Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Variational Sparse Inverse Cholesky Approximation for Latent Gaussian Processes via Double Kullback-Leibler Minimization
Jian Cao, Myeongjong Kang, Felix Jimenez, Huiyan Sang, Florian Tobias Schaefer, Matthias Katzfuss
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 |
|---|---|---|---|
| reduced ancestor sets used instead of full ancestor sets in ELBO (ablation) | DKLGP · synthetic 5D Matern | ELBO approximation error | No clear effect |
| adding global sparsity pattern instead of nearest-neighbor SIC (DKL-G) vs full DKLGP (ablation) | DKLGP prior and posterior sparsity · synthetic 5D Matern | RMSE and NLL | Hurt |
| diagonal variational covariance (DKL-D, mean-field) vs SIC posterior (ablation) | DKLGP posterior family · synthetic 5D Matern | RMSE and NLL | Hurt |
| increasing complexity parameter rho (larger sparsity set) (ablation) | DKLGP · synthetic 5D Matern | RMSE and NLL | Improved |
| DKLGP vs VNNGP on UCI datasets (except Covtype) | DKLGP · UCI datasets | RMSE | Improved |
| DKLGP vs SVIGP (matched m) on KEGGU | DKLGP · UCI KEGGU | RMSE | Hurt |
| Vecchia-Laplace + IC0 initialization only (no ELBO optimization) (ablation) | DKLGP · synthetic 5D Matern | RMSE | Hurt |
Closest papers by meaning
- Variational Inference for Gaussian Process Models with Linear Complexity (NeurIPS 2017 · similarity 0.90)
- Convergence of Sparse Variational Inference in Gaussian Processes Regression (JMLR 2020 · similarity 0.89)
- Scalable Variational Bayesian Kernel Selection for Sparse Gaussian Process Regression (AAAI 2020 · similarity 0.89)
- New Bounds for Sparse Variational Gaussian Processes (ICML 2025 · similarity 0.87)
- Sparse Gaussian Processes: Structured Approximations and Power-EP Revisited (NeurIPS 2025 · similarity 0.87)
- Least squares variational inference (NeurIPS 2025 · similarity 0.87)
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.
