Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
A variational approximate posterior for the deep Wishart process
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 DWP inference instead of DGP inference with the same prior | five-layer model · Boston | ELBO: -0.45 $\pm$ 0.01 -> -0.38 $\pm$ 0.01 | Helped |
| Use DWP inference instead of DGP inference with the same prior | five-layer model · Concrete | ELBO: -0.50 $\pm$ 0.00 -> -0.49 $\pm$ 0.00 | Helped |
| Use DWP inference instead of DGP inference with the same prior | five-layer model · Energy | ELBO: 1.39 $\pm$ 0.00 -> 1.41 $\pm$ 0.00 | Helped |
| Use DWP inference instead of DGP inference with the same prior | five-layer model · Kin8nm | ELBO: -0.13 $\pm$ 0.00 -> -0.14 $\pm$ 0.00 | Hurt |
| Use DWP inference instead of DGP inference with the same prior | five-layer model · Naval | ELBO: 3.91 $\pm$ 0.10 -> 3.65 $\pm$ 0.07 | Hurt |
| Use DWP inference instead of DGP inference with the same prior | five-layer model · Power | ELBO: 0.02 $\pm$ 0.00 -> 0.03 $\pm$ 0.00 | Helped |
| Use DWP inference instead of DGP inference with the same prior | five-layer model · Protein | ELBO: -1.00 $\pm$ 0.00 -> -1.01 $\pm$ 0.00 | Hurt |
| Use DWP inference instead of DGP inference with the same prior | five-layer model · Wine | ELBO: -1.19 $\pm$ 0.00 -> -1.19 $\pm$ 0.00 | No clear effect |
| Use DWP inference instead of DGP inference with the same prior | five-layer model · Yacht | ELBO: 1.37 $\pm$ 0.03 -> 1.65 $\pm$ 0.01 | Helped |
| Use DWP inference instead of DGP inference with the same prior | five-layer model · Boston | test log likelihood: -2.44 $\pm$ 0.04 -> -2.39 $\pm$ 0.04 | Helped |
| Use DWP inference instead of DGP inference with the same prior | five-layer model · Concrete | test log likelihood: -3.14 $\pm$ 0.02 -> -3.13 $\pm$ 0.01 | Helped |
| Use DWP inference instead of DGP inference with the same prior | five-layer model · Energy | test log likelihood: -0.70 $\pm$ 0.03 -> -0.70 $\pm$ 0.03 | No clear effect |
Closest papers by meaning
- Deep kernel processes (ICML 2021 · similarity 0.95)
- Sparse Inducing Points in Deep Gaussian Processes: Enhancing Modeling with Denoising Diffusion Variational Inference (ICML 2024 · similarity 0.87)
- Convolutional Deep Kernel Machines (ICLR 2024 · similarity 0.86)
- Self-Attention through Kernel-Eigen Pair Sparse Variational Gaussian Processes (ICML 2024 · similarity 0.86)
- State-of-art minibatches via novel DPP kernels: discretization, wavelets, and rough objectives (NeurIPS 2026 · similarity 0.86)
- Deep Kernel Posterior Learning under Infinite Variance Prior Weights (ICLR 2025 · similarity 0.86)
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.