Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Quantile Propagation for Wasserstein-Approximate Gaussian Processes
Rui Zhang, Christian Walder, Edwin V. Bonilla, Marian-Andrei Rizoiu, Lexing Xie
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 |
|---|---|---|---|
| Replace forward-KL projection with L₂ Wasserstein quantile projection | Gaussian-process classification · IonoS, Cancer, Pima, Crabs, Sonar, Glass, Wine1–Wine3 | test error: IonoS 7.9 -> 7.9; Cancer 3.2 -> 3.2; Pima 20.3 -> 20.3; Crabs 2.7 -> 2.7; Sonar 14.0 -> 14.0; Glass 1.1 -> 1.0; Wine1 1. | No clear effect |
| Replace forward-KL projection with L₂ Wasserstein quantile projection | Gaussian-process classification · Pima | NTLL (× 10⁻³): 424.7 -> 424.0 | Improved |
| Replace forward-KL projection with L₂ Wasserstein quantile projection | Gaussian-process classification · Glass | NTLL (× 10⁻³): 29.5 -> 29.0 | Improved |
| Replace forward-KL projection with L₂ Wasserstein quantile projection | Gaussian-process classification · Wine1 | NTLL (× 10⁻³): 48.0 -> 47.4 | Improved |
| Replace forward-KL projection with L₂ Wasserstein quantile projection | Gaussian-process classification · Wine2 | NTLL (× 10⁻³): 18.0 -> 17.8 | Improved |
| Replace forward-KL projection with L₂ Wasserstein quantile projection | Gaussian-process classification · Wine3 | NTLL (× 10⁻³): 52.1 -> 51.8 | Improved |
| Replace forward-KL projection with L₂ Wasserstein quantile projection | Gaussian-process classification · IonoS, Cancer, Crabs, Sonar | NTLL (× 10⁻³): IonoS 215.9 -> 215.9; Cancer 88.2 -> 88.2; Crabs 64.4 -> 64.3; Sonar 306.7 -> 306.2 | No clear effect |
| Replace forward-KL projection with L₂ Wasserstein quantile projection | Gaussian-process Poisson regression · Mining | NTLL (× 10⁻³): 1606.8 -> 1606.5 | Improved |
| Replace forward-KL projection with L₂ Wasserstein quantile projection | Gaussian-process Poisson regression · Mining | TE: 118.6 -> 118.6 | No clear effect |
| Use QP rather than variational Bayes | Gaussian-process classification · All classification datasets except Cancer | test error | Improved |
| Use QP rather than variational Bayes | Gaussian-process classification · Cancer | test error: 3.2 vs 3.1 | Hurt |
| Use QP rather than variational Bayes | Gaussian-process classification · All classification datasets | NTLL | Improved |
Closest papers by meaning
- Efficient Bayesian Inference of Sigmoidal Gaussian Cox Processes (JMLR 2018 · similarity 0.88)
- A Unifying Framework for Gaussian Process Pseudo-Point Approximations using Power Expectation Propagation (JMLR 2017 · similarity 0.88)
- Efficiently sampling functions from Gaussian process posteriors (ICML 2020 · similarity 0.87)
- Turbocharging Gaussian Process Inference with Approximate Sketch-and-Project (NeurIPS 2025 · similarity 0.86)
- Conditioning Gaussian Processes on Almost Anything (NeurIPS 2026 · similarity 0.86)
- Bayesian Optimization through Gaussian Cox Process Models for Spatio-temporal Data (ICLR 2024 · 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.
