Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Implicit Posterior Variational Inference for Deep Gaussian Processes
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 IPVI DGP 4 instead of the strongest reported baseline | DGP classification · MNIST | test accuracy (%): 97.55 -> 97.80 | Helped |
| Use IPVI DGP 4 instead of the strongest reported baseline | DGP classification · MNIST (M=800) | test accuracy (%): 98.05 -> 98.23 | Helped |
| Use IPVI DGP 4 instead of the strongest reported baseline | DGP classification · Fashion-MNIST | test accuracy (%): 87.99 -> 88.90 | Helped |
| Use IPVI DGP 4 instead of the strongest reported baseline | DGP classification · CIFAR-10 | test accuracy (%): 52.81 -> 53.27 | Helped |
| Add parameter tying (ablation) | IPVI DGP 4 · MNIST | test accuracy (%): 97.45 -> 97.80 | Helped |
| Add parameter tying (ablation) | IPVI DGP 4 · Fashion-MNIST | test accuracy (%): 88.01 -> 88.90 | Helped |
| Add parameter tying (ablation) | IPVI DGP 4 · CIFAR-10 | test accuracy (%): 52.76 -> 53.27 | Helped |
| Add parameter tying (ablation) | IPVI DGP 2 · Boston | test MLL: -2.37 -> -2.08 | Helped |
| Add parameter tying (ablation) | IPVI DGP 4 · Energy | test MLL: -0.47 -> -0.01 | Helped |
| Use parameter tying in the generator and discriminator | IPVI · Airline dataset, 4-layer DGP | training speed: 9 times faster during training | Helped |
| Generate posterior samples in parallel and use parameter tying | IPVI · Airline dataset, 4-layer DGP | posterior sample generation speed: 500 times faster | Helped |
| Increase generator parameter count | IPVI synthetic multi-modal posterior estimation · Diamond synthetic dataset | Jensen-Shannon divergence | Helped |
Closest papers by meaning
- Doubly Stochastic Variational Inference for Deep Gaussian Processes (NeurIPS 2017 · similarity 0.89)
- Sparse Inducing Points in Deep Gaussian Processes: Enhancing Modeling with Denoising Diffusion Variational Inference (ICML 2024 · similarity 0.88)
- Revisiting Unbiased Implicit Variational Inference (ICML 2025 · similarity 0.86)
- Bayesian Basis Function Approximation for Scalable Gaussian Process Priors in Deep Generative Models (ICML 2025 · similarity 0.86)
- VIKING: Deep variational inference with stochastic projections (NeurIPS 2025 · similarity 0.86)
- Natural Gradient VI: Guarantees for Non-Conjugate Models (NeurIPS 2025 · 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.