When Foundation Models Meet Astronomical Data

Jan 14, 2025·
Matthieu Le Lain
Matthieu Le Lain
· 1 min read
Abstract
Evaluation of the performance of existing visual foundation models, including ViT, CLIP and DINOv2, for astronomical tasks. We adopt a common methodology based on fine-tuning and report results on galaxy classification datasets.
Date
Jan 14, 2025 — Jan 16, 2025
Location

Tromsø, Norway

events

Poster presented at the Northern Lights Deep Learning Conference 2025 in Tromsø, Norway. This work evaluates the performance of existing visual foundation models (ViT, DINOv2, CLIP) for astronomical data analysis, with a focus on galaxy morphological classification.

Foundation Models Computer Vision Galaxy Classification Poster
Matthieu Le Lain
Authors
Matthieu Le Lain
AI for astronomy & astrophysics
PhD student at IRISA, Université Bretagne Sud (expected 2026), and lecturer in computer science, working on foundation models for astronomy and astrophysics, with a broader interest in deep learning applied to scientific data. Also contributing to the UniverseTBD collaboration on vision–language models for astronomy.