A Classification-Based Approach to Semi-Supervised Clustering with Pairwise Constraints

In this paper, we introduce a neural network framework for semi-supervised\nclustering (SSC) with pairwise (must-link or cannot-link) constraints. In\ncontrast to existing approaches, we decompose SSC into two simpler\nclassification tasks/stages: the first stage uses a pair of Siamese neural\nnetworks to label the unlabeled pairs of points as must-link or cannot-link;\nthe second stage uses the fully pairwise-labeled dataset produced by the first\nstage in a supervised neural-network-based clustering method. The proposed\napproach, S3C2 (Semi-Supervised Siamese Classifiers for Clustering), is\nmotivated by the observation that binary classification (such as assigning\npairwise relations) is usually easier than multi-class clustering with partial\nsupervision. On the other hand, being classification-based, our method solves\nonly well-defined classification problems, rather than less well specified\nclustering tasks. Extensive experiments on various datasets demonstrate the\nhigh performance of the proposed method.\n