</> Code with Papers by Syntology

Papers › Contrastive Hierarchical Clustering

Contrastive Hierarchical Clustering

3 Mar 2023arXiv:2303.03389archive 2025-07-28

Michał Znaleźniak, Przemysław Rola, Patryk Kaszuba, Jacek Tabor, Marek Śmieja

Deep clustering has been dominated by flat models, which split a dataset into a predefined number of groups. Although recent methods achieve an extremely high similarity with the ground truth on popular benchmarks, the information contained in the flat partition is limited. In this paper, we introduce CoHiClust, a Contrastive Hierarchical Clustering model based on deep neural networks, which can be applied to typical image data. By employing a self-supervised learning approach, CoHiClust distills the base network into a binary tree without access to any labeled data. The hierarchical clustering structure can be used to analyze the relationship between clusters, as well as to measure the similarity between data points. Experiments demonstrate that CoHiClust generates a reasonable structure of clusters, which is consistent with our intuition and image semantics. Moreover, it obtains superior clustering accuracy on most of the image datasets compared to the state-of-the-art flat clustering models.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

michalznalezniak/contrastive-hierarchical-clustering officialmentioned in papermentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

ClusteringDeep ClusteringImage ClusteringSelf-Supervised Learning

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Clustering CIFAR-10 CoHiClust ARI 0.731 #24 of 40 Archive leaderboard report
Image Clustering CIFAR-10 CoHiClust Accuracy 0.839 #24 of 40 Archive leaderboard report
Image Clustering CIFAR-10 CoHiClust Backbone ResNet-50 #24 of 40 Archive leaderboard report
Image Clustering CIFAR-10 CoHiClust NMI 0.779 #24 of 40 Archive leaderboard report
Image Clustering CIFAR-10 CoHiClust Train set Train #24 of 40 Archive leaderboard report
Image Clustering CIFAR-100 CoHiClust ARI 0.299 #19 of 30 Archive leaderboard report
Image Clustering CIFAR-100 CoHiClust Accuracy 0.437 #19 of 30 Archive leaderboard report
Image Clustering CIFAR-100 CoHiClust NMI 0.467 #19 of 30 Archive leaderboard report
Image Clustering Fashion-MNIST CoHiClust Accuracy 0.65 #5 of 13 Archive leaderboard report
Image Clustering ImageNet-10 CoHiClust ARI 0.899 #6 of 18 Archive leaderboard report
Image Clustering ImageNet-10 CoHiClust Accuracy 0.953 #6 of 18 Archive leaderboard report
Image Clustering ImageNet-10 CoHiClust Backbone ResNet-50 #6 of 18 Archive leaderboard report
Image Clustering ImageNet-10 CoHiClust NMI 0.907 #6 of 18 Archive leaderboard report
Image Clustering Imagenet-dog-15 CoHiClust ARI 0.232 #15 of 20 Archive leaderboard report
Image Clustering Imagenet-dog-15 CoHiClust Accuracy 0.355 #15 of 20 Archive leaderboard report
Image Clustering Imagenet-dog-15 CoHiClust Backbone ResNet-50 #15 of 20 Archive leaderboard report
Image Clustering Imagenet-dog-15 CoHiClust NMI 0.411 #15 of 20 Archive leaderboard report
Image Clustering MNIST CoHiClust Accuracy 0.99 #2 of 3 Archive leaderboard report
Image Clustering STL-10 CoHiClust ARI 0.474 #22 of 29 Archive leaderboard report
Image Clustering STL-10 CoHiClust Accuracy 0.613 #22 of 29 Archive leaderboard report
Image Clustering STL-10 CoHiClust Backbone ResNet-50 #22 of 29 Archive leaderboard report
Image Clustering STL-10 CoHiClust NMI 0.584 #22 of 29 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

1x1 ConvolutionAverage PoolingBASEBatch NormalizationBottleneck Residual BlockConvolutionGlobal Average PoolingKaiming InitializationMax PoolingReLUResidual BlockResidual Connection

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections