Machine Learning Optimization Improves Engineering Model Accuracy

Abstract

Optimization in machine learning provides a game changing approach for engineering applications, where machine learning models can achieve greater than 99 % accuracy, and less than 1 % mean relative error in classification task. Traditional mathematical optimization techniques are strong, however machine learning allows for an unprecedented precision in hard to optimize analysis tasks. Artificial intelligence optimization flexes its power well beyond the basics of model improvement. In fact, the use of machine learning optimization algorithms in conjunction with conventional methods has reduced computational times considerably as compared to times in excess of 1000 seconds in some instances to less than 0.1 seconds in others. This breakthrough efficiency applies broadly in electricity distribution, finance, logistics, and spans many other sectors of engineering challenges. In this paper, we investigate how machine learning optimization extends to engineering model accuracy and what has been done in terms of deploying different algorithms, how they have been implemented and applications to real world problems.