engleski

Interoperability of Machine Learning Services: A Use Case

Machine learning has become very popular because the supporting infrastructure is now available for reasonable price and most companies have big data sources for solving problems in a high-dimensional space. There are many business use cases that show successful application of machine learning methods. However, machine learning methods are complex, so many cloud providers (e.g. Google, Amazon, Microsoft) have recently started to offer different machine learning services. SMEs often want to use multiple machine learning services of different cloud providers. In this work, we present current state-of-the-art of machine learning services interoperability research. Additionally, a real use case is implemented to show how SMEs can use two machine learning services of two different cloud providers in one application. Concretely, a web application has been developed by implementing machine learning services of two different cloud service providers. The following services are used within the practical example: text detection with the support of the Google Cloud Platform and text translation services with the support of Microsoft Azure. The idea of the practical example is to show the interoperability of the mentioned platforms.

cloud ; interoperability ; machine learning ; services

nije evidentirano