Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
NGBoost: Natural Gradient Boosting for Probabilistic Prediction
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 NGBoost with multiparameter boosting and natural gradients rather than 2nd-order boosting (ablation) | NGBoost · UCI regression datasets | NLL: Boston: 3.57 $\pm$ 0.20 -> 2.43 $\pm$ 0.15; Concrete: 4.21 $\pm$ 0.05 -> 3.04 $\pm$ 0.17 | Helped |
| Use natural gradients rather than ordinary gradients for multiparameter boosting (ablation) | NGBoost · UCI regression datasets | NLL: Boston: 3.17 $\pm$ 0.13 -> 2.43 $\pm$ 0.15; Naval: -3.46 $\pm$ 0.00 -> -5.34$\pm$ 0.04 | Helped |
| Allow variance to vary with covariates instead of assuming homoscedasticity (ablation) | NGBoost · UCI regression datasets | NLL: Boston: 2.79 $\pm$ 0.42 -> 2.43 $\pm$ 0.15; Power: 2.66 $\pm$ 0.11 -> 2.79 $\pm$ 0.11 | Mixed |
| Use NGBoost rather than the best listed probabilistic baseline | NGBoost · Boston, Concrete, Kin8nm, Naval, Power, Protein, Yacht, Year MSD | NLL: Boston: 2.43 $\pm$ 0.15 vs 2.37 $\pm$ 0.24; Kin8nm: -0.49 $\pm$ 0.02 vs -1.20 $\pm$ 0.02 | Hurt |
| Use NGBoost rather than the best listed probabilistic baseline | NGBoost · Energy, Wine | NLL: Energy: 0.60 $\pm$ 0.45 vs 0.66 $\pm$ 0.17; Wine: 0.91 $\pm$ 0.06 vs 0.95 $\pm$ 0.06 | Helped |
| Use NGBoost rather than point-estimation methods | NGBoost · UCI regression benchmarks | RMSE: Boston: NGBoost 2.94 $\pm$ 0.53; Concrete: NGBoost 5.06 $\pm$ 0.61 | Mixed |
Closest papers by meaning
- Statistical Inference for Gradient Boosting Regression (NeurIPS 2025 · similarity 0.86)
- Wasserstein Gradient Boosting: A Framework for Distribution-Valued Supervised Learning (NeurIPS 2024 · similarity 0.85)
- Boosting Methods for Interval-censored Data with Regression and Classification (ICLR 2025 · similarity 0.84)
- CoffeeBoost: Gradient Boosting Native Conformal Inference for Bayesian Optimization (AAAI 2025 · similarity 0.84)
- Robust Minimax Boosting with Performance Guarantees (NeurIPS 2025 · similarity 0.84)
- Understanding Stochastic Natural Gradient Variational Inference (ICML 2024 · similarity 0.84)
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.