Semi-supervised discriminative clustering with graph regularization

Śmieja, Marek; Myronov, Oleksandr; Tabor, Jacek

doi:10.1016/j.knosys.2018.03.019

Cited by 16 publications

(7 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our work extends recent discriminative SCC methods [11,10] by learning additional pairwise relations. Moreover, the approach is implemented using Siamese neural networks [12,32,13], allowing for higher flexibility.…”

Section: Related Workmentioning

confidence: 78%

“…• d-graph: this is a DNN-based implementation of d-graph [11]. The network architecture is identical to CluNet (the batch structure is also the same).…”

Section: Comparison With Related Modelsmentioning

confidence: 99%

“…This requires transferring class-label knowledge into a clustering algorithm, which is often unnatural and puts a higher weight on clustering structure than on class labels. It has been recently shown that discriminative clustering methods, which approach clustering problems using classification tools, are usually more effective in taking advantage of label constraints/information [10,11]. While those formulations assume that class labels are the primary source of semi-supervision, it might be difficult to produce satisfactory results in the presence of a large number of clusters.…”

Section: Introductionmentioning

confidence: 99%

“…shown that discriminative clustering methods, which approach clustering problems using classification tools, are usually more effective in taking advantage of label constraints/information [10,11]. While those formulations assume that class labels are the primary source of semi-supervision, it might be difficult to produce satisfactory results in the presence of a large number of clusters.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A classification-based approach to semi-supervised clustering with pairwise constraints

2020

Self Cite

View full text Add to dashboard Cite

In this paper, we introduce a neural network framework for semi-supervised clustering (SSC) with pairwise (must-link or cannot-link) constraints. In contrast to existing approaches, we decompose SSC into two simpler classification tasks/stages: the first stage uses a pair of Siamese neural networks to label the unlabeled pairs of points as must-link or cannot-link; the second stage uses the fully pairwise-labeled dataset produced by the first stage in a supervised neural-network-based clustering method.The proposed approach, S 3 C 2 (Semi-Supervised Siamese Classifiers for Clustering), is motivated by the observation that binary classification (such as assigning pairwise relations) is usually easier than multi-class clustering with partial supervision. On the other hand, being classification-based, our method solves only well-defined classification problems, rather than less well specified clustering tasks. Extensive experiments on various datasets demonstrate the high performance of the proposed method.

show abstract

Section: Related Workmentioning

confidence: 78%

“…• d-graph: this is a DNN-based implementation of d-graph [11]. The network architecture is identical to CluNet (the batch structure is also the same).…”

Section: Comparison With Related Modelsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A classification-based approach to semi-supervised clustering with pairwise constraints

2020

Self Cite

View full text Add to dashboard Cite

show abstract

“…for a (small) subset of pairs (M ∪ N ) ⊂ X 2 . Pairwise constraints have been utilized for discriminative clustering with graph regularization [18], GMMs [12], or spectral clustering [11].…”

Section: Problem Statement and Related Workmentioning

confidence: 99%

Semi-Supervised Clustering via Information-Theoretic Markov Chain Aggregation

Steger,

Geiger,

Smieja

2021

Preprint

View full text Add to dashboard Cite

We connect the problem of semi-supervised clustering to constrained Markov aggregation, i.e., the task of partitioning the state space of a Markov chain. We achieve this connection by considering every data point in the dataset as an element of the Markov chain's state space, by defining the transition probabilities between states via similarities between corresponding data points, and by incorporating semi-supervision information as hard constraints in a Hartigan-style algorithm. The introduced Constrained Markov Clustering (CoMaC) is an extension of a recent information-theoretic framework for (unsupervised) Markov aggregation to the semisupervised case. Instantiating CoMaC for certain parameter settings further generalizes two previous information-theoretic objectives for unsupervised clustering. Our results indicate that CoMaC is competitive with the state-of-the-art. Furthermore, our approach is less sensitive to hyperparameter settings than the unsupervised counterpart, which is especially attractive in the semi-supervised setting characterized by little labeled data.

show abstract