An AI Model for Industrial Operations and Maintenance (Video)
Predictive analysis can deliver more efficient industrial operations and maintenance but that does not eliminate the need for human experience and knowledge transfer, according to Serious Insights founder & principal analyst Daniel W. Rasmus.
“The way to get to the smaller window is predictive analytics. How do we get to the point that we can predict the most likely time for a machine failure, and then on top of that, either advise and/or predict when we're going to actually get to the point that it's the best time to make the repair?” said Rasmus, who presented a talk on KM in the age of AI and IoT at KMWorld 2018.
“So, we want to get to the repair before the piece fails, before the piece of the machine fails, and when it's least impactful on the organization. And so, these are all kinds of things that have to happen from the knowledge standpoint, from figuring out the best use cases to understanding the data, to deploying the sensors and what sensors do we use, and what's that network look like--how does that integrate with all the other data that we have around the organization? How do we format it? How do we make it clean? How do we get it to the point that it's useful?”
There is the requirement to get to the sources of data, he noted, “so we can have the predictive models, and the predictive models themselves have feedback loops that require human feedback, right? Because you're going to have this data come in, you're going to have the machine make some predictions, and sometimes they're going to be wrong. Many times early on they're going to be wrong, and you have to go, ‘Why are they wrong?’ and then collaborate.”
The processes involved in actually putting AI and machine learning to work are very complex and requires a great deal of human experience, said Rasmus. “This is complex stuff. This is not something that you're going to go hire somebody out of Stanford or MIT or Carnegie Mellon and just have them walk into your facility and go, "We have this really big factory, and we got all this data, and we're going to have even more data because we're going to put sensors all over stuff and we don't even know how those work, and since you have a Ph.D. in Analytics and Computer Science, we want you to fix that.
Not going to happen. I mean, how many new graduates out of MIT, Carnegie Mellon, or Stanford that you can put in any job and expect them to come in and go, ‘Yeah, I know how to ... that's easy. I can do that.’ It's experience. It's knowledge transfer.”
Learn more at KMWorld 2019, coming to Washington, DC, Nov. 5-7.
Watch the complete video of this presentation, Rethinking KM for an Age of AI & IoT, in the KMWorld Conference 2018 Video Portal.
Many speakers have made their presentations available at www.kmworld.com/Conference/2018/Presentations.aspx.