Research map / Learning theory and algorithms / Clustering
A Unified Framework to BRIDGE Complete and Incomplete Deep Multi-View Clustering under Non-IID Missing Patterns
Xiaorui Jiang, Buyun He, Peng Yuan Zhou, Xinyue Chen, Jingcai Guo, Jie Xu, Yong Liao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use the full set of components (training, distribution alignment, view prediction, weighted imputation) (ablation) | BRIDGE · NUSWIDE and BDGP, IID and Non-IID | ACC (%): NUSWIDE CL 44.2 with all components | Improved |
| remove multi-view training on complete data (component a) (ablation) | BRIDGE · NUSWIDE and BDGP | ACC (%) | Hurt |
| add domain adversarial distribution alignment (ablation) | BRIDGE · NUSWIDE, BDGP, Hdigit, Cifar10 visualizations | cluster structure | Improved |
| add weighted intra- and inter-view imputation (component d) (ablation) | BRIDGE · NUSWIDE and BDGP | clustering quality | Improved |
| choose weighting coefficient lambda too large or too small (ablation) | Imputation fusion · BBCSport and Reuters (M=0.5) | ACC | Hurt |
| use contrastive training (BRIDGE+CL) under Non-IID missing patterns | BRIDGE · NUSWIDE Non-IID | ACC (%) | Improved |
| plug BRIDGE into MFLVC, CVCL, GCFAgg | DMVC methods · BDGP, BBCSport, NUSWIDE | FCR (%): FCR 90.3 to 119.0 across settings | Improved |
| use classifier predictions instead of imputation plus k-means clustering (ablation) | BRIDGE · Hdigit, BDGP, BBCSport, Reuters, NUSWIDE | ACC (%): 97.02 vs 97.00 on Hdigit | No clear effect |
| vary training objective (MI, DD, CL) across datasets | BRIDGE · eight datasets | ACC (%) | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- COMPLETER — COMPLETER: Incomplete Multi-View Clustering via Contrastive Prediction (CVPR 2021) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Adaptive Feature Imputation with Latent Graph for Deep Incomplete Multi-View Clustering (AAAI 2024 · similarity 0.93)
- Deep Variational Incomplete Multi-View Clustering: Exploring Shared Clustering Structures (AAAI 2024 · similarity 0.93)
- Incomplete Contrastive Multi-View Clustering with High-Confidence Guiding (AAAI 2024 · similarity 0.92)
- Incomplete Multi-view Clustering via Diffusion Contrastive Generation (AAAI 2025 · similarity 0.91)
- Towards Resource-friendly, Extensible and Stable Incomplete Multi-view Clustering (ICML 2024 · similarity 0.91)
- Diffusion-based Missing-view Generation With the Application on Incomplete Multi-view Clustering (ICML 2024 · similarity 0.90)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
