Parallel Support Vector Machines on Multi-Core and Multiprocessor Systems

Abstract

Abstract-This paper proposes a new and efficient parallel implementation of support vector machines based on decomposition method for handling large scale datasets. The parallelizing is performed on the most time-and-memory consuming work of training, i.e., to update the vector f . The inner problems are dealt by sequential minimal optimization solver. Since the underlying parallelism is realized by the shared memory version of Map-Reduce paradigm, our system is easy to build and particularly suitable to apply to multi-core and multiprocessor systems. Experimental results show that on most of the tested datasets, our system offers higher than fourfold increase in speed compared to Libsvm, and it is also far more efficient than the MPI implementation Pisvm.

Embedded Hardware-Efficient Real-Time Classification With Cascade Support Vector Machines

“…The speedups achieved by software implementations of cascade SVM classification schemes over monolithic, although significant, do not offer adequate performance for real-time resource-constrained applications [4,5,6,15,16]. This is because the latter stages become the bottleneck since they require processing an increased number of SVs and the requirement for parallel processing arises.…”

Boosting the Hardware-Efficiency of Cascade Support Vector Machines for Embedded Classification Applications

“…Software implementations of cascade SVM classification schemes [6], [7], [8], [15], [16], [21], have shown speedups over monolithic SVMs and although noteworthy and suitable for some applications, are yet to offer adequate performance for real-time resource-constrained applications. This is due to the fact that the latter stages become the bottleneck since they require processing an increased number of SVs and the requirement for parallel processing arises.…”

A Novel Contactless Human Machine Interface Based on Machine Learning

“…As indicated previously, a first step of FIZI aims at initializing the parameters of the FIZI module by learning the main discriminative features of the background. This learning step consists of a machine learning algorithm similar to the one presented in [24,19] for energy applications [20,18,21,9] and allows to detect key features of the background [10,7,23,22]. This enable to improve the background removal procedure.…”