Practicum Experience(s)
Brookhaven National Laboratory (2024)
Summary of Research
Machine learning holds great promise for automating decision-making in fields such as healthcare, climate science, and public policy. However, most existing approaches are designed to recognize patterns and correlations in data, not to understand the underlying causes. This limits their ability to generalize, interpret results, and make reliable decisions, especially in high-stakes or data-limited settings.
My research lies at the intersection of machine learning, causal inference, and robust data-driven decision-making. I design algorithms that go beyond prediction to identify and reason about cause-and-effect relationships in complex environments. This includes developing methods for experiments with imperfect compliance, risk and sensitivity analysis in observational studies, and reinforcement learning in dynamic, interactive settings. I also focus on settings where data is structured across time or space, such as climate or health systems. Ultimately, my goal is to make machine learning models more trustworthy, interpretable, and effective by grounding them in causal reasoning—laying the foundation for secure and reliable decision-making across a range of real-world applications.
Annual Program Review Abstracts
Publications
M. Oprescu, D.K. Park, X. Luo, S. Yoo, and N. Kallus. "GST-UNet: Spatiotemporal Causal Inference with Time-Varying Confounders". Under Review, 2025.
M. Oprescu, N. Kallus. "Estimating Heterogeneous Treatment Effects by Combining Weak Instruments and Observational Data". In Advances in Neural Information Processing Systems, 2024.
A. Bennett, N. Kallus, M. Oprescu, W. Sun, K. Wang. "Efficient and Sharp Off-Policy Evaluation in Robust Markov Decision Processes". In Advances in Neural Information Processing Systems, 2024.
A. Bennett, N. Kallus, M. Oprescu. "Low-Rank MDPs with Continuous Action Spaces". In International Conference on Artificial Intelligence and Statistics, 2024.
M. Oprescu, J. Dorn, M. Ghoummaid, A. Jesson, N. Kallus, U. Shalit. "B-Learner: Quasi-Oracle Bounds on Heterogeneous Causal Effects Under Hidden Confounding". In International Conference on Machine Learning, 2023.
N. Kallus and M. Oprescu. "Robust and agnostic learning of conditional
distributional treatment effects". In International Conference on Artificial Intelligence and Statistics, 2023.
S. Mouatadid, P. Orenstein, G. Flaspohler, M. Oprescu, J. Cohen, F. Wang, S. Knight, M. Geogdzhayeva, S. Levang, E. Fraenkel, L. Mackey. "SubseasonalClimateUSA: a dataset for subseasonal forecasting and benchmarking". In Advances in Neural Information Processing Systems, 2023.
S. Mouatadid, P. Orenstein, G. Flaspohler, J. Cohen, M. Oprescu, E. Fraenkel, and L. Mackey. "Adaptive Bias Correction for Improved Subseasonal Forecasting". Nature Communications 14.1 (2023): 3482.
K. Battocchi, E. Dillon, M. Hei, G. Lewis, M. Oprescu, V. Syrgkanis. "Estimating the Long-Term Effects of Novel Treatments". In Advances in Neural Information Processing Systems, 2021.
S. Mouatadid, P. Orenstein, G. Flaspohler, M. Oprescu, J. Cohen, et al. "Learned Benchmarks for Subseasonal Forecasting". arXiv preprint arXiv:2109.10399 (2021). In Submission.
G. Flaspohler, F. Orabona, J. Cohen, S. Mouatadid, M. Oprescu, P. Orenstein, L. Mackey. "Online Learning with Optimism and Delay". In International Conference on Machine Learning. PMLR, 2021.
V. Syrgkanis, G. Lewis, M. Oprescu, M. Hei, K. Battocchi, et al. "Causal Inference and Machine Learning in Practice with EconML and CausalML: Industrial Use Cases at Microsoft, TripAdvisor, Uber". Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2021.
V. Syrgkanis, V. Lei, M. Oprescu, M. Hei, K. Battocchi, G. Lewis. "Machine learning estimation of heterogeneous treatment effects with instruments". In Advances in Neural Information Processing Systems, 2019. Spotlight presentation.
M. Oprescu, V. Syrgkanis, K. Battocchi, M. Hei, and G. Lewis. "EconML: A Machine Learning Library for Estimating Heterogeneous Treatment Effects". In CausalML Workshop, NeurIPS, 2019. Spotlight presentation.
M. Oprescu, V. Syrgkanis, Z. S. Wu. "Orthogonal random forest for causal inference". In International Conference on Machine Learning, 2019.
K. Arbour, M. Oprescu, J. Hakim, et al. "Multifactorial Model to Predict Response to PD-(L) 1 Blockade in Patients with High PD-L1 Metastatic Non-Small Cell Lung Cancer". Journal of Thoracic Oncology, 2019.
M. Hamilton, S. Raghunathan, M. Oprescu et al. "Flexible and Scalable Deep Learning with MMLSpark". In International Conference on Predictive Applications and APIs, 2018.
Awards
Meta PhD Research Fellowship Finalist, 2022
Cum Laude, Harvard University, 2015
High Honors, Harvard University Physics Department, 2015
Derek C. Bok Award for Distinction in Teaching (Data Science), Harvard, 2014
Excellence in Summer Research Award, Johns Hopkins University, 2014