Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Hierarchical Gaussian Process Priors for Bayesian Neural Network Weights
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 |
|---|---|---|---|
| add input-dependent periodic kernel to the GP weight prior | MetaGP · Sinusoid extrapolation | RMSE: 2.59 -> 0.12 | Helped |
| add input-dependent RBF kernel to the GP weight prior | MetaGP · Sinusoid extrapolation | RMSE: 2.59 -> 0.74 | Helped |
| add input-dependent RBF kernel to the GP weight prior | MetaGP · Sinusoid interpolation | RMSE: 0.24 -> 0.11 | Helped |
| use periodic instead of RBF input kernel | MetaGP-Local · Sinusoid extrapolation | NLL: 1.50 -> -0.69 | Helped |
| use periodic instead of RBF input kernel | MetaGP-Local · Sinusoid interpolation | NLL: -0.80 -> -0.76 | No clear effect |
| replace independent-prior BNN-MFVI with MetaGP-Local[Periodic] | BNN · Sinusoid extrapolation | NLL: 88.12 -> -0.69 | Helped |
| apply MetaGP-local to last layer of MNIST BNN | BNN classifier · MNIST OOD predictive entropy | entropy CDF | Mixed |
| use diagonal approximation to q(w|z) instead of full covariance (ablation) | MetaGP · toy regression/classification | qualitative fit | No clear effect |
| use MetaGP prior instead of MFVI for active learning | BNN · UCI regression active learning | RMSE | Mixed |
Closest papers by meaning
- Bayesian Neural Network Priors Revisited (ICLR 2022 · similarity 0.90)
- Understanding Priors in Bayesian Neural Networks at the Unit Level (ICML 2019 · similarity 0.89)
- Deep Kernel Posterior Learning under Infinite Variance Prior Weights (ICLR 2025 · similarity 0.86)
- Bayesian Optimization via Continual Variational Last Layer Training (ICLR 2025 · similarity 0.86)
- Linearization Turns Neural Operators into Function-Valued Gaussian Processes (ICML 2025 · similarity 0.86)
- Beyond IID weights: sparse and low-rank deep Neural Networks are also Gaussian Processes (ICLR 2024 · similarity 0.85)
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.