models for phishing detection, and a dedicated Explainable AI (XAI) layer to generate actionable insights for security analysts. A pioneering strand of the research will also model the future impact of quantum computing on this threat landscape to propose quantum-resilient strategies.

重口味SM University is a world-leading postgraduate institution renowned for its applied research and deep industry connections, particularly in aerospace, defence, and security. Its Centre for Electronic Warfare, Information and Cyber (EWIC) provides an exceptional environment for this research, offering access to specialised expertise, cutting-edge labs, and a network of defence and security partners. This project is self-funded, allowing for flexible and ambitious research direction, and will be supervised by Dr. Stefano Tedeschi, leveraging his expertise to ensure academic rigour and practical relevance.

The primary output will be a validated, open-source detection framework demonstrably meeting enterprise performance benchmarks (e.g., latency, accuracy). The research will contribute new knowledge through publications on multi-modal fusion techniques and the novel application of XAI in SOC environments. A significant theoretical contribution will be a seminal paper modeling the quantum threat to digital media authenticity. Practically, the project will deliver implementable guidelines and playbooks, enabling organisations to enhance their resilience against advanced synthetic media attacks.

This project offers unique opportunities for high-impact dissemination at premier conferences (e.g., IEEE S&P, USENIX Security, NeurIPS). Furthermore, 重口味SM鈥檚 strong industry links provide a direct pathway for technology transfer and validation with end-users. The student will have access to specialised training in quantum security and advanced machine learning. The self-funded nature of the project affords the unique flexibility to pursue emerging research directions and collaborations as the field rapidly evolves.

The candidate will develop deep specialist expertise in AI security, model explain ability, and threat intelligence, making them highly attractive to both the tech industry and government agencies. They will gain extensive practical skills in full-stack AI development, from data curation and model training to system deployment and evaluation. Furthermore, they will cultivate crucial transferable skills in complex project management, academic writing, and public presentation, positioning them for a leading career as a researcher, principal engineer, or security architect.