Research map / Label-efficient and robust learning / Recognition and long-tail
A Probabilistic Hierarchical Model for Multi-View and Multi-Feature Classification
Jinxing Li, Hongwei Yong, Bob Zhang, Mu Li, Lei Zhang, David Zhang
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 hierarchical fusion (J,Kj)=(3,3) instead of single view with all 9 features (ablation) | HMMF · synthetic | accuracy: 68.6% -> 86.8% | Improved |
| use hierarchical fusion instead of concatenating features per view (ablation) | HMMF · synthetic (3,3) | accuracy: 85.0% -> 86.8% | Improved |
| use hierarchical fusion instead of treating each feature as separate view (ablation) | HMMF · synthetic (3,3) | accuracy: 84.8% -> 86.8% | Improved |
| increase number of views and features per view from (2,2) to (4,4) (ablation) | HMMF · synthetic | accuracy: 82.2% -> 93.6% | Improved |
| increase latent dimension Dj from 1 to 8 (ablation) | HMMF · Wiki Text-Image | accuracy: 44.0% -> 71.1% | Improved |
| increase latent dimension Dj from 8 to 10 (ablation) | HMMF · Wiki Text-Image | accuracy: 71.1% -> 69.4% | Hurt |
| random vs PCA-based initialization of parameters (ablation) | HMMF · synthetic (3,3) | accuracy: 85.2%-86.6% over 10 random inits | No clear effect |
| HMMF vs best baseline | HMMF · Biomedical DM (70 train) | AUC: 0.8703 -> 0.9027 | Improved |
| HMMF vs best baseline | HMMF · Wiki Text-Image | accuracy: 69.1% -> 71.1% | Improved |
Closest papers by meaning
- Infinite Latent Feature Selection: A Probabilistic Latent Graph-Based Ranking Approach (ICCV 2017 · similarity 0.86)
- Multi-view Anomaly Detection via Robust Probabilistic Latent Variable Models (NeurIPS 2016 · similarity 0.85)
- Multi-View Correlated Feature Learning by Uncovering Shared Component (AAAI 2017 · similarity 0.85)
- Multi-view Learning as a Nonparametric Nonlinear Inter-Battery Factor Analysis (JMLR 2021 · similarity 0.84)
- Discrete Variational Autoencoders (ICLR 2017 · similarity 0.84)
- Multi-View Multi-Label Classification via View-Label Matching Selection (AAAI 2025 · similarity 0.83)
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.
