Machine learning applications for mobile devices

Abstract

Research background: This paper explores and compares the possibilities of creating machine-to-machine applications for mobile devices. In today's globalized world, we encounter many areas where machine learning is suitable for various human activities. Machine learning helps organizations in analyzing the reality around us and one of the activities that machine learning deals with is the analysis of digital content and especially digital images. Purpose of the article: The aim of this paper is to bring new perspectives on the possibilities of creating machine learning applications for mobile phones. Currently, there are a large number of options for creating such an application, and not all of them are easy to understand or usable for practical applications. Methods: The methods used in the paper are mainly observation and surveys of existing papers. It is possible to use users’ local hardware and data to solve machine learning problems on mobile devices and paper show some possibilities, advantages, disadvantages and challenges that are connected with this phenomena. Findings & Value added: Machine learning applications are currently on the rise, as there are an ever-increasing number of opportunities to use these applications. There is also an ever-increasing number of different libraries and SDKs for creating such applications. The contribution of this paper is then to compare the different libraries and another contribution is to find out how such applications can be used and more importantly created.