Marina Meila
About
Appointed Canada CIFAR AI Chair – 2025
Marina Meilă’s primary research interest lies in statistical learning and the discovery of geometric and combinatorial structure within data using efficient algorithms. Her current work focuses on the validation of unsupervised structure learning (such as dimension reduction and dependency structure). She pursues this through two main directions: developing rigorous algorithms to interpret low-dimensional structures using meaningful scientific concepts, and creating new paradigms that guarantee a learned structure is correct, requiring minimal assumptions about the data-generating process.
Additionally, she has developed new models for the statistical analysis of preferences and is currently working on a quantitative theory of explanation. Collaborations often inspire her research questions in applied inverse problems, materials science, and theoretical chemistry.
Relevant Publications
- Meilă, M., & Zhang, H. (2024). “Manifold Learning: What, How and Why”. Annual Reviews of Statistics and Its Application.
- Meilă, M., & Chen, Y.-C. (2022). “Manifold Coordinates with Physical Meaning.” Journal of Machine Learning Research, 23(343), 1–68.
- Meilă, M. (2017). “How to Tell When a Clustering Is (Approximately) Correct Using Convex Relaxations.” In Neural Information Processing Systems (NeurIPS).
- Meilă, M. (2007). “Comparing clusterings—an information based distance”. Journal of multivariate analysis, 98(5), 873–895.
- Meilă, M., & Shi, J. (2001). “A random walks view of spectral segmentation.” In International workshop on artificial intelligence and statistics (AISTATS).