lix.polytechnique.fr

I am a professor in the Computer Science Department at École Polytechnique in France, and a Visiting Research Scientist at Google DeepMind. At École Polytechnique, I am a member of the GeomeriX team, and my research is currently supported by my ERC Consolidator Grant (VEGA). At DeepMind, I am involved in efforts related to video understanding.

My group's research focuses on developing mathematically principled approaches for geometric data analysis and synthesis. We tackle problems such as learning on surface meshes, 3D point clouds, and graphs; generative modeling of complex scenes; non-rigid shape matching; 3D reconstruction; and transfer learning, among others. A recurring theme in our work is learning structure-preserving maps and invariant operators across different domains. Beyond computer graphics and vision, I am also interested in using deep learning to address challenges in other scientific fields, and to facilitate scientific discovery. You can find some of my work on the Publications page.

Recent News

July 2026

June 2026

    I gave a keynote at the 3DS (Dassault Systèmes) Corporate Research Day, on The journey across representations: 3D data and deep learning over the years.

    I gave a talk at the PEPR IA Days 2026 on Measuring and Improving Alignment Between Neural Representations and the Physical World.

    We have 2 papers accepted at ECCV 2026:

    • Beyond Prompts: Unconditional 3D Inversion for Out-of-Distribution Shapes with Victoria Yue Chen, Emery Pierson, and Léopold Maillard.
    • From Blobs to Spokes: High-Fidelity Surface Reconstruction via Oriented Gaussians with Diego Gomez, Antoine Guedon, Nissim Maruani, and Bingchen Gong.

May 2026

    Jing Ren, who did a part of her PhD in our group has just won the Eurographics Young Researcher Award. Additionally, she has just received a position as an Assistant Professor at the Université de Montréal. Congratulations to Jing!

    We have 4 papers accepted at ICML 2026:

    • Beyond ReLU: Bifurcation, Oversmoothing, and Topological Priors with Erkan Turan, Gaspard Abel, Maysam Behmanesh, and Emery Pierson (as a Spotlight).
    • Unfolding Generative Flows with Koopman Operators: Trajectory-Preserving Linearization with Erkan Turan, Ari Siozopoulos, Louis Martinez, Julien Gaubil and Emery Pierson.
    • Graph Alignment via Dual-Pass Spectral Encoding and Latent Space Communication with Maysam Behmanesh and Erkan Turan.
    • The Art of Interrogation: Consistency Amplifies Factuality in Spatial Reasoning with Théo Uscidda, Marta Tintore Gazulla, Federico Tombari, and Leonidas Guibas.

February 2026

    I'm very happy to be part of five papers accepted at CVPR 2026:

    • Unique Lives, Shared World: Learning from Single-Life Videos with Tengda Han, Sayna Ebrahimi, Dilara Gokay, Li Yang Ku, Iva Babukova, Daniel Zoran, Viorica Patraucean, Joao Carreira, Andrew Zisserman, and Dima Damen.
    • A Mixed Diet Makes DINO an Omnivorous Vision Encoder with Rishabh Kabra, Drew A. Hudson, Ye Xia, Skanda Koppula, Andre Araujo, Joao Carreira and Niloy Mitra (as a Highlight).
    • PatchAlign3D: Local Feature Alignment for Dense 3D Shape understanding with Souhail Hadgi, Bingchen Gong, Ramana Sundararaman, Emery Pierson, Lei Li, and Peter Wonka.
    • FILTR: Extracting Topological Features from Pretrained 3D Models with Louis Martinez (as an Oral).
    • PaNDaS: Learnable Shape Interpolation Modeling with Localized Control with Thomas Besnier, Emery Pierson, Sylvain Arguillere, and Mohamed Daoudi.

January 2026

    We have two papers accepted at ICLR 2026:

archive >>

Read the original on lix.polytechnique.fr ↗