Efficient support vector machines implementation on Intel/Movidius Myriad 2
Charalampos Marantos, Nikolaos Karavalakis, Vasileios Leon, Vasileios Tsoutsouras, Kiamal Pekmestzi, Dimitrios Soudris · 2018
Support Vector Machines (SVM) classifiers are widely used as inference tools in Internet of Things (IoT) and Edge Computing applications. To achieve high classification accuracy, the SVM classifier can turn out to be the computationally intensive and power hungry component of the application. In this paper, we enable an efficient SVM implementation, in terms of performance and power dissipation, on an ultra-low-power multi-core SoC, Intel/Movidius Myriad 2. Experimental results highlight the efficiency of the proposed solution, as it achieves 105 × speed-up compared to its initial porting and up to 40% energy savings against state-of-the-art relevant approaches.