Rubin Difference Detectives: Finding Truth and Expertise
Citizen science and expert labelling for LSST transient classification
The Rubin Difference Detectives is a Zooniverse citizen science project launched in November 2025, tasking volunteers with real/bogus classification of LSST transient alert stamp triplets. A parallel expert labelling project ran alongside it, restricted to experienced annotators from the QUB/Oxford Transient Group.
The project investigates how citizen scientist labelling compares to expert annotation, using Cohen’s κ to measure inter-rater agreement. The resulting gold standard label set will be used to train and evaluate machine learning classifiers for the LSST alert stream.
Status: Work in progress — paper in preparation. Zooniverse: Rubin Difference Detectives