Photometric classification and out-of-taxonomy detection of ZTF transients with FLARE
Abstract
Photometric classifiers decide which of the million nightly alerts of time-domain surveys receive one of the few thousand spectra taken each year, and they are the first place where a transient outside the trained classes can be noticed. We present flare, the photometric classification and anomaly-detection module of AppleCiDEr, evaluated on 10 497 spectroscopically confirmed ZTF transients in six classes and 53 objects from families outside the taxonomy. Catalogue context raises the macro F1 score, with the largest gains where the defining information is absent from the light curve. Blind detection recovers few out-of-taxonomy objects at a 1% false-alarm rate, while a simulated review loop recovers 83% of them after inspecting 15% of the stream. The first confirmations of a new family are most useful in the feature space: with the same eleven labels, a head on the frozen features detects 68% of the tidal disruption events hidden from the training against 57% for a combiner over the anomaly scores, and retraining with them as a new class reaches 85%. The module runs in production on Fritz, the ZTF instance of SkyPortal, and its alarm is reported as a false-alarm rate per unit time on the stream, calibrated on the photometry that the service receives
Related
- Unveiling photometric transients with the ILMT and their follow-up using DOT
- Photometric characterization of near-Earth asteroids: rotation, taxonomy, and candidate evidence for internal cohesion
- Exoplanet detection through photometric orbital phase modulations
- FOMO-Z. Optimal galaxy classification, photometric redshifts, and reconstruction of population distributions
- Image-catalog complementarity in quasar-galaxy ambiguous regimes of photometric source classification
BibTeX
@misc{sasli2026photometric,
title = {{Photometric classification and out-of-taxonomy detection of ZTF transients with FLARE}},
author = {Argyro Sasli and Felipe Fontinele Nunes and Xuanrui Zhang and Dimitrios Kourtesis and Saina Nikmehr and Alexandra Junell and Jesper Sollerman and Antoine Le Calloch and Sushant Sharma Chaudhary and Sneha Maharjan and Maojie Xu and Avyukt Raghuvanshi and Steven L. Groom and Joahan Castaneda Jaimes and Josiah Purdum and Roger Smith and Mansi M. Kasliwal and Matthew J. Graham and Ashish A. Mahabal and Michael Coughlin},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.30656},
eprint = {2610.30656},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.30656},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Photometric classification and out-of-taxonomy detection of ZTF transients with FLARE
AU - Argyro Sasli
AU - Felipe Fontinele Nunes
AU - Xuanrui Zhang
AU - Dimitrios Kourtesis
AU - Saina Nikmehr
AU - Alexandra Junell
AU - Jesper Sollerman
AU - Antoine Le Calloch
AU - Sushant Sharma Chaudhary
AU - Sneha Maharjan
AU - Maojie Xu
AU - Avyukt Raghuvanshi
AU - Steven L. Groom
AU - Joahan Castaneda Jaimes
AU - Josiah Purdum
AU - Roger Smith
AU - Mansi M. Kasliwal
AU - Matthew J. Graham
AU - Ashish A. Mahabal
AU - Michael Coughlin
PY - 2026
DA - 2026/10/22
PB - arXiv
DO - 10.48550/arxiv.2610.30656
UR - https://arxiv.org/abs/2610.30656
AB - Photometric classifiers decide which of the million nightly alerts of time-domain surveys receive one of the few thousand spectra taken each year, and they are the first place where a transient outside the trained classes can be noticed. We present flare, the photometric classification and anomaly-detection module of AppleCiDEr, evaluated on 10 497 spectroscopically confirmed ZTF transients in six classes and 53 objects from families outside the taxonomy. Catalogue context raises the macro F1 score, with the largest gains where the defining information is absent from the light curve. Blind detection recovers few out-of-taxonomy objects at a 1% false-alarm rate, while a simulated review loop recovers 83% of them after inspecting 15% of the stream. The first confirmations of a new family are most useful in the feature space: with the same eleven labels, a head on the frozen features detects 68% of the tidal disruption events hidden from the training against 57% for a combiner over the anomaly scores, and retraining with them as a new class reaches 85%. The module runs in production on Fritz, the ZTF instance of SkyPortal, and its alarm is reported as a false-alarm rate per unit time on the stream, calibrated on the photometry that the service receives
ER -