Abstract
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.
Highlights
Decision trees are a light and efficient machine learning technique that have proved their effectiveness in several classification problems
In this paper we present an accelerator for Gradient Boosting Decision Trees (GBDT) that can execute the GBDT trained with LightGBM
[10] compared the results of Random Forest and GBDT and the results show that GBDT provided better accuracy while using smaller models, we believe that it is a better approach for embedded systems
Summary
Decision trees are a light and efficient machine learning technique that have proved their effectiveness in several classification problems. In the context of embedded systems, energy efficiency is as much important as accuracy, so it is necessary to search for efficient algorithms liable to be accelerated. This makes the decision trees a perfect target to develop an FPGA accelerator. With the support of LightGBM, GBDTs are currently considered one of the most powerful machine learning models due to its efficiency and accuracy. Recently they have been used for many winning solutions in several machine learning competitions [5]. In this work the authors identified that GBDTs present a very interesting tradeoff between the use of computational and hardware resources and the obtained accuracy
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
