21

Jun

Data Virtualization: Unlocking Data for AI and Machine Learning

This post is authored by Robert Alexander, Senior Software Engineer at Microsoft.

For reliability, accuracy and performance, both AI and machine learning heavily rely on large sets. Because the larger the pool of data, the better you can train the models. That’s why it’s critical for big data platforms to efficiently work with different data streams and systems, regardless of the structure of the data (or lack thereof), data velocity or volume.

However, that’s easier said than done.

Today every big data platform faces these systemic challenges:

Compute / Storage Overlap: Traditionally, compute and storage were never delineated. As data volumes grew, you had to invest in compute as well as storage.
Non-Uniform Access of Data: Over the years, too much dependency on business operations and applications have led companies to acquire, ingest and store data in different physical systems like file systems, databases and data warehouses (e.g. SQL Server or Oracle), big data systems (e.g. Hadoop). This results in disparate systems, each with its own method of accessing data.
Hardware-Bound Compute: You have your data in nice storage schema (e.g. SQL Server), but then you’re hardware constrained to execute your query as it takes several hours to complete.

21

Jun

From cranes to shipping containers, IoT helps ZPMC transform the world’s ports

Source: https://blogs.microsoft.com/iot/2017/06/21/from-cranes-to-shipping-containers-iot-helps-zpmc-transform-the-worlds-ports/     Looking to transform the world’s shipping ports, Shanghai-based ZPMC is using the Internet of Things (IoT) to connect equipment ranging from giant cranes to shipping containers. Already the world’s largest manufacturer of port cranes, ZPMC is READ MORE

21

Jun

Help Improve Power BI Desktop
Help Improve Power BI Desktop

Source: https://powerbi.microsoft.com/en-us/blog/help-improve-power-bi-desktop-4/           We love hearing feedback from the Power BI community.​ This helps us to understand what we’re doing well, and what we can change to make your experience even better. Once a quarter, we READ MORE

21

Jun

Announcing Data Refresh APIs in the Power BI Service
Announcing Data Refresh APIs in the Power BI Service

Source: https://powerbi.microsoft.com/en-us/blog/announcing-data-refresh-apis-in-the-power-bi-service/           Today, I am excited to announce the general availability of new APIs for managing data refresh in the Power BI service. These new APIs will allow you to programmatically trigger data refreshes and READ MORE

20

Jun

Running BigDL Apache Spark Deep Learning Library on Microsoft Data Science Virtual Machine

This post is co-authored by Gopi Kumar at Microsoft and Sergey Ermolin at Intel.

Introduction

BigDL is a distributed deep learning library for Apache Spark. It has both Python and Scala interfaces and takes advantage of Spark-enabled distributed compute infrastructure, allowing users to write Deep Learning applications in a familiar native Spark context format. The Microsoft Data Science Virtual Machine (DSVM) is a customized VM image on Azure built specifically for doing data science and deep learning. It has many popular AI and data science tools pre-installed and pre-configured, helping jump-start the process of building intelligent apps that use predictive analytics and deep learning. DSVM is available on Windows Server 2012 and on Linux in both Ubuntu and CentOS7 editions.

This post shows how BigDL can be easily integrated into DSVM for Linux, and provides an example with LeNet5 character recognition implementation in Python. We first provide manual setup steps and then an automated mechanism to deploy BigDL on the DSVM with just a few clicks or commands. 

Provisioning DSVM

Before you start, you need to provision DSVM for Linux (Ubuntu) by visiting the product detail page and following the directions in the VM creation wizard:

 

When DSVM is

19

Jun

Microsoft R Server 9.1 on HDInsight is available!
Microsoft R Server 9.1 on HDInsight is available!

Source: https://blogs.msdn.microsoft.com/azuredatalake/2017/06/19/microsoft-r-server-9-1-on-hdinsight-is-available/   Today, we are excited to announce that Microsoft R Server 9.1 on Azure HDInsight is generally available. With this, we bring the power and innovation of our latest 9.1 release to the cloud on Spark 2.1 on HDInsight READ MORE

19

Jun

Loading a Trained Model Dynamically in an Azure ML Web Service

This post is authored by Ahmet Gyger, Program Manager at Microsoft.

We are very excited to announce the availability of a new module named “Load Trained Model”. As its name indicates, this module allows you to load a trained model (also referred to as an iLearner file) in Azure ML into an experiment at runtime. When used on a published predictive web service, in combination with web service parameters, it allows you to dynamically select a trained model for scoring.

Training multiple models using different datasets but with identical schema and identical training workflow is a very common scenario. (This article explains how this can be done using automation.) After you have obtained the models, you want to put them in operation. But you might not want to create individual web services for each trained model, particularly when the number of models are high, and the scoring workflow is identical except the models are different. You could accomplish that using PowerShell automation and web service endpoints as suggested in the aforementioned article, but it can still be cumbersome and difficult to manage.

With the new “Load Trained Model”, you can just create a single predictive web service and dynamically

19

Jun

Webinar Wednesday: Data Visualization Design Discussion Lab

Source: https://powerbi.microsoft.com/en-us/blog/webinar-wednesday-data-visualization-design-discussion-lab/           Take your dashboard and reports to the next level with a hands-on look at improving and updating your visual design in Wednesday’s webinar! Join Microsoft Design and Data Visualization Lead Miranda Li as READ MORE

19

Jun

Interactive R custom visuals support is here!
Interactive R custom visuals support is here!

Source: https://powerbi.microsoft.com/en-us/blog/interactive-r-custom-visuals-support-is-here/           R is a strong and popular language, enabling developers to create great analytics on data, as well as visualization. Developers can already create R custom visuals that plugs into Power BI reports, to READ MORE

16

Jun

Community Blog Highlights
Community Blog Highlights

Source: https://powerbi.microsoft.com/en-us/blog/community-blog-highlights-2/           Have you visited the Power BI Community blog lately? Along with rapidly growing forums, event listings, and ways to leave feedback, the Community blog is a platform for you to share ideas with READ MORE