ORC Seminar: OR with a White Hat: Evidencing Privacy Vulnerabilities in ML Models

Thibaut Vidal Professor École Polytechnique de Montréal Abstract: The deployment of machine learning models in high-risk domains (e.g., finance, medicine) raises questions about the privacy of the data used to train them. In this talk, I will show that operations research methods can provide a rigorous methodological backbone for auditing and mitigating certain privacy risks in ML models. I will first discuss a white-box reconstruction attack that formulates the recovery of a random forest’s training data as a combinatorial problem solved with constraint programming. This approach reconstructs entire datasets with high accuracy, even from forests with only a few trees. Next, we turn to black-box access and explainability-driven interfaces. Counterfactual explanations (increasingly needed and exposed through ML APIs) represent a powerful attack surface. Using tools from online optimization and competitive analysis, we derive tight bounds on the number of counterfactual queries required to extract tree-based models and introduce new algorithms achieving provably perfect fidelity. Finally, we will examine the protection offered by differential privacy. Focusing on ε-DP random forests, we demonstrate that even models satisfying strict DP guarantees can still leak meaningful, dataset-specific information in practice, unless the privacy noise is increased to the point where the model loses most of its predictive value. Overall, the talk highlights critical tensions between predictive performance, explainability, and privacy protection, and showcases OR-based techniques as powerful tools for navigating these trade-offs. Related references : Ferry, J., Fukasawa, R., Pascal, T., & Vidal, T. (2024). Trained random forests completely reveal your dataset. ICML’24 (oral). http://arxiv.org/abs/2402.19232 Khouna, A., Ferry, J., & Vidal, T. (2025). From counterfactuals to trees: Competitive analysis of model extraction attacks. NeurIPS’25 (spotlight). http://arxiv.org/abs/2502.05325 Gorgé, A., Ferry, J., Gambs, S., & Vidal, T. (2025). Training set reconstruction from differentially private forests: How effective is DP? SATML’26 http://arxiv.org/abs/2502.05307 Bio: Thibaut Vidal, PhD, holds the SCALE-AI Chair in Data-Driven Supply Chains and is a full professor at the Department of Mathematics and Industrial Engineering (MAGI) of Polytechnique Montréal, Canada. He is also a member of CIRRELT and Hi! PARIS International Visiting Chair. His expertise lies in combinatorial optimization and trustworthy machine learning, with applications spanning transportation, supply chain management, resource allocation, and information processing. He has authored over 80 peer-reviewed studies published in leading journals and conferences in operations research and machine learning, such as NeurIPS, ICML, Operations Research, Transportation Science, and SIAM Journal on Optimization, among many others. His academic contributions include developing state-of-the-art algorithms, many of which are accessible through open-source libraries, facilitating advancements in data-driven logistics operations. He has also collaborated with various stakeholders to enhance data analytics and logistics performance through consulting projects. Among others, he has been the recipient of two best paper awards from the Transportation Science and Logistics section of INFORMS, as well as the Robert Faure prize from the French Operations Research society.