Enregistré dans:
| Auteurs principaux: | , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | https://arxiv.org/abs/2511.11787 |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866909903481733120 |
|---|---|
| author | Hassan, Sultan Andrianomena, Sambatra Wandelt, Benjamin D. |
| author_facet | Hassan, Sultan Andrianomena, Sambatra Wandelt, Benjamin D. |
| contents | Systematics contaminate observables, leading to distribution shifts relative to theoretically simulated signals-posing a major challenge for using pre-trained models to label such observables. Since systematics are often poorly understood and difficult to model, removing them directly and entirely may not be feasible. To address this challenge, we propose a novel method that aligns learned features between in-distribution (ID) and out-of-distribution (OOD) samples by optimizing a feature-alignment loss on the representations extracted from a pre-trained ID model. We first experimentally validate the method on the MNIST dataset using possible alignment losses, including mean squared error and optimal transport, and subsequently apply it to large-scale maps of neutral hydrogen. Our results show that optimal transport is particularly effective at aligning OOD features when parity between ID and OOD samples is unknown, even with limited data-mimicking real-world conditions in extracting information from large-scale surveys. Our code is available at https://github.com/sultan-hassan/feature-alignment-for-OOD-generalization. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_11787 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Towards Mitigating Systematics in Large-Scale Surveys via Few-Shot Optimal Transport-Based Feature Alignment Hassan, Sultan Andrianomena, Sambatra Wandelt, Benjamin D. Instrumentation and Methods for Astrophysics Computer Vision and Pattern Recognition Machine Learning Systematics contaminate observables, leading to distribution shifts relative to theoretically simulated signals-posing a major challenge for using pre-trained models to label such observables. Since systematics are often poorly understood and difficult to model, removing them directly and entirely may not be feasible. To address this challenge, we propose a novel method that aligns learned features between in-distribution (ID) and out-of-distribution (OOD) samples by optimizing a feature-alignment loss on the representations extracted from a pre-trained ID model. We first experimentally validate the method on the MNIST dataset using possible alignment losses, including mean squared error and optimal transport, and subsequently apply it to large-scale maps of neutral hydrogen. Our results show that optimal transport is particularly effective at aligning OOD features when parity between ID and OOD samples is unknown, even with limited data-mimicking real-world conditions in extracting information from large-scale surveys. Our code is available at https://github.com/sultan-hassan/feature-alignment-for-OOD-generalization. |
| title | Towards Mitigating Systematics in Large-Scale Surveys via Few-Shot Optimal Transport-Based Feature Alignment |
| topic | Instrumentation and Methods for Astrophysics Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2511.11787 |