Correlation Clustering with Active Learning of Pairwise Similarities

TL;DR

This paper introduces an active learning framework for correlation clustering that efficiently queries pairwise similarities, allowing flexible feedback, noise robustness, and adaptability to various algorithms, demonstrated through extensive experiments.

Contribution

It presents a novel active learning approach for correlation clustering, including new query strategies and a flexible, noise-robust framework adaptable to different algorithms.

Findings

Effective in reducing the number of queries needed

Flexible feedback mechanisms improve clustering accuracy

Robust to noisy similarity data

Abstract

Correlation clustering is a well-known unsupervised learning setting that deals with positive and negative pairwise similarities. In this paper, we study the case where the pairwise similarities are not given in advance and must be queried in a cost-efficient way. Thereby, we develop a generic active learning framework for this task that benefits from several advantages, e.g., flexibility in the type of feedback that a user/annotator can provide, adaptation to any correlation clustering algorithm and query strategy, and robustness to noise. In addition, we propose and analyze a number of novel query strategies suited to this setting. We demonstrate the effectiveness of our framework and the proposed query strategies via several experimental studies.