Research map / Probabilistic and scientific modeling / Variational and generative sampling
Gaussian Process Conditional Density Estimation
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 GP-CDE instead of C-KDE | Conditional density estimation · Manhattan taxi drop-offs, 1K training points | NLPP: 2.40 -> 2.31 | Helped |
| Use GP-CDE instead of C-KDE | Conditional density estimation · Manhattan taxi drop-offs, 5K training points | NLPP: 2.38 -> 2.22 | Helped |
| Use MDN-50 instead of GP-CDE | Conditional density estimation · Manhattan taxi drop-offs, 1M training points | NLPP: 2.13 -> 1.97 | Helped |
| Use GP-CDE instead of CVAE | Image density estimation · MNIST, 2 images per class | Test log-likelihood, optimized σ²: \text{-}1296.63 -> 74.01 | Helped |
| Use CVAE instead of GP-CDE | Image density estimation · MNIST, 512 images per class | Test log-likelihood, optimized σ²: 124.2 -> 244.88 | Helped |
| Use natural gradients instead of Adam for variational parameters (ablation) | Unconditional GP-LVM · 100 MNIST '1' digits | Test log-likelihood: -0.13 -> 1.02 | Helped |
| Use the analytically optimal free-form latent posterior instead of Gaussian q(w) (ablation) | GP-CDE · 10 UCI regression datasets | Test log-likelihood | Helped |
| Use natural gradients instead of ordinary gradients for GP variational parameters (ablation) | GP-CDE inference · Mini-batched training with uncertain inputs | Optimization performance | Helped |
| Condition the density estimator on pick-up location and time instead of ignoring conditions | KDE · Manhattan taxi drop-offs | NLPP | Helped |
| Remove the condition inputs (ablation) | GP-CDE, yielding GP-LVM · Manhattan taxi drop-offs | NLPP: 2.31 vs 2.61; 2.22 vs 2.52; 2.13 vs 2.43 | Hurt |
| Remove the latent variables (ablation) | GP-CDE, yielding GP regression · Manhattan taxi drop-offs | NLPP: 2.31 vs 2.68; 2.22 vs 2.67; 2.13 vs 2.67 | Hurt |
Closest papers by meaning
- Deconvolutional Density Network: Modeling Free-Form Conditional Distributions (AAAI 2022 · similarity 0.87)
- Conditional Density Estimation with Histogram Trees (NeurIPS 2024 · similarity 0.87)
- Deep Gaussian Processes with Importance-Weighted Variational Inference (ICML 2019 · similarity 0.86)
- Amortized Structured Stochastic Variational Inference for Gaussian Process Latent Variable Models (NeurIPS 2026 · similarity 0.86)
- Learning from Summarized Data: Gaussian Process Regression with Sample Quasi-Likelihood (AAAI 2025 · similarity 0.85)
- Sparse Variational Student-t Processes (AAAI 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.