RUC. Repositorio da Universidade da Coruña
Publicación
Artículo científico (JournalArticle)
CUDA acceleration of MI-based feature selection methods
[Abstract]: Feature selection algorithms are necessary nowadays for machine learning as they are capable of removing irrelevant and redundant information to reduce the dimensionality of the data and improve the quality of subsequent analyses. The problem with current feature selection approaches is that they are computationally expensive when processing large datasets. This work presents parallel implementations for Nvidia GPUs of three highly-used feature selection methods based on the Mutual Information (MI) metric: mRMR, JMI and DISR. Publicly available code includes not only CUDA implementations of the general methods, but also an adaptation of them to work with low-precision fixed point in order to further increase their performance on GPUs. The experimental evaluation was carried out on two modern Nvidia GPUs (Turing T4 and Ampere A100) with highly satisfactory results, achieving speedups of up to 283x when compared to state-of-the-art C implementations., This work was supported by grants PID2019-104184RB-I00, PID-
2019-109238GB-C22, TED2021-130599A-I00 and PID2022-136435NB-
I00, funded by MCIN/AEI/ 10.13039/501100011033 (TED2021 also
funded by “NextGenerationEU”/PRTR and PID2022 by “ERDF A way of
making Europe”, EU). Grant TSI-100925-2023-1, funded by Ministry
for Digital Transformation and Civil Service. FPU predoctoral grant
of Bieito Beceiro ref. FPU20/00997, funded by the Ministry of Sci-
ence, Innovation and Universities. We gratefully thank the Galician
Supercomputing Center (CESGA) for the access granted to its super-
computing resources. Funding for open access charge: Universidade da
Coruña/CISUG., Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG
Proyecto:
AEI / Plan Estatal de Investigación Científica y Técnica y de Innovación 2017-2020 PID2019-104184RB-I00
,
AEI / Plan Estatal de Investigación Científica y Técnica y de Innovación 2017-2020 PID2019-109238GB-C22
,
AEI / Plan Estatal de Investigación Científica y Técnica y de Innovación 2021-2023 PID2022-136435NB-I00
,
AEI / Plan Estatal de Investigación Científica y Técnica y de Innovación 2021-2024 TED2021-130599A-I00
,
MINECO / Plan Estatal de Investigación Científica y Técnica y de Innovación 2021-2023 TSI-100925-2023-1
,
MECD / Plan Estatal de Investigación Científica y Técnica y de Innovación 2021-2023 FPU20-00997
HANDLE: http://hdl.handle.net/2183/36386
Ver en: http://hdl.handle.net/2183/36386