Marco Cuturi

Marco Cuturi is a Research Scientist at Apple ML Research in Paris and Professor of Statistics at CREST-ENSAE, Institut Polytechnique de Paris. His work bridges machine learning, optimal transport, and optimization, with applications in time-series analysis, kernels, and multiresolution methods. He has held academic roles at Kyoto University and Princeton University, and previously worked in the financial industry.

Research Interests:

• Optimal transport theory and computational methods
• Kernel design for structured data and histograms
• Time-series alignment and soft-DTW
• Entropic regularization in optimization
• Applications to computer vision and genomics

Teaching: Cuturi has taught courses on linear optimization at Princeton, geometric methods in machine learning at Kyoto, and scientific English. He has also organized machine learning summer schools in Kyoto, Les Houches, and other international venues.

Recent Trends: His 2024-2025 publications focus on entropic optimal transport solvers, disentangled representation learning via Gromov-Monge gaps, and applications to text-to-image diffusion models. Collaborative work with institutions like Google Research, MIT, and University of Tokyo highlights his interdisciplinary impact.