Research map / Probabilistic and scientific modeling / Variational and generative sampling
Copula-SVI: Vine-Copula Variational Inference with Stein Refining for Instance-Level Correlation Capturing
Junxi Xiao, Qinliang Su
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 |
|---|---|---|---|
| add SVGD refinement steps (0 to 5) (ablation) | Copula-SVI 10-order · MNIST constrained clustering | ACC (%): 97.6 -> 98.6 | Improved |
| add SVGD refinement steps (0 to 5) (ablation) | Copula-SVI 10-order · Fashion-MNIST constrained clustering | ACC (%): 81.6 -> 83.8 | Improved |
| add SVGD refinement steps (0 to 5) (ablation) | Copula-SVI 10-order · MNIST constrained clustering | ARI (%): 94.2 -> 96.7 | Improved |
| raise vine dependence order from 5 to 100 (ablation) | Copula-SVI · MNIST constrained clustering | ACC (%): 98.3 -> 99.0 | Improved |
| raise vine dependence order from 5 to 100 (ablation) | Copula-SVI · STL-10 constrained clustering | ACC (%): 92.7 -> 93.3 | Improved |
| raise vine dependence order from 3 to 100 (ablation) | Copula-SVI · SMAP anomaly detection | F1: 0.8628 -> 0.8871 | Improved |
| raise vine dependence order from 3 to 100 (ablation) | Copula-SVI · MSL anomaly detection | ELBO: -154.5485 -> -91.1058 | Improved |
| raise vine dependence order from 3 to 100 (ablation) | Copula-SVI · Multivariate forecasting average | MSE: 0.487 -> 0.326 | Improved |
| use Copula-SVI instead of HoT-VI at 3-order | Copula-SVI · Multivariate forecasting average | MSE: 0.573 -> 0.487 | Improved |
| use Copula-SVI 10-order instead of HoT-VI 10-order | Copula-SVI · MNIST constrained clustering | ACC (%): 98.5 -> 98.6 | Improved |
| use Copula-SVI 10-order instead of HoT-VI 10-order | Copula-SVI · Reuters constrained clustering | ACC (%): 97.6 -> 98.0 | Improved |
| use Copula-SVI instead of tree-restricted TreeVI | Copula-SVI 10-order · Fashion-MNIST constrained clustering | ACC (%): 81.4 -> 83.8 | Improved |
Closest papers by meaning
- HoT-VI: Reparameterizable Variational Inference for Capturing Instance-Level High-Order Correlations (NeurIPS 2025 · similarity 0.90)
- TreeVI: Reparameterizable Tree-structured Variational Inference for Instance-level Correlation Capturing (NeurIPS 2024 · similarity 0.89)
- Copula-like Variational Inference (NeurIPS 2019 · similarity 0.86)
- Nearly Dimension-Independent Convergence of Mean-Field Black-Box Variational Inference (NeurIPS 2025 · similarity 0.85)
- Variational Inference with Mixtures of Isotropic Gaussians (NeurIPS 2025 · similarity 0.84)
- Least squares variational inference (NeurIPS 2025 · similarity 0.84)
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.
