I am a postdoctoral researcher at Meta FAIR, working on world models with Yann LeCun. My research interests lie at the intersection of machine learning and physics; my current focus is on developing efficient algorithms that enable agents to perceive, understand, and predict the world around them, in order to interact with it. Previously, I completed my PhD at Ecole normale supérieure (DIENS) and Inria (WILLOW & THOTH), advised by Jean Ponce, Julien Mairal and Olivier Flasseur. There, I designed physically informed deep learning architectures for scientific data, with a particular emphasis on the detection of exoplanets via direct imaging. In collaboration with astronomers Anne-Marie Lagrange and Maud Langlois, I applied these methods to observations from the ESO's Very Large Telescope (VLT) in Chile.
News
- September 2025: I started my postdoc at Meta FAIR
- August 2025: Our work was featured by Jean Ponce on the tech channel Underscore_
- June 2025: Our paper Evidence for a sub-jovian planet in the young TWA7 disk was accepted in Nature 🎉
- June 2025: I successfully defended my PhD! 🎉 My thesis is available here
- March 2025: Our paper ExoMILD was accepted to CVPR 2025 🍾
- September 2024: Our paper MODEL&CO was accepted to MNRAS 🎊
- August 2023: Our paper deep PACO was accepted to MNRAS 🥳
- November 2021: I started my PhD
Publications
Main Publications
| | A New Statistical Model of Star Speckles for Learning to Detect and Characterize Exoplanets in Direct Imaging Observations Théo Bodrito, Olivier Flasseur, Julien Mairal, Jean Ponce, Maud Langlois, Anne-Marie Lagrange. CVPR, 2025 [arxiv] [code] |
| | MODEL&CO: Exoplanet detection in angular differential imaging by learning across multiple observations Théo Bodrito, Olivier Flasseur, Julien Mairal, Jean Ponce, Maud Langlois, Anne-Marie Lagrange. MNRAS, 2024 [arxiv] |
| | deep PACO: Combining statistical models with deep learning for exoplanet detection and characterization in direct imaging at high contrast Olivier Flasseur, Théo Bodrito, Julien Mairal, Jean Ponce, Maud Langlois, Anne-Marie Lagrange. MNRAS, 2024 [arxiv] |
| | A Trainable Spectral-Spatial Sparse Coding Model for Hyperspectral Image Restoration Théo Bodrito*, Alexandre Zouaoui*, Jocelyn Chanussot, Julien Mairal. NeurIPS, 2021 [arxiv] [code] |
Other Publications
| | Evidence for a sub-jovian planet in the young TWA7 disk A.-M. Lagrange, C. Wilkinson, M. Malin, A. Boccaletti, C. Perrot, L. Matrà, F. Combes, D. Rouan, H. Beust, A. Chomez, B. Charnay, S. Mazevet, O. Flasseur, J. Olofsson, A. Bayo, Q. Kral, G. Chauvin, P. Thebault, P. Rubini, J. Milli, F. Kiefer, A. Carter, K, Crotts, A. Radcliffe, J. Mazoyer, T. Bodrito , S. Stasevic, P. Delorme, M. Langlois Nature, 2025 [paper] [arxiv] |
| | Fine Dense Alignment of Image Bursts through Camera Pose and Depth Estimation Bruno Lecouat*, Yann Dubois de Mont-Marin,∗ Théo Bodrito* , Julien Mairal, Jean Ponce arxiv, 2024 [arxiv] |
| | Physical Simulation Layer for Accurate 3D Modeling Mariem Mezghanni, Théo Bodrito , Malika Boulkenafed, Maks Ovsjanikov CVPR, 2022 (oral) [paper] |
