Interest in machine learning hasn't peaked yet. It will be one of the key technologies for wringing value from data generated by the industrial internet of things (IIoT). We are starting to generate so much data, so frequently, from so many sources. Automated learning algorithms will become critical to identify important patterns in a timely way.
According to angel.co, there are now 916 machine learning
startups. In May of this year, there were just - ha, just - 624. (That's a 47% increase to justify my headline). To be honest, many of those startups are not very applicable to industrial applications. For example, some are using algorithms to improve ad targeting, or reduce credit fraud. But plenty of others have potential for industry - you just have to sort the wheat from the chaff...
My recent
research report, Machine Learning and the Industrial Internet of Things, highlights the potential for industrial companies. The report includes a quick overview of what exactly machine learning is, and includes profiles of some of the emerging and established technology platforms. Beyond that, I also profile a number of vendors in what will be the first in a new wave of machine learning applications for industry - predictive and prescriptive maintenance. Unfortunately the full report is only available to subscribers, or for cold hard cash. But, if you're only interested in the section that defines machine learning, you can find most of that
in this post, and a bit more
here too.
Beyond that, there are other signs of warming too:
- Google and Microsoft both open sourced part of their machine learning portfolio. But, as this article in Wired notes, one of the reasons they can be so magnanimous is because most of the value lies in the data, not the code.
- And, to back up the assertion that predictive maintenance is the beachhead, Microsoft just announced a predictive maintenance solution based on the Azure IoT Suite. Alas, too late to be included in my research report, but ya gotta draw the line somewhere...