From Simulations to Surveys: Domain Adaptation for Galaxy Observations
Not on the map. Why?
Abstract (arXiv)
Large photometric surveys will image billions of galaxies, but we currently lack quick, reliable automated ways to infer their physical properties like morphology, stellar mass, and star formation rates. Simulations provide galaxy images with ground-truth physical labels, but domain shifts in PSF, noise, backgrounds, selection, and label priors degrade transfer to real surveys. We present a preliminary domain adaptation pipeline that trains on simulated TNG50 galaxies and evaluates on real SDSS galaxies with morphology labels (elliptical/spiral/irregular). We train three backbones (CNN, E(2)-steerable CNN, ResNet-18) with focal loss and effective-number class weighting, and a feature-level domain loss LD built from GeomLoss (entropic Sinkhorn OT, energy distance, Gaussian MMD, and related metrics). We show that a combination of these losses with an OT-based "top_k soft matching" loss that focuses LD on the worst-matched source-target pairs can further enhance domain alignment. With Euclidean distance, scheduled alignment weights, and top-k matching, target accuracy (macro F1) rises from ~46% (~30%) at no adaptation to ~87% (~62.6%), with a domain AUC near 0.5, indicating strong latent-space mixing.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 27 of its 47 references that are in the map. First posted on arXiv 23 Nov 2025.
Builds on
27 of 47 references.
References · 27
- 2015Deep Residual Learning for Image Recognitioncited by 52,096
- 2014Unsupervised Domain Adaptation by Backpropagationcited by 1,917
- 2017Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semanticscited by 1,358
- 2019Class-Balanced Loss Based on Effective Number of Samplescited by 1,245
- 2018Deep Visual Domain Adaptation: A Surveycited by 635
- 2015Optimal Transport for Domain Adaptationcited by 463
- 2019General E(2)-Equivariant Steerable CNNscited by 446
- 2019Contrastive Adaptation Network for Unsupervised Domain Adaptationcited by 407
- 2018A Survey of Unsupervised Deep Domain Adaptationcited by 361
- 2021Domain Adaptation for Medical Image Analysis: A Surveycited by 352
- 2020A Brief Review of Domain Adaptationcited by 288
- 2016Correlation Alignment for Unsupervised Domain Adaptationcited by 147
15 more
- 2017Domain Adaptation for Visual Applications: A Comprehensive Surveycited by 128
- 2022Deep Unsupervised Domain Adaptation: A Review of Recent Advances and Perspectivescited by 98
- 2021Recent Advances in Domain Adaptation for the Classification of Remote Sensing Datacited by 70
- 2017Learning with Average Top-k Losscited by 52
- 2021Tune it the Right Way: Unsupervised Validation of Domain Adaptation via Soft Neighborhood Densitycited by 44
- 2021DeepAdversaries: Examining the Robustness of Deep Learning Models for Galaxy Morphology Classificationcited by 10
- 2021Unsupervised Domain Adaptation for Constraining Star Formation Historiescited by 8
- 2023DeepAstroUDA: Semi-Supervised Universal Domain Adaptation for Cross-Survey Galaxy Morphology Classification and Anomaly Detectioncited by 8
- 2020Domain adaptation techniques for improved cross-domain study of galaxy mergerscited by 7
- 2023Domain Adaptation based Object Detection for Autonomous Driving in Foggy and Rainy Weathercited by 7
- 2023E(2) Equivariant Neural Networks for Robust Galaxy Morphology Classificationcited by 7
- 2018A General Approach to Domain Adaptation with Applications in Astronomycited by 6
- 2023Domain Adaptive Graph Neural Networks for Constraining Cosmological Parameters Across Multiple Data Setscited by 6
- 2024Neural Network Prediction of Strong Lensing Systems with Domain Adaptation and Uncertainty Quantificationcited by 2
- 2025SIDDA: SInkhorn Dynamic Domain Adaptation for Image Classification with Equivariant Neural Networkscited by 2
Built on by
Not cited yet.