Very fast Bayesian Additive Regression Trees on GPU

Giacomo Petrillo
10/30/2024
stat.MLcs.LG

No official code found. Best implementation: bartz-org/bartz (102 stars, MIT, last commit 2026-04-26). 1 more below.

Abstract

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.

View on arXivView PDF

Code Implementations(2)

bartz-org/bartzCommunity100%
MIT

Super-fast BART (Bayesian Additive Regression Trees) in Python

1027R, HTMLFeb 19, 20245 months agoMIT
Gattocrucco/bart-gpu-articleCommunity61%

Code for Petrillo (2024), "Very fast Bayesian Additive Regression Trees on GPU"

00TeX, PythonOct 29, 20245 months ago

Similar Papers

ASBART:Accelerated Soft Bayes Additive Regression Trees

84% match2023

On the Gaussian process limit of Bayesian Additive Regression Trees

81% match1 repo2024

The Computational Curse of Big Data for Bayesian Additive Regression Trees: A Hitting Time Analysis

80% match2024

Oblique Bayesian additive regression trees

78% match2024

Bayesian Additive Regression Networks

78% match1 repo2322024

Bayesian Additive Regression Trees for functional ANOVA model

77% match2025

Cite this paper

@article{petrillo2024very,
  title  = {Very fast Bayesian Additive Regression Trees on GPU},
  author = {Giacomo Petrillo},
  year   = {2024},
  eprint = {2410.23244},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2410.23244v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2410.23244.svg)](https://codeofpaper.com/paper/2410.23244)

Discussion