A Layperson’s Explanation of How Machine Learning in Google’s Algorithm Works
Digital Marketer – Podcaster – Jason Barnard | Ep. 360 Send It Rising Internet Marketing
Want to know how Google makes decisions?
Check out this clip! In this clip, Jason Barnard explains that Google’s algorithms are a “black box” and no one knows exactly what’s inside. Even the Google engineers do not know how the machine makes its decisions. Mathematical formulas, data and a goal are being fed to it on one side. Then, humans at Google take the results the machine has delivered and review them. They then feed it back into the machine as corrective data: basically showing the machine where it got things wrong, and where it got things right so it can adapt and improve future results.
00:00 Machine is a “black box”
00:23 Google is fed with mathematical formulas and data
00:53 Humans guide and influence the machine
01:19 A shoutout to the future generation
Transcript from: A Layperson’s Explanation of How Machine Learning in Google’s Algorithm Works
And I think the important thing that we need to remember is the machine isn’t just running out there all on its own. Right? You have a black box that nobody knows what’s inside. As you said earlier on, even the Google engineers don’t actually know how the machine is making its decisions. But you do have mathematical formulas and data being fed in one side. The black box of the machine figuring out how it’s going to arrive at the goal we gave it. And then human beings take the results that machine has given. And they say, “This is good. This is bad.”
And they feed that data back into the machine to say, (basically corrective data) like feeding back in. You did this one right, this one wrong. Now try to aim more for the right one. So we are guiding the machine, but we just don’t know how the machine works in the core of it. How it makes those decisions. So I think it’s important to remember that the machine isn’t just running wild. It has data coming in where humans can influence it. And it has data coming out where humans can judge its results and then ask it to correct itself.
I tell you someone 36,000 years from now is going to be watching this. Shout out to you, by the way. I know we sound very weird. Our language was very different back then. We were doing our best. That’s my message to you, person in the future.