Research map / Optimization / Optimizers
Stein Variational Gradient Descent With Matrix-Valued Kernels
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 matrix-valued kernels with mixture preconditioning instead of vanilla SVGD | SVGD · MR, CR, SUBJ | Classification error: MR: 19.73 -> 19.09; CR: 18.07 -> 17.13; SUBJ: 6.67 -> 6.59 | Helped |
| Use matrix-valued kernels with mixture preconditioning instead of vanilla SVGD | SVGD · MPQA | Classification error: 10.58 -> 10.71 | Hurt |
| Use mixture rather than average (constant) preconditioning (ablation) | Matrix-SVGD · UCI regression benchmarks | Test RMSE | Mixed |
| Use mixture rather than average (constant) preconditioning (ablation) | Matrix-SVGD · UCI regression benchmarks | Test Log-Likelihood | Mixed |
| Use second-order preconditioning information | Matrix-SVGD · Protein | Test RMSE: 4.186 -> 3.852 | Helped |
| Use second-order preconditioning information | Matrix-SVGD · Protein | Test Log-Likelihood: -2.846 -> -2.755 | Helped |
| Use mixture preconditioning instead of vanilla SVGD | SVGD · Concrete | Test Log-Likelihood: -3.064 -> -3.207 | Hurt |
| Use mixture preconditioning instead of vanilla SVGD | SVGD · Boston | Test RMSE: 2.785 -> 2.717 | Hurt |
Closest papers by meaning
- Stein Variational Gradient Descent as Moment Matching (NeurIPS 2018 · similarity 0.95)
- Message Passing Stein Variational Gradient Descent (ICML 2018 · similarity 0.94)
- Stein Variational Gradient Descent Without Gradient (ICML 2018 · similarity 0.94)
- A Stein variational Newton method (NeurIPS 2018 · similarity 0.94)
- ELBOing Stein: Variational Bayes with Stein Mixture Inference (ICLR 2025 · similarity 0.92)
- Improved Finite-Particle Convergence Rates for Stein Variational Gradient Descent (ICLR 2025 · 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.