Discrete Tensorized Label Learning with Anchor Graphs
Keywords: multi-view clustering, tensorized label learning, Anchor graph, nuclear norm
Abstract: Many discrete multi-view clustering methods based on anchor graphs use the anchor graph decomposition or spectral clustering to obtain the final clustering labels, such methods achieve good results but lack interpretability. Morever, some of them are poorly balanced. To this end, first, we start from the perspective of label transmission to convert labels of the anchors to the labels of the samples, which has better interpretability. Second, we find a new and remarkable use of the nuclear norm, i.e., maximizing the nuclear norm can ensure the balanced clusters, which has the rigorous theoretical proof. Simultaneously, a novel optimisation method based on the first order Taylor expansion is proposed for the nuclear norm. Finally, we introduce the tensor Schatten $p$-norm to fully exploit the spatial structural and complementary information between views, which can obtain aligned label matrices. Extensive experiments have verified the superiority of the proposed method compared with other state-of-the-art methods.
Primary Area: unsupervised, self-supervised, semi-supervised, and supervised representation learning
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 130
Loading