GPUCSL: GPU-Based Library for Causal Structure Learning

被引：1

作者：

Braun, Tom ^{[1
]}

Hurdelhey, Ben ^{[1
]}

Meier, Dominik ^{[1
]}

Tsayun, Petr ^{[1
]}

机构：

[1] Univ Potsdam, Hasso Plattner Inst, Potsdam, Germany

来源：

2022 IEEE INTERNATIONAL CONFERENCE ON DATA MINING WORKSHOPS, ICDMW | 2022年

关键词：

Causal Structure Learning; PC Algorithm; GPU Acceleration; CUDA; !text type='Python']Python[!/text;

D O I：

10.1109/ICDMW58026.2022.00159

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

GPUCSL is a maintainable and extensible Python library for GPU-accelerated causal structure learning (CSL) based on the PC algorithm. The library supports multivariate normal and discrete distributed data, and implements multi-GPU support for multivariate normal distributed data. GPUCSL combines several stand-alone independent research implementations to allow a unified entry point into GPU-accelerated CSL. The library outperforms CPU-based implementations with an average speedup factor of 9.5 against pcalg and 19.8 against bnlearn and remains within the order of magnitude of existing GPU-accelerated CSL research. The source code is available at https://github.com/hpi- epic/gpucsl.

引用

页码：1236 / 1239

页数：4

共 27 条

[21] Bayesian Network Constraint-Based Structure Learning Algorithms: Parallel and Optimized Implementations in the bnlearn R Package [J].