July 2017
Volume 32 Number 7
[Editor's Note]
Machine Learning Goes Mainstream
By Michael Desmond | July 2017
.jpg "Michael Desmond")
At the Microsoft Build 2017 conference in Seattle in May, Microsoft detailed its deep investments in artificial intelligence (AI), showing how solutions like Cognitive Services and Azure Machine Learning can be leveraged to change the way people work, live and interact. At the core of this push is machine learning—the science of applying data to enable systems to make increasingly better decisions over time.
Machine learning is addressing prominent challenges, from financial fraud detection to manufacturing quality control to autonomous cars. This month in MSDN Magazine, we dedicate four articles to the topic of machine learning and explore how developers can employ it in their software development.
Hiren Patel leads off the coverage with “Cognition at Scale with U-SQL on ADLA,” which dives into the U-SQL Studio IDE to show how the Microsoft Azure Data Lake Analytics distributed engine can be used to train and score models at scale.
James McCaffrey then provides a look at the Microsoft Cognitive Toolkit (CNTK) v2.0 Library, which lets developers create machine learning prediction models based on deep neural networks—technology that’s at the forefront of AI efforts such as Cortana and self-driving automobiles.
“Doing Data Science and AI with SQL Server” by Wee Hyong Tok shows how SQL Server 2016 is bursting with intelligence capabilities that allow developers to run R code right where the data is. Tok walks through building an intelligent application and shows how the R code developed by data scientists can be used in database projects.
Finally, “Scale Applications with Azure Redis and Machine Learning” presents design best practices and code examples for implementing the Azure Redis Cache and tuning the performance of ASP.NET MVC applications to optimize cache hit ratio. Stefano Tempesta shows how smart algorithms processed by machine learning can help reduce cache “miss rate.”
‘A Must-Have Skill’
Our coverage this month reflects the growing importance of machine learning in software development—and not just in specialized environments. As McCaffrey notes, machine learning is at the foundation of deep learning techniques and AI development. It’s also coupled tightly with Big Data and Internet of Things development.
“There’s a general consensus among my senior dev colleagues that having at least basic machine learning knowledge is quickly becoming a must-have skill with regard to career growth,” McCaffrey says, adding, “Our HR people at Microsoft say that developers with machine learning skills are among the most requested, and most difficult to find.”
That value is reflected in the growing membership of Microsoft’s internal machine learning community, which stands at more than 5,500 developers, data scientists, architects, consultants and others since its formation in 2014. Alex Blanton is a senior program manager in the Data Group at Microsoft and community manager of the internal machine learning community at Microsoft. He says that more than 3,000 people attended the group’s latest Machine Learning, Analytics and Data Science (MLADS) conference in Redmond, which had presentations covering topics like Azure Data Lake, deep learning tools and techniques, Microsoft Bot Framework, security analytics, and more.
The group also hosts frequent online and in-person talks and events, semi-annual machine learning competitions, and an active discussion list used by the community to pose questions and share knowledge across the company. These engagements offer an opportunity to “discover pockets of excellence” in the community, Blanton says, and explore topics as they emerge in the fast-moving machine learning space.
“We pay some attention to connecting machine learning to other domains—for example, an upcoming talk is from a designer talking about how AI and machine learning can impact design,” Blanton says.
This issue is just scratching the surface of what developers can expect from Microsoft in the realm of machine learning and AI. I expect we’ll see a lot more on these topics in the months and years to come.
Michael Desmond is the Editor-in-Chief of MSDN Magazine.