Reproduce “A Model for Imbalanced Label Aggregation: A Focus on Minority-Class Detection” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Gabriel Singer, Samuel Gruffaz, Olivier Vo Van, Nicolas Vayatis, Argyris Kalogeratos
CC-Rasch aggregates crowdsourced labels by modeling workers’ class-specific ability and items’ class-specific difficulty. It estimates unknown labels from votes using an EM procedure. On three multiclass datasets with imbalance of at least three, it reports mean minority recall of 0.370 ± 0.257, versus 0.296 ± 0.128 for Dawid–Skene. This is a 7.4-point gain, not a uniform win across subgroups or metrics.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.