Implementation of Parallel Programming Method that Support Various Machine Learning Algorithms

Abstract

As we are towards the start of the multiprocessing period were there are lot of advances in the field of technology, PCs will have progressively numerous centers and computing capacity however there is still horrible programming structure for these designs, and therefore no straightforward and brought together path for AI to exploit the possibly accelerate. In this paper, we build up a method that is comprehensively applicable equal programming technique, one that is effortlessly applied to a wide range of learning calculations in the field of machine learning. Our work is in unmistakable differentiation to the convention in AI of structuring approaches to accelerate a solitary calculation at once. In particular, we show that calculations that fit the Statistical Query model can be written in a specific "summation structure," which permits them to be effectively standard allelized on multicore PCs.  We propose the methods for parallel programming that direct relapse k-implies, strategic regressioninnocent Bayes, Gaussian discriminant examination and back propagation.