Machine learning has recently emerged from being a relatively obscure part of the enterprise computing world to becoming a hot topic. At our company, ManyWorlds, we’ve actually been involved with machine learning-based R&D for quite some time—over ten years, in fact. Our CEO, Steven Flinn even wrote a book a few years ago, The Learning Layer: Building the Next Level of Intellect in Your Organization, which outlined the inevitability of a machine learning-based layer in enterprise IT architectures and the amazing opportunities for business improvement that this layer would engender.
What is Machine Learning?
The term “machine learning” covers a lot of territory, so it is helpful to put it in perspective with related terminology that is also becoming increasingly topical. Machine learning can be thought of as a particular branch of predictive analytics. The distinction is that predictive analytics can mean human analysts using various mathematical models to make predictions from data, typically big data. Machine learning also is about making predictions from data, particularly big data, but in this case the system itself automatically and continuously learns from data and gets better and better with respect to what it is designed to predict. Now, there are also many different applications of machine learning. We specialize in one particular type of application, which can best be termed “anticipatory computing.” Anticipatory computing refers to systems that are focused on anticipating what is most relevant to users and acting accordingly, rather than only just reacting to users. To do this requires making high quality predictions of the user’s interests, preferences, and even expertise from social big data—that is, without needing to be explicitly told by the user.
The Opportunity to Increase Enterprise Growth & Productivity
Anticipatory computing is the answer to the problem of users drowning in information, much of it personally irrelevant at any given time. In the consumer world, major e-commerce and social platforms have addressed this problem some time ago with various forms of recommendation engines that surface personally relevant suggestions. With the increasing importance of collaborative and social systems in the enterprise and the large amounts of additional information they generate, the need for maintaining the best possible relevant-to-personally irrelevant information ratios has become an imperative. That’s where our Synxi-brand personalized discovery apps come in. Synxi integrates with systems such as SharePoint and Yammer and continuously surface suggestions of content that is both contextually relevant to what the user is currently doing and that is also personally relevant to the particular user. The apps can even recommend other users who are inferred to have contextually relevant expertise to a particular item of SharePoint or Yammer content that the user is viewing. And the apps can automatically learn to cross-contextualize across platforms, so that SharePoint content that is relevant to a specific Yammer post is automatically surfaced. Microsoft, of course, has also recently announced new capabilities specifically for Office365, Office Graph and its first app, Delve, that allow you to explore and discover content from your organization’s social graph.
Beginning the Era of Adaptation
Our and Microsoft’s products only scratch the surface of what is possible with respect to machine learning opportunities in the enterprise. The newly announced capability of Azure Machine Learning creates opportunities for Microsoft ISV partners to integrate machine learning capabilities within their products, and for SI partners to integrate ISV and Microsoft machine learning-based applications with other applications. For example, there is no reason going forward that processes and workflow shouldn’t automatically adapt to personal requirements. The same is true of educational and eLearning applications.
There is also an opportunity to apply machine learning to applications that are associated with the “Internet of Things,” such as automatically learning and acting on the vast amounts of process control-type big data that are generated by sensors and other devices that are prevalent in many industries. For example at WPC 14, Satya Nadella’s keynote included an example of ThyssenKrupp, an elevator company who manage 1.1 million elevators globally. For them, reliability is critical. The ability to predict when maintenance will be required is a source of competitive advantage for ThyssenKrupp and essential to the people who live and work in buildings, such as in the newly constructed 102 story One World Trade Center in New York City that will carry 3.5 million people a year.
Most generally, it now seems pretty inevitable that we are entering an “era of adaptation” in the enterprise driven by the possibilities that machine learning promises. It’s an era in which most systems automatically sense and learn from their environment, including their users, and beneficially adapt accordingly. As Microsoft partners we are collectively well positioned to make this inevitable future a reality for our customers sooner rather than later.
Naomi Moneypenny is CTO at Synxi (a ManyWorlds brand) where she leads the development team for the adaptive recommendations and machine learning engine for SharePoint, Yammer and other social systems. She holds over 30 patents in the area of adaptive systems and was named a Top 25 SharePoint Influencer in 2014. Naomi is a frequent speaker at international events on enterprise social and collaboration, anticipatory computing and business growth from innovative technology. Follow Naomi at @nmoneypenny and read her latest thinking at NaomiMoneypenny.com.