We study active correlation clustering where pairwise similarities are not provided upfront and must be queried in a cost-efficient manner through active learning. Specifically, we focus on the cold-start scenario, where no true initial pairwise similarities are available for active learning. To address this challenge, we propose a coverage-aware method that encourages diversity early in the process. We demonstrate the effectiveness of our approach through several synthetic and real-world experiments.
@article{arxiv.2509.25376,
title = {Cold-Start Active Correlation Clustering},
author = {Linus Aronsson and Han Wu and Morteza Haghir Chehreghani},
journal= {arXiv preprint arXiv:2509.25376},
year = {2026}
}