Real-Time Active Learning for optimised spectroscopic follow-up: Enhancing early SN Ia classification with the Fink broker
A. Möller, E. E. O. Ishida, J. Peloton, O. Vidal Velázquez, J. Soon, B. Martin, M. Cluver, M. Leoni, E. Taylor · 2025
Read Real-Time Active Learning for optimised spectroscopic follow-up: Enhancing early SN Ia classification with the Fink broker by A. Möller, E. E. O. Ishida, J. Peloton, O. Vidal Velázquez, J. Soon, B. Martin, M. Cluver, M. Leoni, E. Taylor (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-19555.txt.
Description
Moller and co-authors report, in a 2025 paper, the first live use of active learning to choose which exploding stars deserve scarce telescope follow up. Surveys find far more transients than spectrographs can confirm. Machine classifiers sort them, but they train on few confirmed events, which caps their quality. Better training sets are the bottleneck.
Working with a real time alert broker on a public survey stream, the team applied an active learning strategy to early type Ia supernova candidates. A photometric classifier flagged the early events, and the strategy picked the most useful ones to observe. Follow up at an Australian telescope yielded 92 newly classified events for the training set. That set, built with a quarter fewer spectra than the public pool it was compared with, improved the classifier's metrics. Picks skewed to fainter events on average.
Catalog record
| CREATOR | A. Möller, E. E. O. Ishida, J. Peloton, O. Vidal Velázquez, J. Soon, B. Martin, M. Cluver, M. Leoni, E. Taylor |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 49 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-19555