Machine Learning – What is it & Why Does it Matter : Ep 11

About the Episode

In this Quick Byte episode, Chief Strategy Officer Brian Haines explains machine learning and its connection to artificial intelligence in this episode. He explains how machine learning is used in building operations and how it evolves with more data over time.

Watch the Episode

Episode Transcription

Brian Haines 0:14

In today’s Quick Byte, we’re going to talk about machine learning. Machine learning is a field of study in artificial intelligence concerned with the development and study of statistical algorithms that can effectively generalize and thus perform tasks without explicit instructions.


Brian Haines 0:32

Now that sounds a little bit more complex than my previous quick byte on artificial intelligence. Although these two components are highly linked, because machine learning is an actual field of study within artificial intelligence. And let me tell you, you’ll often see them reference with one another artificial intelligence and machine learning.


Brian Haines 0:54

So let me give you an example of what machine learning is as it relates to facility operations. And this example is around occupancy utilization. So machine learning is really an algorithm that learns over time, very similar to a human. When you’re in elementary school or you’re young and you’re in first grade, you don’t have a lot of knowledge. But as you go through time, and you go through high school, and you maybe go on to college, and you go on to your career, you develop a large dataset of knowledge. So machine learning is really a part of artificial intelligence, where it’s learning over time.


Brian Haines 1:35

And I’ll give you an example of how that could be applied. So when we’re looking at utilization, let’s say, I’m going to apply a machine learning algorithm to how well we’re utilized within one of our facilities. And I start on a Friday, and that’s the only data that the machine learning algorithm has available to it. And I have a bunch of meetings in the office on Friday. Now, most of you know, in the hybrid workplace environment of today, Friday is the least day used. So in this example, Friday is a day, that looks really heavily utilized, because we’ve had a bunch of meetings, which is highly unusual. So AI and machine learning is going to tell me if I ask it, what’s the most utilized day of the week, it’s going to tell me Friday, because that’s the only information it has available to it. As time goes on, and we go through the weekend, we come to Monday, and we start getting back to a more normal utilization pattern, Monday through Friday, most of us know that when you look at a lot of the data that’s being tracked on utilization globally, Tuesday, Wednesday, Thursdays are actually the most utilized days of the work week, for the most part in the hybrid workplace environment. With Tuesday being the most often utilized day and then Monday and Friday being the least. So going back to that first question, machine learning if it only had that one day of Friday, it would believe that Friday would always be our most utilized day of the week. But as this algorithm learns, and it starts to look at ongoing weeks, Monday through Friday, Monday through Friday, and it sees that more expected curve, it’s going to start to understand that that first Friday was probably an outlier. And that based upon the data that it’s analyzing, Tuesday, Wednesday, Thursday, are most likely going to be that day of the week, where utilization is the highest.


Brian Haines 3:35

So there you are, that’s machine learning. It’s not as difficult as it seems. It’s really the way an artificial intelligence algorithm can learn and get more refined based upon larger datasets over time.

FM:Systems in the news
See our upcoming live podcasts
Check out our latest blogs