Breaking down the BI Buzz
In a previous blog, our esteemed colleague Mikko Muurinen discussed in broad strokes the ideas of Artificial Intelligence and Robotics in the field of Business Intelligence and this got us thinking about the confusion surrounding these ideas. So, we thought we would take a minute and help clear up the fog surrounding the 3 biggest buzz words in the BI field, letting you to move forward with a clear base understanding of these ideas and their principals.
Let’s start with the easier ones:
API Application Program Interface
Starting at the buzz word to beat all buzz worlds in our tech world these days: APIs or Application Program Interfaces.
Without a doubt you have heard this word before but what does it mean/do? Put simply, any API is a set of rules that allow programs to talk to each other. Think of it as a universal translator. For example, if you want to use Amazon Web Services or similar products to help improve your own systems, how does your system or Amazon’s talk to each other? Sure, you could go line by line and learn their exact data specifications to send exactly what their system wants exactly how they want it, but that would make the cost of doing so in work hours alone an unviable option, not to mention Amazon would never want to release this kind of detail information on their systems, and lets just hope you don’t make any mistakes along the way.
So how do you communicate? The answer: API’s. This simple set of rules means that as long as both you and Amazon follow them, your systems will be able to talk to each other and you both can benefit massively in a multitude of different ways.
What we would pay to see the mental image that pops into most people’s heads when they hear these words! As odd as it sounds, Data Mining is actually a rather simple idea at its core. All it is at its heart is a way of looking for patterns in data that is not immediately obvious. Think of it like a Monet painting. If you’re close to it you can’t see the forest through the trees, it is all just a jumble of colours. Take a step back and the patterns immerge and next thing you know you’ve got a work of art!
Now take that painting and make it the world’s biggest Excel spreadsheet, terrifying I know! There is just no way a human could step far enough back and still see the individual cells well enough to make out anything useful, but a computer can. Algorithms are written to do just this, they provide a computer with the ability to compare and contrast vast quantities of data to turn our excel sheet into a work of art!
And now for the big one:
I know I know, we have talked about this before but stick with me! Trust me, A.I. is much less complicated to understand than you might think at its core despite what you may believe. Of course, please understand this is all a massive oversimplification but this theory stays the same.
One of the basis of A.I. is something called a neural network. Basically, these are layers of decision making that lay on top of one another like a sandwich, and each layer has a specific task, and is made up of vast numbers of dials for each question kind of like a giant switchboard used by musicians.
So let’s say we want to use this computer to decide where it should route an incoming call to your company. Our first layer could be set up to determine if this number has ever called before and gather all the data about that last call, the next layer could look at that information and decide if the last call was a positive one or not, the layer after that could determine if this call should be moved up to a higher queue position or not based on its history, and so on and so on and so on.
Now I know what you’re thinking, how does this A.I. know what is correct or not? How does it know the right thing to do? And the answer is very much like humans, practice makes perfect! You see an A.I. is really two parts, the neural network with all the layers filled with dials, and a little robot turning those dials. We take million upon millions of example situations and feed it into this neural network and tell that little robot we want this kind of result and let it get to work.
It will turn thousands and thousands of those little dials, and each time improving the networks success rate just a little, and over the course of those millions upon millions of tries it finds the right combination of dial turns to go from getting 1/100 right to getting 99/100 right. And the best part is the more you use it the better it gets just for you! Every single time it tries it gets better, every single time. It actually encourages you to give it more to do!
If you’re a little taken back by the possibilities this could allow, you have the right idea! Imagine what this could do for even the most basic tasks, not to mention the more complex ones. It staggers the mind.
Thank you so much for taking the time to let us here at Benemen help all of our understanding grow together! We honestly love doing this and look forward to next time!