Built independently by an author, for readers. Read the story and support ChapterPal
ChapterPal Logo
Log inSign up

keyword

co-regularized multi-view spectral clustering

Co-regularized multi-view spectral clustering is an unsupervised machine learning technique that groups data objects described across multiple distinct feature representations, or views, by jointly learning spectral embeddings that remain consistent across those views. Rather than clustering each view independently or concatenating all features into a single representation, the method simultaneously optimizes graph-based spectral clustering objectives for each individual view while applying a co-regularization penalty. This penalty explicitly discourages disagreement among the low-dimensional eigenvectors or cluster indicator matrices derived from the separate graph Laplacians, encouraging the different views to converge toward a shared consensus partition. By harmonizing view-specific spectral structures and exploiting the complementary information contained in diverse data modalities, this approach produces a robust and unified clustering solution that outperforms individual single-view partitions.

1 item

Co-regularized Multi-view Spectral Clustering

Co-regularized Multi-view Spectral Clustering

Abhishek Kumar, Piyush Rai, Hal Daumé

OrganizationsUniversity of MarylandUniversity of Utah

Why you should read this

Proposes a multi-view spectral clustering framework that co-regularizes graph Laplacians across different data representations to find consistent cluster assignments across diverse views.

In many clustering problems, we have access to multiple views of the data each of which could be individually used for clustering. Exploiting information from multiple views, one can hope to find a clustering that is more accurate than the ones obtained using the individual views. Often these different views admit same underlying clustering of the data, so we can approach this problem by looking for clusterings that are consistent across the views, i.e., corresponding data points in each view should have same cluster membership. We propose a spectral clustering framework that achieves this goal by co-regularizing the clustering hypotheses, and propose two co-regularization schemes to accomplish this. Experimental comparisons with a number of baselines on two synthetic and three real-world datasets establish the efficacy of our proposed approaches.

Added

2026-09-25