Abstract
Galaxy mergers are essential to understanding galaxy evolution. We are developing a pipeline that aggregates volunteer annotations into segmentation masks, fine-tunes Zoobot on Euclid images to predict feature-level vote fractions and uses Zoobot’s attention maps (refined with a lightweight U-Net) to generate input masks for all training images. These masks support a semantic-segmentation model that maps merger substructures and predicts feature vote fractions. Leveraging MSI for large-scale training, this framework enables construction of a Euclid-based merger catalogue and future studies of gas flows, AGN fueling, and feedback.
| Original language | English (US) |
|---|---|
| Title of host publication | PEARC 2026 - Practice and Experience in Advanced Research Computing 2026 |
| Subtitle of host publication | Resilient Roots + Empowered Communities |
| Publisher | Association for Computing Machinery, Inc |
| ISBN (Electronic) | 9798400723773 |
| DOIs | |
| State | Published - Jul 26 2026 |
| Event | Practice and Experience in Advanced Research Computing, PEARC26 - Minneapolis, United States Duration: Jul 26 2026 → Jul 30 2026 |
Publication series
| Name | PEARC 2026 - Practice and Experience in Advanced Research Computing 2026: Resilient Roots + Empowered Communities |
|---|
Conference
| Conference | Practice and Experience in Advanced Research Computing, PEARC26 |
|---|---|
| Country/Territory | United States |
| City | Minneapolis |
| Period | 7/26/26 → 7/30/26 |
Bibliographical note
Publisher Copyright:© 2026 Copyright held by the owner/author(s).
Keywords
- citizen science
- Euclid
- Galaxy mergers
- Galaxy Zoo
- HPC Link to poster: PEARC Poster 2026
- machine learning
- semantic segmentation
Fingerprint
Dive into the research topics of 'A Scalable Machine Learning Pipeline for Identifying Galaxy Merger Features Using Volunteer Annotations'. Together they form a unique fingerprint.Cite this
- APA
- Standard
- Harvard
- Vancouver
- Author
- BIBTEX
- RIS