Lucas Degeorge

PhD Student working on Generative Modeling

I am a PhD student in Computer Vision at VISTA (Ecole polytechnique) and IMAGINE (Ecole des Ponts) labs, under the supervision of Vicky Kalogeiton and David Picard.

My research focuses on generative models for images, especially text-to-image generation and flow matching. I am interested in making these models more data- and compute-efficient and training competitive text-to-image models. I am also interested in vision representation learning.

News

Oct 2026
Our new work Dino Forcing Flow Models: Do not denoise what you can predict is out! Check out the code.
Sep 2026
Our work Balancing Frequencies and Pixels in Flow Matching has been accepted at NeurIPS 2026. See you in Paris!
Jul 2026
I attended ICML 2026 in Seoul to present MIRO.
Apr 2026
Our work MIRO has been accepted at ICML 2026. See you in Seoul!

Show all news (10)

Feb 2026
Our paper PoM has been accepted as a Findings paper at CVPR 2026!
Dec 2025
I am attending EurIPS 2025 to present our work at the Workshop on Principles of Generative Modeling (PriGM)
Oct 2025
Our new work MIRO: MultI-Reward cOnditioned pretraining improves T2I quality and efficiency is online. Check it out!
Oct 2025
We release the dataset used in the paper How far can we go with ImageNet for text-to-image generation?. Download them on HuggingFace and use them!
Nov 2024
I started my PhD at VISTA (Ecole polytechnique) and IMAGINE (Ecole des Ponts) labs.
Sep 2024
I attended ECCV 2024 in Milano.

Publications

Also on Google Scholar.

2026

Dino Forcing Flow Models: Do not denoise what you can predict

Arijit Ghosh^*, Lucas Degeorge^*, Paul Couairon, Alexei A. Efros, Vicky Kalogeiton, David Picard

arXiv preprint · 2026^*equal contribution

Paper Code

Balancing Frequencies and Pixels in Flow Matching

Lucas Degeorge^*, Paul Couairon^*, Arijit Ghosh^*, Alexei A. Efros, David Picard, Vicky Kalogeiton

NeurIPS · 2026^*equal contribution

Project page Paper

MIRO: MultI-Reward cOnditioned pretraining improves T2I quality and efficiency

Nicolas Dufour, Lucas Degeorge, Arijit Ghosh, Vicky Kalogeiton, David Picard

ICML · 2026

Project page Paper Code Models Demo

PoM: A Linear-Time Replacement for Attention with the Polynomial Mixer

David Picard, Nicolas Dufour, Lucas Degeorge, Arijit Ghosh, Davide Allegro, Tom Ravaud, Yohann Perrion, Corentin Sautier, Zeynep Sonat Baltaci, Fei Meng, Syrine Kalleli, Marta López-Rauhut, Thibaut Loiseau, Ségolène Albouy, Raphael Baena, Elliot Vincent, Loic Landrieu

CVPR Findings · 2026

Paper Code

2025

How far can we go with ImageNet for text-to-image generation?

Lucas Degeorge^*, Arijit Ghosh^*, Nicolas Dufour, David Picard, Vicky Kalogeiton

arXiv preprint · 2025^*equal contribution

Project page Paper Code Models Dataset

One-step Diffusion Models with Bregman Density Ratio Matching

Yuanzhi Zhu, Eleftherios Tsonis, Lucas Degeorge, Vicky Kalogeiton

EurIPS 2025 Workshop on Principles of Generative Modeling (PriGM) · 2025

Paper

Misc

Teaching

2024 – 2026
Advanced Deep LearningTeaching Assistant · Ecole polytechnique
2024 – 2026
Deep Learning for Computer VisionTeaching Assistant · Ecole polytechnique

Reviewer

2027
ICLR
2026
NeurIPS Top ReviewerECCVCVPR