EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Learning Representation for Clustering Via Prototype Scattering and Positive Sampling

Zhizhong Huang, Jie Chen, Junping Zhang, Hongming Shan

IEEE Transactions on Pattern Analysis and Machine Intelligence · 01.06.2023 · 45 (6) · 7509-7524

DOI: 10.1109/tpami.2022.3216454 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Pattern Analysis and Machine Intelligence, erschienen am 01.06.2023. Verfasst von Zhizhong Huang und 3 weiteren Autor:innen. Die Arbeit wurde bisher 124-mal zitiert (Crossref) und nennt 68 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: National Natural Science Foundation of China, Shanghai Municipal Science and Technology, ZJ Lab.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Existing deep clustering methods rely on either contrastive or non-contrastive representation learning for downstream clustering task. Contrastive-based methods thanks to negative pairs learn uniform representations for clustering, in which negative pairs, however, may inevitably lead to the class collision issue and consequently compromise the clustering performance. Non-contrastive-based methods, on the other hand, avoid class collision issue, but the resulting non-uniform representations may cause the collapse of clustering. To enjoy the strengths of both worlds, this paper presents a novel end-to-end deep clustering method with prototype scattering and positive sampling, termed ProPos. Specifically, we first maximize the distance between prototypical representations, named prototype scattering loss, which improves the uniformity of representations. Second, we align one augmented view of instance with the sampled neighbors of another view-assumed to be truly positive pair in the embedding space-to improve the within-cluster compactness, termed positive sampling alignment. The strengths of ProPos are avoidable class collision issue, uniform representations, well-separated clusters, and within-cluster compactness. By optimizing ProPos in an end-to-end expectation-maximization framework, extensive experimental results demonstrate that ProPos achieves competing performance on moderate-scale clustering benchmark datasets and establishes new state-of-the-art performance on large-scale datasets. Source code is available at https://github.com/Hzzone/ProPos.

Förderung

  • National Natural Science Foundation of China
  • Shanghai Municipal Science and Technology
  • ZJ Lab
  • Shanghai Municipal of Science and Technology
  • Shanghai Sailing Program
  • Natural Science Foundation of Shanghai
  • Shanghai Center for Brain Science and Brain-inspired Technology

Zitieren

APA

Huang, Z., Chen, J., Zhang, J., & Shan, H. (2023). Learning Representation for Clustering Via Prototype Scattering and Positive Sampling. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(6), 7509-7524. https://doi.org/10.1109/tpami.2022.3216454

Vancouver

Huang Z, Chen J, Zhang J, Shan H. Learning Representation for Clustering Via Prototype Scattering and Positive Sampling. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2023;45(6):7509-7524. doi:10.1109/tpami.2022.3216454

BibTeX
@article{Huang2023,
  author = {Huang, Zhizhong and Chen, Jie and Zhang, Junping and Shan, Hongming},
  title = {Learning Representation for Clustering Via Prototype Scattering and Positive Sampling},
  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year = {2023},
  volume = {45},
  number = {6},
  pages = {7509--7524},
  doi = {10.1109/tpami.2022.3216454}
}
Kostenlos starten →