Machine Learning as a Service: Challenges in Research and Applications

Abstract

This study aims to evaluate the current state of research with regards to Machine Learning as a Service (MLaaS) and to identify challenges and research fields of this novel topic. First, a literature review on a basket of eight leading journals was performed. We motivate this study by identifying a lack of studies in the field of MLaaS. The structured literature review was further extended to established scientific databases relevant in this field. We found 30 contributions on MLaaS. As a result of the analysis we grouped them into four key concepts: Platform, Applications; Performance Enhancements and Challenges. Three of the derived concepts are discussed in detail to identify future research areas and to reveal challenges in research as well as in applications.