Marc Juarez
School of Informatics,
University of Edinburgh
Office: IF-4.05A
10 Crichton St
Edinburgh, UK
EH8 9AB
Email: <first>.<last>@ed.ac.uk
I am a Lecturer in Cyber Security and Privacy in the University of Edinburgh’s School of Informatics. Within the school, I am a member of the Security, Privacy, and Trust group (SecPrivTru). Full bio
News
Our new paper on traffic matching on mixnets will appear in PETS'24. Read our blog post!
Our work on the effects of retraining on AI-generated data will be featured in the Springer magazine.
New paper on the fair exposure problem under homophily was presented at AAAI'23 by Jakob Schoeffer.
I presented our new work on Privately Measuring Demographic Performance Disparities in Federated Learning. See tweet.
Research
My research topics and a sample of relevant work:
ML-Based Traffic Analysis: the development and evaluation of traffic analysis defenses from both a practical and theoretical point of view.
S. Siby, M. Juarez, C. Diaz, N. Vallina-Rodriguez, and C. Troncoso. "Encrypted DNS ⇒Privacy? A Traffic Analysis Perspective." NDSS, 2020.
Security and Privacy of ML: the design of methods to audit privacy and security of ML models, and the development of privacy-aware ML techniques.
M. Juarez, S. Yeom, and M. Fredrikson. "Black-Box Audits for Group Distribution Shifts." arXiv pre-print, 2022.
Privacy and Fairness: the privacy challenges that arise from identifying and mitigating algorithmic bias.
M. Juarez, and A. Korolova. "'You Can’t Fix What You Can’t Measure': Privately Measuring Demographic Performance Disparities in Federated Learning." Proceedings of the Algorithmic Fairness through the Lens of Causality and Privacy Workshop (NeurIPS), 2022.
publications.