Microsoft R Server 9.0 enhances machine learning efforts

Microsoft has released R Server 9.0 to bring machine learning close to servers. The new update includes some cutting-edge algorithms within an all-new MicrosoftML package to make server technologies smarter than ever before.

Microsoft R Server 9.0 allows a combination of the latest algorithms with the conventional open source CRAN R packages and preloaded parallel external memory algorithms like the RevoScaleR package. There are offerings like a fast linear learner, GPUaccelerated Deep Neural Networks with convolutions and fast random forest.

Alongside the algorithms, the updated R Server improves some existing operationalisation capabilities. There is an option to convert existing R models and scripts into Web services using a single-line code on IDE such as R Tools for Visual Studio, RStudio or Jupyter Notebooks. Also, the R Server simplifies the application integration experience from the Swagger standard and brings crossplatform support to train models.

Microsoft has also introduced support for Spark 2.0, an upgrade from

Spark 1.6. This change offers new methodologies for handling streaming data and includes an advanced memory management sub-system to improve performance.

“With this latest release, you have access to a powerful tool, one that supports popular operating systems and a variety of data sources, helping you to create sophisticated analytics models and deploy them in the real world, efficiently and at scale,” said Nagesh Pabbisetty, partner director of programme management, Microsoft, in a statement.

Microsoft R Server 9.0 can be deployed on Ubuntu in addition to other platforms. This means the new R Server can be a part of SUSE and Red Hat