Research map / Learning theory and algorithms / Clustering
Incomplete multi-view clustering
Clusters multi-view data with missing views using semantic alignment and contrastive prototype matching.
125 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Clustering, keyword cluster “correlation clustering · hierarchical · spectral clustering”
State History Benchmarks Tried Checklist Preprints Who Export
Where it stands
Leveling off Several papers from 2025–26 still report on 1 tracked benchmark, but the best result has barely moved since 2024.
16 papers in 2023–24 and 79 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 11.4% to 38.3%.
What looks settled
- BDGP (ACC): best up to 2024 98.4, best in 2025–26 80.82 from 2 papers, at most 0.2% better, if at all.
No longer reported by papers here: MSRC-v1 (NMI, last in 2020); MSRCv1 (ACC, last in 2020); Yale (NMI, last in 2020); Yale (ACC, last in 2020).
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2016 · Discriminatively Embedded K-Means for Multi-View Clustering (CVPR 2016 · 155 citations)
- 2017 · Multi-View Clustering via Deep Matrix Factorization (AAAI 2017 · 514 citations)
- 2018 · Consistent and Specific Multi-View Subspace Clustering (AAAI 2018 · 434 citations)
- 2019 · Unified Embedding Alignment with Missing Views Inferring for Incomplete Multi-View Clustering (AAAI 2019 · 259 citations)
- 2020 · Multi-View Clustering in Latent Embedding Space (AAAI 2020 · 282 citations)
- 2021 · COMPLETER: Incomplete Multi-View Clustering via Contrastive Prediction (CVPR 2021 · 441 citations)
- 2022 · Multi-Level Feature Learning for Contrastive Multi-View Clustering (CVPR 2022 · 388 citations)
- 2023 · GCFAgg: Global and Cross-View Feature Aggregation for Multi-View Clustering (CVPR 2023 · 192 citations)
- 2024 · Incomplete Contrastive Multi-View Clustering with High-Confidence Guiding (AAAI 2024 · 109 citations)
- 2025 · Collaborative Similarity Fusion and Consistency Recovery for Incomplete Multi-view Clustering (AAAI 2025 · 16 citations)
- 2026 · Mamba-Driven Multi-View Discriminative Clustering via Global-Local Cross-View Sequence Modeling (AAAI 2026 · 5 citations)
Earliest papers here
- Discriminatively Embedded K-Means for Multi-View Clustering (CVPR 2016 · 155 citations)
- Multi-View Clustering via Deep Matrix Factorization (AAAI 2017 · 514 citations)
- Exclusivity-Consistency Regularized Multi-View Subspace Clustering (CVPR 2017 · 285 citations)
Benchmark records
The best number each year that beat every earlier one, among papers of this approach reporting their own method. Numbers are extracted by AI from each paper's text; check the paper before citing.
BDGP · ACC (higher is better)
5 papers, 2022–2025 · flat since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2022 | 98.4 | Deep Safe Incomplete Multi-view Clustering: Theorem and Algorithm (ICML 2022) |
MSRC-v1 · NMI (higher is better)
4 papers, 2016–2020 · not reported since 2020
| Year | Best so far | Set by |
|---|---|---|
| 2016 | 0.865 | Discriminatively Embedded K-Means for Multi-View Clustering (CVPR 2016 · 155 citations) |
MSRCv1 · ACC (higher is better)
4 papers, 2016–2020 · not reported since 2020
| Year | Best so far | Set by |
|---|---|---|
| 2016 | 0.924 | Discriminatively Embedded K-Means for Multi-View Clustering (CVPR 2016 · 155 citations) |
Yale · NMI (higher is better)
4 papers, 2017–2020 · not reported since 2020
| Year | Best so far | Set by |
|---|---|---|
| 2017 | 0.782 | Multi-View Clustering via Deep Matrix Factorization (AAAI 2017 · 514 citations) |
| 2018 | 0.784 | Consistent and Specific Multi-View Subspace Clustering (AAAI 2018 · 434 citations) |
Yale · ACC (higher is better)
4 papers, 2017–2020 · not reported since 2020
| Year | Best so far | Set by |
|---|---|---|
| 2017 | 0.771 | Exclusivity-Consistency Regularized Multi-View Subspace Clustering (CVPR 2017 · 285 citations) |
What has been tried
65 papers here report 646 measured changes (a component added, removed or swapped, compared with a baseline): helped 81%, no clear effect 4%, hurt 12%, mixed 3%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Add instance-level contrastive loss | MSRC-V1 (30% missing) · ACC | 80.38 -> 87.72 | Helped | Incomplete Contrastive Multi-View Clustering with High-Confidence Guiding (AAAI 2024 · 109 citations) |
| Replace no FN identification with random-walk FN identification | Scene15 · ACC | 48.0 -> 49.1 | Helped | Decoupled Contrastive Multi-View Clustering with High-Order Random Walks (AAAI 2024 · 92 citations) |
| Add cluster-structure regularization and use cluster-wise anchors instead of the Ours-a anchor variant | MSRC · ACC | 0.7507 -> 0.8418 | Helped | Learning Cluster-Wise Anchors for Multi-View Clustering (AAAI 2024 · 68 citations) |
| add anchor alignment | Dermatology, ForestTypes, BDGP · ACC | 90.87±3.60 -> 91.82±3.78; 70.76±0.05 -> 75.51±0.17; 54.38±0.10 -> 57.97±0.09 | Helped | Learn from View Correlation: An Anchor Enhancement Strategy for Multi-view Clustering (CVPR 2024 · 52 citations) |
| add graph Laplacian regularization | MSRCV · ACC | 66.43±56.21 -> 79.21±0.78 | Helped | Sample-Level Cross-View Similarity Learning for Incomplete Multi-View Clustering (AAAI 2024 · 42 citations) |
| Add the coherence objective loss to DVIMC (vanilla) | Four datasets at different missing view rates · clustering performance | Helped | Deep Variational Incomplete Multi-View Clustering: Exploring Shared Clustering Structures (AAAI 2024 · 28 citations) | |
| add intra-view graph similarity (IGS) exploration | CCV, Caltech102, NUS-WIDE-OBJ, AwA, Cifar-10, YoutubeFace sel · ACC | 15.49 -> 54.79; 17.27 -> 55.47; 17.17 -> 63.95; 8.08 -> 62.53; 65.92 -> 99.94; 11.38 -> 57.33 | Helped | S2MVTC: a Simple yet Efficient Scalable Multi-View Tensor Clustering (CVPR 2024 · 21 citations) |
| Remove view-to-sample weights and implicit view weights | Six real-world datasets, 20% missing · ACC | Hurt | Collaborative Similarity Fusion and Consistency Recovery for Incomplete Multi-view Clustering (AAAI 2025 · 16 citations) | |
| add diffusion loss | CUB, missing rate 0.3 · ACC | 70.96 -> 77.17 | Helped | Incomplete Multi-view Clustering via Diffusion Contrastive Generation (AAAI 2025 · 16 citations) |
| Replace separate view-specific extractors with a shared MoE learner | ALOI · ACC | 54.90±2.42 -> 91.54±0.59 | Helped | Mixture of Experts as Representation Learner for Deep Multi-View Clustering (AAAI 2025 · 13 citations) |
| add re-alignment | Caltech-101, FP ratio 20% · ACC | 46.9 -> 49.9 | Helped | Robust Contrastive Multi-view Clustering against Dual Noisy Correspondence (NeurIPS 2024 · 12 citations) |
| Use granular-ball contrastive loss (p=2) instead of reconstruction loss only | Caltech101-20 · ACC (%) | 40.28 -> 72.63 | Helped | Multi-view Granular-ball Contrastive Clustering (AAAI 2025 · 11 citations) |
What reviewers will expect
Datasets and baselines named in the 34 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- Reuters · 6 of 34 papers
- ALOI-100 · 6 of 34 papers
- HandWritten · 5 of 34 papers
- NoisyMNIST · 4 of 34 papers
- LandUse-21 · 4 of 34 papers
- Scene-15 · 4 of 34 papers
- ORL · 3 of 34 papers
- 100Leaves · 3 of 34 papers
Baselines
- MFLVC · 7 of 34 papers
- ICMVC · 4 of 34 papers
- DCP · 4 of 34 papers
- SURE · 4 of 34 papers
- GCFAgg · 3 of 34 papers
- DealMVC · 3 of 34 papers
- CVCL · 3 of 34 papers
- DIVIDE · 3 of 34 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 95 papers; from OpenAlex and the papers' affiliations).
- National University of Defense Technology · 13 papers
- Intelligent Game and Decision Lab · 8 papers
- Sichuan University · 6 papers
- Nanjing University of Science and Technology · 3 papers
- Southeast University · 3 papers
- Beijing Jiaotong University · 3 papers
- Harbin Institute of Technology · 2 papers
- Shanghai Jiao Tong University · 2 papers
- Nanjing University · 2 papers
- Xidian University · 2 papers
Use it
Download BibTeX of 38 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Clustering approximation algorithms (150 papers)
- Correlation clustering (102 papers)
- Subspace clustering (102 papers)
- Fairness-constrained clustering (19 papers)
- Spectral clustering algorithms (43 papers)
Elsewhere in the topic
- Coresets and dynamic clustering (77 papers)
- k-means++ seeding analysis (76 papers)
- Fair clustering approximation (31 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).