Very fast Bayesian Additive Regression Trees on GPU
Not on the map. Why?
Abstract (arXiv)
Bayesian Additive Regression Trees (BART) is a nonparametric Bayesian regression technique based on an ensemble of decision trees. It is part of the toolbox of many statisticians. The overall statistical quality of the regression is typically higher than other generic alternatives, and it requires less manual tuning, making it a good default choice. However, it is a niche method compared to its natural competitor XGBoost, due to the longer running time, making sample sizes above 10,000-100,000 a nuisance. I present a GPU-enabled implementation of BART, faster by up to 200x relative to a single CPU core, making BART competitive in running time with XGBoost. This implementation is available in the Python package bartz.
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 10 of its 43 references that are in the map. First posted on arXiv 30 Oct 2024.
Builds on
10 of 43 references.
References · 10
- 2016XGBoost: A Scalable Tree Boosting Systemcited by 5,166
- 2017Automated versus do-it-yourself methods for causal inference: Lessons learned from a data analysis competitioncited by 106
- 2017Bayesian Regression Tree Ensembles that Adapt to Smoothness and Sparsitycited by 34
- 2013bartMachine: Machine Learning with Bayesian Additive Regression Treescited by 24
- 2019Atlantic Causal Inference Conference (ACIC) Data Analysis Challenge 2017cited by 24
- 2020Stochastic tree ensembles for regularized nonlinear regressioncited by 14
- 2022A Mixing Time Lower Bound for a Simplified Version of BARTcited by 7
- 2018XBART: Accelerated Bayesian Additive Regression Treescited by 4
- 2024On the Gaussian process limit of Bayesian Additive Regression Treescited by 4
- 2024On the Computational Efficiency of Bayesian Additive Regression Trees: An Asymptotic Analysiscited by 1
Built on by
Citing papers per year: 2026: 2.
2 papers.