![Sanyam Vyas](https://profiles-cardiff.imgix.net/attachments/3465?w=200&h=200&auto=format&crop=faces&fit=crop&q=90")
Sanyam Vyas
Research student
School of Computer Science and Informatics
- VyasS3@cardiff.ac.uk
- Abacws, Floor 4, Room 4.34, Senghennydd Road, Cathays, Cardiff, CF24 4AG
Overview
Occupation:
PhD researcher in the School of Computer Science and Informatics, specifically focusing on Autonomous Cyber Network Defence (see Research for more detail).
Funded by: Airbus Defence and EPSRC.
Background:
MEng with Honours Computer Science and Electronics, University of Bristol (2017-2021)
- Individual Research Project (4th year): Privacy-preserving Fall Detection System for Digital Health.
- Industrial Research Placement (3rd year): AI-based Network Intrusion Detection System using Transfer Learning
Work Experience:
- Machine Learning Research Intern at The Alan Turing Institute's AICD Research Centre (09/2023-04/2024)
- Teaching Assistant at Cardiff School of Computer Science and Informatics (10/2022-06/2023)
- Cyber Security Outreach Officer at Cardiff School of Computer Science and Informatics (11/2021-07/2022)
- Industrial Research Placement (3rd year) and Summer Intern at Toshiba Europe (01/2020-09/2020)
Publications:
- [1] Vyas S., Hannay J., Bolton A., and Burnap P. 2023. Automated Cyber Defence: A Review, 2023 (Accepted at ACM Computing Surveys)
- [2] Vyas S., Hicks C., Mavroudis V.. Mitigating Deep Reinforcement Learning Backdoors in the Neural Activation Space. Deep Learning Security and Privacy Workshop (DLSP), 2024
Research
Research Elaborated:
My area of research lies in Autonomous Cyber Network Defence, which revolves around a game-playing environment featuring fully autonomous AI-based red and blue (attacker and defender) agents. The agents are to combat one another through the involvement of various adversarial AI and cyber attacks for a holistic purpose of strengthening the defence system of an enterprise network.
Research areas of interest within Autonomous Cyber Network Defence:
- Detection and sanitisation of poisoning within autonomous AI agents
- Scalability of autonomous AI agents
- Explainabliity and Interpretability of the autonomous AI agents
Publications:
- [1] Vyas S., Hannay J., Bolton A., and Burnap P. 2023. Automated Cyber Defence: A Review, 2023 (Accepted at ACM Computing Surveys)
- [2] Vyas S., Hicks C., Mavroudis V.. Mitigating Deep Reinforcement Learning Backdoors in the Neural Activation Space.Deep Learning Security and Privacy Workshop (DLSP), 2024
Teaching
- CMT121 Penetration Testing and Malware Analysis (10/2022-02/2023)
- CMT316 Applications of Machine Learning (10/2022-06/2023)
- CMT307 Applied Machine Learning (02/2023-06/2023)
