This post is authored by Nagesh Pabbisetty, Partner Director of Program Management at Microsoft.

The pace of adoption of Advanced Analytics (AA) and Machine Learning (ML) among expert data scientists in Tier-1 enterprises is increasing fast. This pace can be significantly increased when enterprise-grade AA and ML are available within the environments where the customers’ data is, infusing intelligence into mission-critical applications is made much easier and, enterprises can turn to a single vendor to make the world of AA and ML is synthesized and supported with the SLAs they have come to expect. At Microsoft, our mission has been to make this vision of ambient intelligence a reality for our customers. We took the first step with Microsoft R Server 9.0, and we are continuing that innovation with this Microsoft R Server 9.1 release.

You can immediately download Microsoft R Server 9.1 from MSDN and Visual Studio Dev Essentials. It comes packed with tons of value built on top of the latest open source R engine, and includes several exciting new capabilities that make R enterprise-class.

State of the Art Machine Learning [Premal]

Bring Machine Learning to where your data is

With MRS 9.0 release, we provided best-of-breed ML algorithms battle-tested by Microsoft as Microsoft ML package, available as a part of Sql Server R Services:

• Fast linear learner, with support for L1 and L2 regularization.
  
• Fast boosted decision tree.
  
• Fast random forest.
  
• Logistic regression, with support for L1 and L2 regularization.
  
• GPU-accelerated Deep Neural Networks (DNNs) with convolutions.
  
• Binary classification using a One-Class Support Vector Machine.
  

We have now made Microsoft ML algorithms portable and distributed, to run on Linux, Windows, and the most popular distributions of Hadoop — Cloudera, Hortonworks, MapR, in addition to SQL Server 2016.

In addition, these capabilities will also be released as Azure VMs in Azure Marketplace, and on Azure HDInsight.

Pre-trained Cognitive Models

We want to make it easy for Enterprises to infuse intelligence into their Line of Business applications. Conventional methods require significant investments of time and effort to hand-craft Machine Learning models from scratch. Harnessing decades of work on cognitive computing in the context of Bing, Office 365 and Xbox, we are delivering the first installment of pre-trained cognitive models that accelerate time to value. Further, these models can be re-trained to optimize for your business domain.

In this release, we are bringing Sentiment Analysis pre-trained cognitive model, where you can assess the sentiment of an English sentence with just one line of code. With Image Featurizer pre-trained cognitive model, you can derive up to 5000 features on a given image, and use that to compare similarity between two images.

Customer Quote here

This blog shows you how to benefit from the power of pre-trained models today.

Combining the best of Microsoft Innovation and Open Source

We are continuing to deliver on the promise of embracing the best of open source, and extending it with the best of Microsoft innovation.

With this release, within the same R script, you can mix and match functions from RevoScaleR and Microsoft ML packages with popular open source packages like SparklyR and through it, H2O.

Customer Quote here

Refer to this blog for examples on how to get the best of both worlds!

Optimized Algorithm for common pattern: Pleasingly Parallel

One of the most popular advanced analytics use cases is Pleasingly Parallel where you run massively parallel computations on partitions that are grouped by one or more attributes. These embarrassingly parallel use cases are very common in a wide variety of industries – Life Sciences simulations to identify the best drug for a give situation, Portfolio analysis to identify the right investment for each portfolio, Utilities to forecast energy consumption for each cohort, Shipping to forecast demand for various container types, etc.,

We have generalized the pattern and have provided a highly performant, and simple yet flexible RxExecBy() function within RevoScaleR, to address all these use cases. Furthermore, this function is portable across all platforms that support Microsoft R Server — Windows, Linux, Hadoop, SQL Server.

Customer Quote here

More details on common patterns, common solutions, and guidance on how to choose the best algorithm for Pleasingly Parallel use-cases are available here.

State-of-the-art Machine Learning Algorithms, across platforms

• Pleasingly Parallel, RxExecBy(), on Windows, Linux, Hadoop, SQL Server
  
• Microsoft ML algos available on Linux, Hadoop (via ensemble methods)
  
• SparklyR – H2O – Rx interop
  
• Pre-trained ML models for Sentiment and Image featurizers
  
• ORC support
  
• RxMerge
  
• Full embrace of Spark2.0 – SparkETL, SparkSQL
  

Where do I bring in the fact that MSL users of MRS will get Python in Preview

SQL Server Machine Learning Services (SQL Server v.next CTP2) [Sumit]

Python (Public Preview)

SQL Server 2016 brought you in-database analytics with SQL Server R Services. With CTP1 of SQL Server v.next, MicrosoftML provided in-database Machine Learning to SQL Server R Services users.

Now, CTP2 of SQL Server v.next, brings you SQL Server Machine Learning Services that embraces both R and Python as first class citizens. Starting from the CTP 2 release, the preview version of Python brings the latest and greatest innovation in the world of Machine Learning to where your data lives. Data Scientists can now choose from a huge collection of algorithms across R and Python communities to execute in-database and get their job done much more effectively.

SQL Server v.next CTP2 enables collaboration between traditional data scientists with strong R backgrounds and computer scientists with strong Python backgrounds, to deliver the best business ROI.

Refer to this blog for examples on how to get the best of both R and Python worlds!

Real-time Scoring – Rx and MML

We set the industry benchmark for scoring at 1 Million predictions per second. Now, we have improved scoring performance significantly, up to two orders of magnitude than earlier versions.

For example, XXX algorithm runs YYY times faster in SQL Server v.next CTP2 compared to CTP1. [ show one example from revo and another from MicrosoftML ]

Flexible Package Management

Assigning permissions

Fail-over

Scope of installed packages

RxSync()

This blog shows you the full set of capabilities delivered as a part of SQL Server v.next CTP2.

…

• SQL Server users now have the world or ML (and most toolkits, if not all, come with Python bindings)….and,
  
• For Python users, they can now run Python scripts in-db within SQL Server….gaining all the efficiencies of in-db analytics.
  
• Python Preview (subset of Rx algos, no MML)
  
• Real-time Scoring – Rx and MML
  
• Improved Package install/uninstall; RxSync package (user-initiated restore)
  

Microsoft R Server Operationalization [CARL]

• Real-time scoring: One to two orders of magnitude improvement in scoring performance….
  
• Role based access control:
  
• Asynchronous batch processing
  
• Dynamic scaling of operationalization grid with Azure VMs
  

This blog shows you all the new Operationalization capabilities introduced by R Server 9.1.

From Kantar (a WPP company): For the very first time we have access to a suite of products from an industry giant that fill in the historical gaps in our vision and can really help our data scientists drive our evolution forward. The Kantar-level vision is to position itself as a game changing, analytics power house and to bring greater value and insight to Kantar clients by embedding consistent, automated analytic techniques and reporting in each of our survey-based products or custom solutions. With Microsoft R Server, we are now able to create advanced analytics models across multiple types of data on a world-wide basis, and operationalize them at scale. Microsoft has been working closely with us on a variety of features, most recently in making operationalization a cinch with asynchronous batch processing and auto-scaling of operationalization grid on Azure.

Azure VMs

Microsoft R Server has been available as an Azure VM on CentOS 7.2 and Ubuntu 16.04. With Microsoft R Server 9.1, we have added support for RHEL 7.2.

These VMs are available in all Azure regions, including China.

We will also refresh the Data Science VM with this latest release.

Customer Quote here

Solution Templates [Bharath]

In our last release, we provided Solutions Template for Campaign Optimization using SQL Server R Services, to make it easy for Data Engineers and Data Scientists to understand how to create full solutions based on SQL Server 2016. We have now created an equivalent solution template for Azure HDInsight platform on Spark compute context as well .

Hospital Length Of Stay (LOS) is the latest solution template built on SQL Server R Services….

The links below walk through each of the solutions and also a video that walkthrough both of these solutions

Campaign Optimization

Hospital LOS
R Solution Templates
Channel 9 Video walkthrough of Solution How-to’s

Customer Quote here

Making it easy to adopt…closed loop development…customer partnership

R Advisors – We have launched R Advisors where customers can have access to preview releases and also have the opportunity to provide feedback before we release. Azure Advisors is an opt-in program that you can sign up as a member of your organization.

We are also launching User Voice and these will be

• UserVoice
  
• MSDN R Blogs, Tiger R Blogs, Revolutions
  
• eBook on SQL
  
• Certifications
  
• Training Partners
  

Development Environments & 3^rd party integrations for Developers and Data Scientists

• RTVS GA
  
• PTVS
  
• Azure Jupyter Notebook Service
  
• Alteryx integration
  
• KNIME integration
  

In Summary [Nagesh]

I am proud of the many MRS enhancements that our team has delivered during this calendar year, including support for Hadoop, Spark, Linux, Teradata and HDInsight, and the addition of R analytics to SQL Server 2016, to name a few.

MRS 9.1 is a culmination of all our hard work this year. With this latest release, you have access to a powerful tool, one that supports popular operating systems and a variety of data sources, helps you create sophisticated analytics models and deploy them in the real world, efficiently and at scale. We invite you to get started with Microsoft R Server 9.0.

Nagesh