FPGA Accelerator for Gradient Boosting Decision Trees
A decision tree is a well-known machine learning technique. Recently their popularity has increased due to the powerful Gradient Boosting ensemble method that allows to gradually increasing accuracy at the cost of executing a large number of decision trees. In this paper we present an accelerator de...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2021-01-01
|
Series: | Electronics |
Subjects: | |
Online Access: | https://www.mdpi.com/2079-9292/10/3/314 |
_version_ | 1797406094051508224 |
---|---|
author | Adrián Alcolea Javier Resano |
author_facet | Adrián Alcolea Javier Resano |
author_sort | Adrián Alcolea |
collection | DOAJ |
description | A decision tree is a well-known machine learning technique. Recently their popularity has increased due to the powerful Gradient Boosting ensemble method that allows to gradually increasing accuracy at the cost of executing a large number of decision trees. In this paper we present an accelerator designed to optimize the execution of these trees while reducing the energy consumption. We have implemented it in an FPGA for embedded systems, and we have tested it with a relevant case-study: pixel classification of hyperspectral images. In our experiments with different images our accelerator can process the hyperspectral images at the same speed at which they are generated by the hyperspectral sensors. Compared to a high-performance processor running optimized software, on average our design is twice as fast and consumes 72 times less energy. Compared to an embedded processor, it is 30 times faster and consumes 23 times less energy. |
first_indexed | 2024-03-09T03:20:11Z |
format | Article |
id | doaj.art-44fea6a06bf542f79ce4150ea0e5ed07 |
institution | Directory Open Access Journal |
issn | 2079-9292 |
language | English |
last_indexed | 2024-03-09T03:20:11Z |
publishDate | 2021-01-01 |
publisher | MDPI AG |
record_format | Article |
series | Electronics |
spelling | doaj.art-44fea6a06bf542f79ce4150ea0e5ed072023-12-03T15:10:59ZengMDPI AGElectronics2079-92922021-01-0110331410.3390/electronics10030314FPGA Accelerator for Gradient Boosting Decision TreesAdrián Alcolea0Javier Resano1Department of Computer Science and Systems Engineering (DIIS), University of Zaragoza, c/Maria de Luna 1, 50018 Zaragoza, SpainEngineering Research Institute of Aragon (I3A), University of Zaragoza, c/Mariano Esquillor SN, 50018 Zaragoza, SpainA decision tree is a well-known machine learning technique. Recently their popularity has increased due to the powerful Gradient Boosting ensemble method that allows to gradually increasing accuracy at the cost of executing a large number of decision trees. In this paper we present an accelerator designed to optimize the execution of these trees while reducing the energy consumption. We have implemented it in an FPGA for embedded systems, and we have tested it with a relevant case-study: pixel classification of hyperspectral images. In our experiments with different images our accelerator can process the hyperspectral images at the same speed at which they are generated by the hyperspectral sensors. Compared to a high-performance processor running optimized software, on average our design is twice as fast and consumes 72 times less energy. Compared to an embedded processor, it is 30 times faster and consumes 23 times less energy.https://www.mdpi.com/2079-9292/10/3/314decision treesGBDTFPGAenergy efficiency |
spellingShingle | Adrián Alcolea Javier Resano FPGA Accelerator for Gradient Boosting Decision Trees Electronics decision trees GBDT FPGA energy efficiency |
title | FPGA Accelerator for Gradient Boosting Decision Trees |
title_full | FPGA Accelerator for Gradient Boosting Decision Trees |
title_fullStr | FPGA Accelerator for Gradient Boosting Decision Trees |
title_full_unstemmed | FPGA Accelerator for Gradient Boosting Decision Trees |
title_short | FPGA Accelerator for Gradient Boosting Decision Trees |
title_sort | fpga accelerator for gradient boosting decision trees |
topic | decision trees GBDT FPGA energy efficiency |
url | https://www.mdpi.com/2079-9292/10/3/314 |
work_keys_str_mv | AT adrianalcolea fpgaacceleratorforgradientboostingdecisiontrees AT javierresano fpgaacceleratorforgradientboostingdecisiontrees |