A Review of the Usage of Machine Learning in Real-time Systems

Information technologies

Abstract

In this work, we supply a general overview over the usage of machine learning techniques in real-time systems. At present, there is a tendency of a full or partial replacement of a human’s intellectual work by computer programs in every sphere and, for that, there is a need to imitate a human brain i.e. create something like artificial intelligence. On the one hand, machine learning has been quite popular and successfully used in various spheres in recent years. Moreover, the discovery and usage of deep neural networks has immensely increased the efficiency of machine learning usage. On the other hand, as the amount of data greatly increases and changes in quality over time, the usage of real-time systems becomes more and more widespread. So it is quite effective and convenient to use machine learning in real time systems for elaborating a huge amount of newly generated data. Although nowadays there are several machine learning algorithms for classification, regression, clustering etc, their traditional usage as supervised or unsupervised machine learning approach in real-time systems will not be efficient enough because of some nuances that we are going to talk about in this work.