About me
I am a postdoc in the SPRING Lab at EPFL, headed by Prof. Carmela Troncoso. My position is funded by the CYD Distinguished Postdoctoral Fellowship of the Cyber-Defense Campus, where my main collaborator is Dr. Raphael Meier.
I completed my PhD in the Computational Privacy Group at Imperial College London, advised by Dr. Yves-Alexandre de Montjoye.
My research sits broadly at the intersection between machine learning, privacy, and security. I develop methods to evaluate the privacy of anonymization techniques (re-pseudonymization, aggregation incl. query-based systems and synthetic data). I focus on automation, scalability, tightness of evaluation, and practical relevance of threat models, with the goal of making privacy evaluations more accessible to practitioners. I also study privacy and security vulnerabilities in systems having the potential to cause harm: machine learning models and perceptual hashing-based client-side scanning.
News
16/10/2024: š Distinguished paper award for our paper QueryCheetah: Fast Automated Discovery of Attribute Inference Attacks Against Query-Based Systems at ACM CCS 2024! Very glad to see this line of work, started with QuerySnout, recognized by the community.
23/08/2024: New paper QueryCheetah: Fast Automated Discovery of Attribute Inference Attacks Against Query-Based Systems accepted at ACM CCS 2024! Joint work with Bozhidar Stevanoski and Yves-Alexandre de Montjoye.
17/07/2024: Our review paper on Anonymization: The imperfect science of using data while preserving privacy has been published in Science Advances. Joint work with Andrea Gadotti, Luc Rocher, Florimond Houssiau and Yves-Alexandre de Montjoye.
10/07/2024: Our paper Correlation inference attacks against machine learning models has been published in Science Advances! Joint work with Florent GuƩpin and Yves-Alexandre de Montjoye.
20/06/2024: š Best paper award for our paper Re-pseudonymization Strategies for Smart Meter Data Are Not Robust to Deep Learning Profiling Attacks at the ACM CODASPY ā24 conference! Miruna Rusu, co-first author on the paper, presented the work in Porto, Portugal.
08/06/2024: New paper A Zero Auxiliary Knowledge Membership Inference Attack on Aggregate Location Data accepted in PoPETS 2024! Joint work with Vincent Guan, Florent GuƩpin and Yves-Alexandre de Montjoye.
More news here.
Publications
Peer-reviewed articles
* denotes joint first authorship.
- Bozhidar Stevanoski, Cretu A.-M. and de Montjoye Y.-A. QueryCheetah: Fast Automated Discovery of Attribute Inference Attacks Against Query-Based Systems. To appear in ACM CCS 2024. [Extended arXiv version] [Code] š Distinguished paper award
- Presented as a poster at the TPDP '24 workshop in Boston, MA, USA.
- Andrea Gadotti, Luc Rocher, Florimond Houssiau, CreČu A.-M. and de Montjoye Y.-A. Anonymization: The imperfect science of using data while preserving privacy. In Science Advances, 2024. [Paper]
- CreČu A.-M.*, GuĆ©pin F.* and de Montjoye Y.-A. Correlation inference attacks against machine learning models. In Science Advances, 2024. [Paper] [Code]
- Guan, V.*, GuĆ©pin F.*, CreČu A.-M. and de Montjoye Y.-A. A Zero Auxiliary Knowledge Membership Inference Attack on Aggregate Location Data. In Proceedings on Privacy Enhancing Technologies 2024(4) (PoPETS 2024). [Paper]
- Cretu A.-M., Jones, D., de Montjoye Y.-A., and Tople, S. Investigating the Effect of Misalignment on Membership Privacy in the White-box Setting. In Proceedings on Privacy Enhancing Technologies 2024(3) (PoPETS 2024). [Paper] [Code]
- Cretu A.-M.*, Rusu, M.*, and de Montjoye Y.-A. Re-pseudonymization Strategies for Smart Meter Data Are Not Robust to Deep Learning Profiling Attacks. In Proceedings of the Fourteenth ACM Conference on Data and Application Security and Privacy (CODASPY '24). [Paper] [Extended arXiv version] š Best paper award
- Cretu A.-M. Evaluating privacy and robustness in modern data processing systems. PhD Thesis, Imperial College London (2023). [Thesis]
- GuĆ©pin, F.*, Meeus, M.*, CreČu A.-M., and de Montjoye Y.-A. Synthetic is all you need: removing the auxiliary data assumption for membership inference attacks against synthetic data. In 18th DPM International Workshop on Data Privacy Management (DPM 2023). [Paper] [Code]
- Meeus, M.*, GuĆ©pin, F.*, CreČu A.-M., and de Montjoye Y.-A. Achilles' Heels: Vulnerable Record Identification in Synthetic Data Publishing. In 28th European Symposium on Research in Computer Security (ESORICS 2023). [Paper] [Code]
- Jain S., CreČu A.-M., Cully, A. and de Montjoye Y.-A. Deep perceptual hashing algorithms with hidden dual-purpose: when client-side scanning does facial recognition. In 2023 IEEE Symposium on Security and Privacy (SP). [Paper]
- Featured in Imperial News and Le soir.
- Cited as evidence in the Open Letter from Security and Privacy Researchers in relation to the Online Safety Bill.
- CreČu A.-M.*, Houssiau, F.*, Cully, A. and de Montjoye Y.-A. QuerySnout: Automating the discovery of attribute inference attacks against query-based systems. In Proceedings of the 2022 ACM SIGSAC Conference on Computer and Communications Security (CCS '22). [Paper] [Extended arXiv version] [Code]
- CreČu A.-M., Monti F., Marrone S., Dong X., Bronstein M. and de Montjoye Y.-A. Interaction data are identifiable even across long periods of time. Nature Communications 13, 313 (2022). [Paper]
- Presented at the ACM CCS Privacy Preserving Machine Learning 2021 workshop (PPML 2021). Contributed talk.
- Presented at the NeurIPS Privacy and Machine Learning 2021 workshop (PriML 2021).
- Featured in TechCrunch and Science News.
- Jain S.*, CreČu A.-M.* and de Montjoye Y.-A. Adversarial Detection Avoidance Attacks: Evaluating the robustness of perceptual hashing-based client-side scanning. 31st USENIX Security Symposium (USENIX Security 22) [Paper] [Extended arXiv version] [Code]
- Cited by Ofcom (UK's communications regulator) in their Overview of Perceptual Hashing Technology report.
- Presented at the NeurIPS Privacy and Machine Learning 2021 workshop (PriML 2021).
- Presented at the Conference on Applied Machine Learning for Information Security 2021 (CAMLIS 2021). Oral presentation.
- Presented as a talk at the 14th Workshop on Hot Topics in Privacy Enhancing Technologies (HotPETS 2021).
- Featured in Imperial College London News and Le soir.
- Kocijan V., Camburu O.-M., CreČu A.-M., Yordanov Y., Blunsom P. and Lukasiewicz T. WikiCREM: A Large Unsupervised Corpus for Coreference Resolution. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP) (2019) [Paper]
- Kocijan V., CreČu A.-M., Camburu O.-M., Yordanov Y. and Lukasiewicz T. A Surprisingly Robust Trick for the Winograd Schema Challenge. Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (ACL 2019) [Paper]
Preprints
Awards and scholarships
- I am a recipient of the Cyber-Defense Campus (CYD) Distinguished Postdoctoral Fellowship, which generously supports my postdoctoral position in the SPRING Lab at EPFL.
- I am a recipient of the USENIX '22 Diversity grant, which generously supported my trip to the conference in Boston.
- I am a recipient of the EPFL Excellence Fellowship (awarded to students with outstanding academic records), which generously supported my MSc studies at EPFL.
- My studies in France were supported by a competitive 2-year full scholarship from the Fondation Odon Vallet and by a 2.5-year scholarship from the Fondation de l'Ecole Polytechnique.
- I was born and grew up in BuzÄu, Romania. There, I participated in many mathematics contests, including the Romanian National Olympiad, where I won one gold medal, three silver medals, and one bronze medal.