Pith. sign in

REVIEW 4 major objections 5 minor 67 references

Intra-view and Inter-view Correlation Guided Multi-view Novel Class Discovery

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper's central claim is that novel class discovery can be extended to multi-view data by transferring view weights learned from known classes to novel classes, and that this gives better clustering than existing multi-view clustering…

desk verdict First multi-view NCD formulation with a coherent optimization, but the headline accuracy claims are undercut by test-set hyperparameter tuning and an overstated view-weight transfer story. read the letter →

arxiv 2507.12029 v1 pith:PK5ILU6M submitted 2025-07-16 cs.CV cs.AI

Xinhang Wan , Jiyuan Liu , Qian Qu , Suyuan Liu , Chuyu Zhang , Fangdi Wang , Xinwang Liu , En Zhu
show 1 more author
Kunlun He
This is my paper · ORCID
classification cs.CVcs.AI
keywords novelclassdiscoverymulti-viewclusteringmatrixfactorizationviewweightspseudo-label-freemulti-omicslabeltransfer
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Novel class discovery asks an algorithm to cluster unlabeled data whose categories are disjoint from a small labeled set, and until now this task has been studied almost exclusively on single-view data. This paper argues that the natural next step is multi-view novel class discovery, where each sample has several complementary views, and it proposes IICMVNCD as the first method aimed at this setting. The central claim is that knowledge from the labeled classes can be transferred in two coordinated ways: within each view, a shared matrix-factorization basis captures the distributional overlap between known and novel classes, and across views, the importance weights learned from known-class supervision guide how the novel classes are fused and labeled. If this transfer works, multi-view NCD can be solved without pseudo-labels, which the paper identifies as a major source of instability in existing NCD methods. Experiments across eight multi-view datasets, including multi-omics data, support that claim with gains in clustering accuracy, NMI, and purity.

What carries the argument

The carrying object is a per-view matrix factorization $X_v = W_v A_v Y$ with an orthonormal shared basis $W_v \in \mathbb{R}^{d_v \times k}$, a centroid matrix $A_v$, and a one-hot predicted label matrix $Y = [Y_l, Y_u]$, together with a view-weight vector $\alpha$ constrained to the simplex. The shared basis is what transfers distributional knowledge between known and novel classes; the weighted fusion $\sum_v \alpha_v \|X_v - W_v A_v Y\|_F^2$ is how inter-view correlation is expressed; and the supervision term $\lambda_1 \|Y_l - G_l\|_F^2$ is what learns $\alpha$ from labeled data before it is handed to the novel partition. Updates alternate in closed form: $W_v$ by SVD, $A_v$ by a linear solve, $Y$ by a discrete per-sample assignment, and $\alpha$ by a Cauchy-Schwarz weighted inverse-residual formula.

What would settle it

Construct a two-view dataset where view 1 cleanly separates the known classes but view 2 alone separates the novel classes, then run IICMVNCD: if view 1 keeps its high weight on the novel partition and accuracy drops below an equal-weight or oracle-weight baseline, the transfer assumption is refuted.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that novel classes in multi-view data can be clustered by borrowing two kinds of structure from known classes. At the intra-view level, features of labeled and unlabeled samples are concatenated and decomposed into a view-specific orthonormal basis matrix shared by both sets and a factor matrix that records pairwise sample relationships; the shared basis is what carries distributional consistency between known and novel classes. At the inter-view level, the factor matrices are fused into one-hot predicted label matrices with per-view weights, and those weights are adjusted so that the predicted labels of known classes match their ground truth. The learned weight vector is then applied directly to the novel-class reconstruction and label assignment, and a repulsion term keeps novel labels disjoint from known labels. The paper's headline empirical claim is that this scheme beats both multi-view clustering baselines and adapted single-view NCD methods on all eight benchmark datasets.

Load-bearing premise

The whole transfer rests on the premise that the relative reliability of views is the same for known and novel classes; if the views inform the two sets differently, the weights learned from labeled data will misdirect the unlabeled clustering.

Editorial extensions

If this is right

  • Multi-view novel class discovery becomes a real setting rather than an implicit single-view assumption, so applications such as multi-omics disease discovery can treat each omics layer as a view.
  • Pseudo-label supervision can be dropped from the multi-view pipeline, removing a known source of instability in existing NCD methods.
  • The learned view weights are interpretable: they say which data view matters most for separating known classes, and the paper assumes that ranking carries over to novel classes.
  • Because each iteration is linear in the number of samples and convergence is guaranteed, the method is scalable to large datasets like the 38,654-sample video dataset used in evaluation.
  • On the eight reported benchmarks, the method achieves the best ACC, NMI, and Purity against both multi-view clustering and NCD baselines.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • An untested but testable extension is to verify view-weight transfer directly, e.g., by measuring each view's clustering accuracy on known versus novel classes and checking whether the relative ranking matches the learned $\alpha$.
  • The framework operates on raw feature matrices, so it could plausibly be composed with learned deep embeddings or graph views; the paper does not explore either.
  • The disjointness repulsion term suggests a natural stress test: datasets where known and novel classes are close in feature space, where that term would need to do most of the work.
  • Because the paper tunes $\lambda_1$ and $\lambda_2$ on labeled data, automated selection from known-class performance is a plausible practical follow-up, which the authors themselves flag.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. This paper proposes IICMVNCD, a matrix-factorization-based method for novel class discovery in multi-view data. The model decomposes each view into a shared orthonormal basis and a factor matrix over known and novel samples, generates one-hot labels via weighted fusion, and uses a known-class supervision term plus a novel-known separation penalty. An alternating optimization with closed-form updates is derived, and experiments on eight datasets compare with multi-view clustering and NCD baselines.

Significance. If the reported gains hold, the paper fills a genuine gap: NCD has been studied almost exclusively in single-view settings, and the proposed closed-form updates and linear-in-n complexity are attractive. The derivations for W_v, A_v, and Y are mostly sound, and the authors provide a convergence argument, ablation study, and parameter-sensitivity analysis. However, the evaluation protocol and an inconsistency in the alpha update currently prevent me from accepting the central empirical claim.

major comments (4)
  1. [5.1.2 and 5.5] The hyperparameters lambda1 and lambda2 are tuned from a grid (10^[0,1,...,5]) with no validation split described, and Section 5.5 states that tuning them on known classes is future work. Because the novel-class labels are the test labels in NCD, selecting the hyperparameters on the test set means the reported ACC/NMI/Purity in Table 2 may be optimistically biased and the 'consistently outperforms' claim is not verifiable. Please provide a validation protocol, such as a held-out split of known classes or a known-class proxy, and report results under that protocol.
  2. [3.4 and 4.4] There is a mismatch between the objective in Eq. (3), which contains alpha_v/2 times the reconstruction error, and the alpha subproblem in Eq. (13), which minimizes sum_v alpha_v^2 r_v^2. The closed-form update in Eq. (15) solves the quadratic surrogate, not the stated objective, so the monotone-decrease argument in Eq. (17) does not follow for the original J. Please align the objective and the subproblem, or justify the quadratic surrogate as an intentional modification.
  3. [3.3 and 4.4] The paper repeatedly states that view weights are learned from known classes and then transferred to novel classes, but the update in Eq. (13) computes r_v^2 = ||X_v - W_v A_v Y||_F^2 over the full X_v = [X_l^v, X_u^v], including novel samples. Thus alpha is influenced directly by novel-class reconstruction errors, and the actual transfer happens through the shared basis and label matrix rather than through the weights themselves. Please correct the narrative and, ideally, test the transfer assumption by comparing alpha learned on known classes only versus the current global update.
  4. [5.4, Table 3] The ablation claim that 'removing any one of the modules results in a decline' is not supported by the table: on STL10 the 'w/o alpha' variant already reaches 99.02 ACC, identical to the full method, and on Wiki and YTB10 the differences are very small (64.35 vs. 65.42 and 94.32 vs. 94.55). Moreover, on uci-digit 'w/o alpha' (95.20) is nearly as good as the full model (95.30). Please temper the claim or provide repeated-trial results with error bars.
minor comments (5)
  1. [4.5.1, Eq. (18)] The bound sum_{g_i in G_l} sum_{y_j in Y_u} ||g_i - y_j||_F^2 <= n_l n_u sqrt(2) is incorrect; each squared distance equals 2, so the sum is 2 n_l n_u. The lower-bound constant in Eq. (19) should be updated accordingly.
  2. [1] The acronym is written as 'IICMVNVD' in the introduction, but 'IICMVNCD' is used elsewhere.
  3. [5.2] The sentence 'These findings highlight the importance of developing NCD techniques that can fully leverage multi-view information...' is duplicated verbatim at the end of the section.
  4. [5.1.2] The adaptation of NCD baselines by replacing their backbone with separate MLPs and concatenating features may disadvantage methods designed for image augmentations; please specify whether any image-specific preprocessing was used and discuss this limitation.
  5. [5.5] The notation '10.^[0, 1, ..., 5]' and the axis labels in Figure 3 are garbled; please clarify the grid and the plotted metric.

Circularity Check

2 steps flagged · score 6.0 of 10

Effectiveness claim is partly fitted: lambda1/lambda2 are tuned on novel-class test labels (admitted by deferring known-class tuning to future work), and the view weights said to transfer from known classes are actually optimized with reconstruction errors over novel data.

  1. fitted input called prediction [Section 5.5 (Parameter Sensitivity), with Section 5.1.2]
    "Moreover, in practical applications, due to the presence of known class label information, we can also adjust the hyperparameters for unknown classes based on the model's performance on known classes. We will explore how to tune them based on the results of known classes in future research."

    lambda_1 and lambda_2 are the trade-off coefficients in Eq. (3), and Section 5.1.2 states they are tuned from 10^[0,1,...,5] without any validation split or known-class-based criterion. Section 5.5 defers known-class-based tuning to future work, so the grid must have been scored on the novel-class labels, which are the test labels. Table 2's ACC/NMI/Purity values are therefore selected on the target, and the claimed 'consistently outperforms' result is partly a fitted value rather than an independent prediction.

  2. fitted input called prediction [Sec. 3.1 and Sec. 4.4, Eqs. (13)-(14), with Eq. (1)]
    "The learned view weights are then applied to the novel classes, ensuring a more effective partitioning of the unlabeled dataset (see Sec. 3.3). ... r2 v = ||Xv - WvAvY||2 F, where Xv = [Xl v, Xu v] in Rdv*n is the feature matrix concatenated of the two sets in v-th view."

    The abstract and Sec. 3.3 claim alpha is learned from known classes and transferred to novel classes, but the only place alpha is updated (Eq. 13) minimizes sum_v alpha_v^2 * ||X_v - W_v A_v Y||_F^2, and X_v contains the novel columns Xu_v (Eq. 1). Hence alpha is fitted on the reconstruction error of the very novel samples whose labels Yu are the target, so the 'transfer' is a joint fit on the target, not a known-class-only learned weight. The predicted partition is thus partially determined by weights already tuned to that partition's reconstruction.

full rationale

The core matrix-factorization and label-generation machinery is self-contained: Eq. (1) and Eq. (3) define a joint objective, and the alternating updates in Section 4 are derived from that objective rather than imported from a self-citation. There is no load-bearing uniqueness theorem and no known result merely renamed. However, the paper's central empirical claim is weakened by two circular elements. First, the hyperparameters lambda1 and lambda2 are tuned on the novel-class test labels; Section 5.5 explicitly says known-class-based tuning is future work, which is an admission that the grid search in Section 5.1.2 was evaluated on the target labels. This makes the reported Table 2 improvements partly a fitted value rather than a prediction. Second, the claimed known-to-novel transfer of view weights is not what the equations implement: Eq. (13) updates alpha from reconstruction residuals over the full concatenated matrix X_v = [X_l_v, X_u_v], so novel-class data directly influence the weights that are then used to cluster the same novel-class data. This is a fitted-input-called-prediction pattern rather than a derivation identity. The view-weight transfer assumption is also unvalidated, but that is an empirical risk, not circularity. Overall, the algorithm itself is not circular by construction, but the reported effectiveness is partially fitted, so the score is 6.

Assumptions & free parameters 2 free parameters · 2 assumptions · 0 invented entities

The central method rests on the NCD distributional-similarity assumption and the transferability of view weights; both are domain assumptions not verified by the experiments. Two hyperparameters are hand-tuned without a validation split.

free parameters (2)
  • lambda1 = tuned from 10^[0,...,5]
    Weight of the known-class supervision loss in Eq. (3); chosen by grid search, likely on the test set.
  • lambda2 = tuned from 10^[0,...,5]
    Weight of the known-novel label repulsion term in Eq. (3); chosen by grid search, likely on the test set.
assumptions (2)
  • domain assumption The known and novel classes share similar data distributions, so a shared low-dimensional subspace (rank k = kl + ku) exists per view.
    Invoked in Section 3.2, Eq. (1), where the same basis matrix Wv is factorized from the concatenation of labeled and unlabeled features.
  • domain assumption View importance learned on known classes transfers to novel classes.
    Invoked in Section 3.3 and Eq. (2), where view weights alpha are optimized using the known-class supervision loss and then applied to the novel-class factorization.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Intra-view and Inter-view Correlation Guided Multi-view Novel Class Discovery." pith.science (2026). https://pith.science/paper/PK5ILU6M

@misc{pith2026250712029,
  author       = {Pith},
  title        = {Pith review of: Intra-view and Inter-view Correlation Guided Multi-view Novel Class Discovery},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PK5ILU6M}},
  note         = {Machine review of arXiv:2507.12029}
}
read the original abstract

In this paper, we address the problem of novel class discovery (NCD), which aims to cluster novel classes by leveraging knowledge from disjoint known classes. While recent advances have made significant progress in this area, existing NCD methods face two major limitations. First, they primarily focus on single-view data (e.g., images), overlooking the increasingly common multi-view data, such as multi-omics datasets used in disease diagnosis. Second, their reliance on pseudo-labels to supervise novel class clustering often results in unstable performance, as pseudo-label quality is highly sensitive to factors such as data noise and feature dimensionality. To address these challenges, we propose a novel framework named Intra-view and Inter-view Correlation Guided Multi-view Novel Class Discovery (IICMVNCD), which is the first attempt to explore NCD in multi-view setting so far. Specifically, at the intra-view level, leveraging the distributional similarity between known and novel classes, we employ matrix factorization to decompose features into view-specific shared base matrices and factor matrices. The base matrices capture distributional consistency among the two datasets, while the factor matrices model pairwise relationships between samples. At the inter-view level, we utilize view relationships among known classes to guide the clustering of novel classes. This includes generating predicted labels through the weighted fusion of factor matrices and dynamically adjusting view weights of known classes based on the supervision loss, which are then transferred to novel class learning. Experimental results validate the effectiveness of our proposed approach.

Figures

Figures reproduced from arXiv: 2507.12029 by the authors.

Figure 1
Figure 1. Framework of the proposed algorithm: We first apply matrix factorization to decompose the features of the two sets into set-shared [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. The objective values of our method vary with iterations [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

67 extracted references · 67 canonical work pages

  1. [1]

    Deep adaptive image cluster- ing

    Jianlong Chang, Lingfeng Wang, Gaofeng Meng, Shiming Xiang, and Chunhong Pan. Deep adaptive image cluster- ing. In Proceedings of the IEEE international conference on computer vision, pages 5879–5887, 2017. 1

    work page 2017

  2. [2]

    M. S. Chen, L. Huang, C. D. Wang, and D. Huang. Multi- view clustering in latent embedding space. InThe 34th AAAI Conference on Artificial Intelligence, 2020. 2

    work page 2020

  3. [3]

    Nearest neighbor matching for deep clustering

    Zhiyuan Dang, Cheng Deng, Xu Yang, Kun Wei, and Heng Huang. Nearest neighbor matching for deep clustering. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 13693–13702, 2021. 1

    work page 2021

  4. [4]

    Cross-view topology based consistent and complemen- tary information for deep multi-view clustering

    Zhibin Dong, Siwei Wang, Jiaqi Jin, Xinwang Liu, and En Zhu. Cross-view topology based consistent and complemen- tary information for deep multi-view clustering. In Proceed- ings of the IEEE/CVF International Conference on Com- puter Vision, pages 19440–19451, 2023. 2

    work page 2023

  5. [5]

    Enhanced then progressive fusion with view graph for multi-view clus- tering

    Zhibin Dong, Meng Liu, Siwei Wang, Ke Liang, Yi Zhang, Suyuan Liu, Jiaqi Jin, Xinwang Liu, and En Zhu. Enhanced then progressive fusion with view graph for multi-view clus- tering. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 15518–15527, 2025. 2

    work page 2025

  6. [6]

    A unified objective for novel class discovery

    Enrico Fini, Enver Sangineto, St ´ephane Lathuili `ere, Zhun Zhong, Moin Nabi, and Elisa Ricci. A unified objective for novel class discovery. InProceedings of the IEEE/CVF inter- national conference on computer vision , pages 9284–9292,

    work page

  7. [7]

    Multi-view low-rank matrix factorization using multi- ple manifold regularization

    Shengxiang Gao, Zhengtao Yu, Taisong Jin, and Ming Yin. Multi-view low-rank matrix factorization using multi- ple manifold regularization. Neurocomputing, 335:143–152,

    work page

  8. [8]

    Learning accurate template match- ing with differentiable coarse-to-fine correspondence refine- ment

    Zhirui Gao, Renjiao Yi, Zheng Qin, Yunfan Ye, Chenyang Zhu, and Kai Xu. Learning accurate template match- ing with differentiable coarse-to-fine correspondence refine- ment. Computational Visual Media , 10(2):309–330, 2024. 1

    work page 2024

Show all 67 references
  1. [9]

    Generic objects as pose probes for few- shot view synthesis

    Zhirui Gao, Renjiao Yi, Chenyang Zhu, Ke Zhuang, Wei Chen, and Kai Xu. Generic objects as pose probes for few- shot view synthesis. IEEE Transactions on Circuits and Sys- tems for Video Technology, 2025. 1

    2025

  2. [10]

    Class- relation knowledge distillation for novel class discovery

    Peiyan Gu, Chuyu Zhang, Ruijie Xu, and Xuming He. Class- relation knowledge distillation for novel class discovery. In 2023 IEEE/CVF International Conference on Computer Vi- sion (ICCV), pages 16428–16437, 2023. 1, 6

    2023

  3. [11]

    Learning to discover novel visual categories via deep transfer cluster- ing

    Kai Han, Andrea Vedaldi, and Andrew Zisserman. Learning to discover novel visual categories via deep transfer cluster- ing. In Proceedings of the IEEE/CVF International Confer- ence on Computer Vision (ICCV), 2019. 2

    2019

  4. [12]

    Automatically discov- ering and learning new visual categories with ranking statis- tics

    Kai Han, Sylvestre-Alvise Rebuffi, Sebastien Ehrhardt, An- drea Vedaldi, and Andrew Zisserman. Automatically discov- ering and learning new visual categories with ranking statis- tics. In International Conference on Learning Representa- tions, 2020. 2

    2020

  5. [13]

    Nc2d: Novel class dis- covery for node classification

    Yue Hou, Xueyuan Chen, He Zhu, Ruomei Liu, Bowen Shi, Jiaheng Liu, Junran Wu, and Ke Xu. Nc2d: Novel class dis- covery for node classification. In Proceedings of the 33rd ACM International Conference on Information and Knowl- edge Management, pages 849–859, 2024. 1

    2024

  6. [14]

    A survey of evolutionary algorithms for clus- tering

    Eduardo Raul Hruschka, Ricardo JGB Campello, Alex A Freitas, et al. A survey of evolutionary algorithms for clus- tering. IEEE Transactions on systems, man, and cybernetics, Part C (applications and reviews), 39(2):133–155, 2009. 1

    2009

  7. [15]

    Learning to cluster in order to transfer across domains and tasks

    Yen-Chang Hsu, Zhaoyang Lv, and Zsolt Kira. Learning to cluster in order to transfer across domains and tasks. In In- ternational Conference on Learning Representations , 2018. 2

    2018

  8. [16]

    Multi-class classification without multi-class labels

    Yen-Chang Hsu, Zhaoyang Lv, Joel Schlosser, Phillip Odom, and Zsolt Kira. Multi-class classification without multi-class labels. In International Conference on Learning Represen- tations, 2019. 2

    2019

  9. [17]

    Data clustering: 50 years beyond k-means

    Anil K Jain. Data clustering: 50 years beyond k-means. Pat- tern recognition letters, 31(8):651–666, 2010. 1

    2010

  10. [18]

    Deep incomplete multi-view clustering with cross-view partial sample and prototype alignment

    Jiaqi Jin, Siwei Wang, Zhibin Dong, Xinwang Liu, and En Zhu. Deep incomplete multi-view clustering with cross-view partial sample and prototype alignment. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 11600–11609, 2023. 2

    2023

  11. [19]

    Novel class discovery without forgetting

    KJ Joseph, Sujoy Paul, Gaurav Aggarwal, Soma Biswas, Piyush Rai, Kai Han, and Vineeth N Balasubramanian. Novel class discovery without forgetting. In European Con- ference on Computer Vision, pages 570–586. Springer, 2022. 1

    2022

  12. [20]

    Large-scale multi-view sub- space clustering in linear time

    Zhao Kang, Wangtao Zhou, Zhitong Zhao, Junming Shao, Meng Han, and Zenglin Xu. Large-scale multi-view sub- space clustering in linear time. In AAAI Conference on Arti- ficial Intelligence, 2019. 6

    2019

  13. [21]

    Multi-view bipartite graph clustering with coupled noisy feature filter

    Liang Li, Junpu Zhang, Siwei Wang, Xinwang Liu, Kenli Li, and Keqin Li. Multi-view bipartite graph clustering with coupled noisy feature filter. IEEE Transactions on Knowl- edge and Data Engineering, 35(12):12842–12854, 2023. 2

    2023

  14. [22]

    Tsang, and Keqin Li

    Liang Li, Yuangang Pan, Jie Liu, Yue Liu, Xinwang Liu, Kenli Li, Ivor W. Tsang, and Keqin Li. Bgae: Auto-encoding multi-view bipartite graph clustering. IEEE Transactions on Knowledge and Data Engineering, 36(8):1–14, 2024. 2

    2024

  15. [23]

    Late fusion multiview clustering via min-max optimization

    Miaomiao Li, Xinwang Liu, Yi Zhang, and Weixuan Liang. Late fusion multiview clustering via min-max optimization. IEEE Transactions on Neural Networks and Learning Sys- tems, pages 1–11, 2023. 2

    2023

  16. [24]

    Modeling inter-class and intra-class constraints in novel class discov- ery

    Wenbin Li, Zhichen Fan, Jing Huo, and Yang Gao. Modeling inter-class and intra-class constraints in novel class discov- ery. In 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 3449–3458, 2023. 1, 6

    2023

  17. [25]

    Multi-view clustering via joint nonnegative matrix factorization

    Jialu Liu, Chi Wang, Jing Gao, and Jiawei Han. Multi-view clustering via joint nonnegative matrix factorization. In Pro- ceedings of the 2013 SIAM international conference on data mining, pages 252–260. SIAM, 2013. 3

    2013

  18. [26]

    Multiview subspace cluster- ing via co-training robust data representation

    Jiyuan Liu, Xinwang Liu, Yuexiang Yang, Xifeng Guo, Mar- ius Kloft, and Liangzhong He. Multiview subspace cluster- ing via co-training robust data representation. IEEE Trans- actions on Neural Networks and Learning Systems , 33(10): 5177–5189, 2021. 1

    2021

  19. [27]

    One-pass multi- view clustering for large-scale data

    Jiyuan Liu, Xinwang Liu, Yuexiang Yang, Li Liu, Siqi Wang, Weixuan Liang, and Jiangyong Shi. One-pass multi- view clustering for large-scale data. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 12344–12353, 2021. 6

    2021

  20. [28]

    Contrastive multi-view kernel learning.IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(8):9552–9566, 2023

    Jiyuan Liu, Xinwang Liu, Yuexiang Yang, Qing Liao, and Yuanqing Xia. Contrastive multi-view kernel learning.IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(8):9552–9566, 2023. 2

    2023

  21. [29]

    Large-scale multi-view tensor cluster- ing with implicit linear kernels

    Jiyuan Liu, Xinwang Liu, Chuankun Li, Xinhang Wan, Hao Tan, Yi Zhang, Weixuan Liang, Qian Qu, Yu Feng, Renxiang Guan, and Ke Liang. Large-scale multi-view tensor cluster- ing with implicit linear kernels. In Proceedings of the Com- puter Vision and Pattern Recognition Conferen...

    2025

  22. [30]

    Learn from view correlation: An anchor enhancement strategy for multi- view clustering

    Suyuan Liu, Ke Liang, Zhibin Dong, Siwei Wang, Xihong Yang, Sihang Zhou, En Zhu, and Xinwang Liu. Learn from view correlation: An anchor enhancement strategy for multi- view clustering. In 2024 IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), pages 2615...

    2024

  23. [31]

    Robust and consistent anchor graph learning for multi- view clustering

    Suyuan Liu, Qing Liao, Siwei Wang, Xinwang Liu, and En Zhu. Robust and consistent anchor graph learning for multi- view clustering. IEEE Transactions on Knowledge and Data Engineering, 36(8):4207–4219, 2024. 6

    2024

  24. [32]

    Alleviate anchor-shift: Explore blind spots with cross- view reconstruction for incomplete multi-view clustering

    Suyuan Liu, Siwei Wang, Ke Liang, Junpu Zhang, Zhibin Dong, Tianrui Liu, En Zhu, Xinwang Liu, and Kunlun He. Alleviate anchor-shift: Explore blind spots with cross- view reconstruction for incomplete multi-view clustering. Advances in Neural Information Processing Systems , 37...

    2024

  25. [33]

    Optimal neighborhood kernel clustering with multiple kernels

    Xinwang Liu, Sihang Zhou, Yueqing Wang, Miaomiao Li, Yong Dou, En Zhu, and Jianping Yin. Optimal neighborhood kernel clustering with multiple kernels. InProceedings of the Thirty-First AAAI Conference on Artificial Intelligence, page 2266–2272. AAAI Press, 2017. 2

    2017

  26. [34]

    Deep graph clustering via dual correlation reduction

    Yue Liu, Wenxuan Tu, Sihang Zhou, Xinwang Liu, Linxuan Song, Xihong Yang, and En Zhu. Deep graph clustering via dual correlation reduction. In Proceedings of the AAAI con- ference on artificial intelligence, pages 7603–7611, 2022. 1

    2022

  27. [35]

    Simple con- trastive graph clustering

    Yue Liu, Xihong Yang, Sihang Zhou, Xinwang Liu, Siwei Wang, Ke Liang, Wenxuan Tu, and Liang Li. Simple con- trastive graph clustering. IEEE Transactions on Neural Net- works and Learning Systems, 2023. 1

    2023

  28. [36]

    Convex sparse spectral clustering: Single-view to multi-view

    Canyi Lu, Shuicheng Yan, and Zhouchen Lin. Convex sparse spectral clustering: Single-view to multi-view. IEEE Trans- actions on Image Processing, 25(6):2833–2843, 2016. 2

    2016

  29. [37]

    Algorithms for hierar- chical clustering: an overview

    Fionn Murtagh and Pedro Contreras. Algorithms for hierar- chical clustering: an overview. Wiley Interdisciplinary Re- views: Data Mining and Knowledge Discovery , 2(1):86–97,

  30. [38]

    Clustering and projected clustering with adaptive neighbors

    Feiping Nie, Xiaoqian Wang, and Heng Huang. Clustering and projected clustering with adaptive neighbors. In Pro- ceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining , pages 977–986,

  31. [39]

    COMIC: Multi-view clustering without parameter selection

    Xi Peng, Zhenyu Huang, Jiancheng Lv, Hongyuan Zhu, and Joey Tianyi Zhou. COMIC: Multi-view clustering without parameter selection. In Proceedings of the 36th International Conference on Machine Learning, pages 5092–5101. PMLR,

  32. [40]

    Deep clustering: A comprehensive survey

    Yazhou Ren, Jingyu Pu, Zhimeng Yang, Jie Xu, Guofeng Li, Xiaorong Pu, S Yu Philip, and Lifang He. Deep clustering: A comprehensive survey. IEEE transactions on neural net- works and learning systems, 2024. 1

    2024

  33. [41]

    Class-incremental novel class discovery

    Subhankar Roy, Mingxuan Liu, Zhun Zhong, Nicu Sebe, and Elisa Ricci. Class-incremental novel class discovery. In European Conference on Computer Vision, pages 317–333. Springer, 2022. 1

    2022

  34. [42]

    Wage: Weight- sharing attribute-missing graph autoencoder

    Wenxuan Tu, Sihang Zhou, Xinwang Liu, Zhiping Cai, Yawei Zhao, Yue Liu, and Kunlun He. Wage: Weight- sharing attribute-missing graph autoencoder. IEEE Trans- actions on Pattern Analysis and Machine Intelligence, 2025. 1

    2025

  35. [43]

    Auto- weighted multi-view clustering for large-scale data

    Xinhang Wan, Xinwang Liu, Jiyuan Liu, Siwei Wang, Yi Wen, Weixuan Liang, En Zhu, Zhe Liu, and Lu Zhou. Auto- weighted multi-view clustering for large-scale data. In Pro- ceedings of the Thirty-Seventh AAAI Conference on Artifi- cial Intelligence and Thirty-Fifth Conference on ...

    2023

  36. [44]

    One-step multi- view clustering with diverse representation

    Xinhang Wan, Jiyuan Liu, Xinbiao Gan, Xinwang Liu, Siwei Wang, Yi Wen, Tianjiao Wan, and En Zhu. One-step multi- view clustering with diverse representation. IEEE Transac- tions on Neural Networks and Learning Systems , pages 1– 13, 2024. 1

    2024

  37. [45]

    Decouple then classify: A dynamic multi-view labeling strategy with shared and specific information

    Xinhang Wan, Jiyuan Liu, Xinwang Liu, Yi Wen, Hao Yu, Siwei Wang, Shengju Yu, Tianjiao Wan, Jun Wang, and En Zhu. Decouple then classify: A dynamic multi-view labeling strategy with shared and specific information. In Forty-first International Conference on Machine Learning, 2024

    2024

  38. [46]

    Contrastive continual multiview clustering with filtered structural fusion

    Xinhang Wan, Jiyuan Liu, Hao Yu, Qian Qu, Ao Li, Xin- wang Liu, Ke Liang, Zhibin Dong, and En Zhu. Contrastive continual multiview clustering with filtered structural fusion. IEEE Transactions on Neural Networks and Learning Sys- tems, 2024

    2024

  39. [47]

    Fast continual multi-view clustering with incomplete views

    Xinhang Wan, Bin Xiao, Xinwang Liu, Jiyuan Liu, Weixuan Liang, and En Zhu. Fast continual multi-view clustering with incomplete views. IEEE Transactions on Image Processing, 33:2995–3008, 2024. 1

    2024

  40. [48]

    Diverse non-negative matrix fac- torization for multiview data representation

    Jing Wang, Feng Tian, Hongchuan Yu, Chang Hong Liu, Kun Zhan, and Xiao Wang. Diverse non-negative matrix fac- torization for multiview data representation. IEEE Transac- tions on Cybernetics, 48(9):2620–2632, 2018. 3

    2018

  41. [49]

    Efficient and effective one-step multi- view clustering

    Jun Wang, Chang Tang, Zhiguo Wan, Wei Zhang, Kun Sun, and Albert Y Zomaya. Efficient and effective one-step multi- view clustering. IEEE Transactions on Neural Networks and Learning Systems, 2023. 2

    2023

  42. [50]

    Multiple kernel clustering with adaptive multi-scale partition selection

    Jun Wang, Zhenglai Li, Chang Tang, Suyuan Liu, Xinhang Wan, and Xinwang Liu. Multiple kernel clustering with adaptive multi-scale partition selection. IEEE Transactions on Knowledge and Data Engineering, 2024. 2

    2024

  43. [51]

    Multi-view clustering via late fusion alignment maximization

    Siwei Wang, Xinwang Liu, En Zhu, Chang Tang, Jiyuan Liu, Jingtao Hu, Jingyuan Xia, and Jianping Yin. Multi-view clustering via late fusion alignment maximization. In Pro- ceedings of the Twenty-Eighth International Joint Confer- ence on Artificial Intelligence, IJCAI-19 , page...

    2019

  44. [52]

    Fast parameter-free multi- view subspace clustering with consensus anchor guidance

    Siwei Wang, Xinwang Liu, Xinzhong Zhu, Pei Zhang, Yi Zhang, Feng Gao, and En Zhu. Fast parameter-free multi- view subspace clustering with consensus anchor guidance. IEEE Transactions on Image Processing, 31:556–568, 2022. 6

    2022

  45. [53]

    Unpaired multi-view graph clustering with cross-view structure matching

    Yi Wen, Siwei Wang, Qing Liao, Weixuan Liang, Ke Liang, Xinhang Wan, and Xinwang Liu. Unpaired multi-view graph clustering with cross-view structure matching. IEEE Trans- actions on Neural Networks and Learning Systems, 2023. 2

    2023

  46. [54]

    Measure domain’s gap: A similar domain se- lection principle for multi-domain recommendation

    Yi Wen, Yue Liu, Derong Xu, Huishi Luo, Pengyue Jia, Yiqing Wu, Siwei Wang, Ke Liang, Maolin Wang, Yiqi Wang, et al. Measure domain’s gap: A similar domain se- lection principle for multi-domain recommendation. arXiv preprint arXiv:2505.20227, 2025. 1, 2

    2025 arXiv

  47. [55]

    Fast multi-view clustering via non- negative and orthogonal factorization

    Ben Yang, Xuetao Zhang, Feiping Nie, Fei Wang, Weizhong Yu, and Rong Wang. Fast multi-view clustering via non- negative and orthogonal factorization. IEEE Transactions on Image Processing, 30:2575–2586, 2021. 2

    2021

  48. [56]

    Multiple kernel k-means clustering by selecting representa- tive kernels

    Yaqiang Yao, Yang Li, Bingbing Jiang, and Huanhuan Chen. Multiple kernel k-means clustering by selecting representa- tive kernels. IEEE Transactions on Neural Networks and Learning Systems, 32(11):4983–4996, 2021. 2

    2021

  49. [57]

    Gzoo: Black-box node injec- tion attack on graph neural networks via zeroth-order opti- mization

    Hao Yu, Ke Liang, Dayu Hu, Wenxuan Tu, Chuan Ma, Si- hang Zhou, and Xinwang Liu. Gzoo: Black-box node injec- tion attack on graph neural networks via zeroth-order opti- mization. IEEE Transactions on Knowledge and Data Engi- neering, 2024. 1

    2024

  50. [58]

    On the adversarial robustness of multi-kernel clustering

    Hao Yu, Weixuan Liang, KE LIANG, Suyuan Liu, Meng Liu, and Xinwang Liu. On the adversarial robustness of multi-kernel clustering. In ICML, 2025. 2

    2025

  51. [59]

    Towards resource-friendly, extensible and stable incom- plete multi-view clustering

    Shengju Yu, Zhibing Dong, Siwei Wang, Xinhang Wan, Yue Liu, Weixuan Liang, Pei Zhang, Wenxuan Tu, and Xinwang Liu. Towards resource-friendly, extensible and stable incom- plete multi-view clustering. In Proceedings of the 41st In- ternational Conference on Machine Learning, pa...

    2024

  52. [60]

    How to construct corre- sponding anchors for incomplete multiview clustering.IEEE Transactions on Circuits and Systems for Video Technology, 34(4):2845–2860, 2024

    Shengju Yu, Siwei Wang, Yi Wen, Ziming Wang, Zhigang Luo, En Zhu, and Xinwang Liu. How to construct corre- sponding anchors for incomplete multiview clustering.IEEE Transactions on Circuits and Systems for Video Technology, 34(4):2845–2860, 2024. 2

    2024

  53. [61]

    Sparse low-rank multi-view subspace clustering with consensus anchors and unified bi- partite graph

    Shengju Yu, Suyuan Liu, Siwei Wang, Chang Tang, Zhigang Luo, Xinwang Liu, and En Zhu. Sparse low-rank multi-view subspace clustering with consensus anchors and unified bi- partite graph. IEEE Transactions on Neural Networks and Learning Systems, 36(1):1438–1452, 2025. 2

    2025

  54. [62]

    Multi-view clustering with incremental in- stances and views

    Chao Zhang, Zhi Wang, Xiuyi Jia, Zechao Li, Chunlin Chen, and Huaxiong Li. Multi-view clustering with incremental in- stances and views. IEEE Transactions on Image Processing,

  55. [63]

    One-stage incomplete multi-view clustering via late fusion

    Yi Zhang, Xinwang Liu, Siwei Wang, Jiyuan Liu, Sisi Dai, and En Zhu. One-stage incomplete multi-view clustering via late fusion. In Proceedings of the 29th ACM International Conference on Multimedia, page 2717–2725, New York, NY , USA, 2021. Association for Computing Machinery. 3

    2021

  56. [64]

    Incomplete multi-view clustering via diffusion contrastive generation

    Yuanyang Zhang, Yijie Lin, Weiqing Yan, Li Yao, Xinhang Wan, Guangyuan Li, Chao Zhang, Guanzhou Ke, and Jie Xu. Incomplete multi-view clustering via diffusion contrastive generation. Proceedings of the AAAI Conference on Artifi- cial Intelligence, 39(21):22650–22658, 2025. 1

    2025

  57. [65]

    Binary multi-view clustering

    Zheng Zhang, Li Liu, Fumin Shen, Heng Tao Shen, and Ling Shao. Binary multi-view clustering. IEEE Transac- tions on Pattern Analysis and Machine Intelligence , 41(7): 1774–1782, 2019. 2

    2019

  58. [66]

    Novel class discovery in semantic segmentation

    Yuyang Zhao, Zhun Zhong, Nicu Sebe, and Gim Hee Lee. Novel class discovery in semantic segmentation. InProceed- ings of the IEEE/CVF conference on computer vision and pattern recognition, pages 4340–4349, 2022. 1

    2022

  59. [67]

    Neighborhood contrastive learning for novel class discovery

    Zhun Zhong, Enrico Fini, Subhankar Roy, Zhiming Luo, Elisa Ricci, and Nicu Sebe. Neighborhood contrastive learning for novel class discovery. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10867–10875, 2021. 1

    2021

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.