November 13 2014

Even the very basic first steps into Artificial Intelligence fascinate me. From the weird unknown world of film where AI is perceived as something extraordinarily difficult to the very realistic use of if in our everyday life thought Siri or Google Now. At this moment I am taking a course in it at my local university and would like to tell you a little about what it is.

AI is the process of understanding of how people think on a fundamental level. Instead of trying to figure out how each individual neuron in your brain works, AI tries to generalized this by think computationally. An easy way of understanding this, is as The Wright Brothers, they didn't look at birds and try to model every single bone, muscle and feather; instead they tried to understand the general basic such as aerodynamics and how to get a similar effect using everyday materials.

Most of the time when you think you apply some kind of a computational scheme, for example if you are trying to find your car keys, you know that they are usually either in your pocket or on the table. So the next logical step would be to check your pocket, if you do not find anything then that means that they are on a table. This is done automatically by your brain without you realising most of the time. If we scale this problem up, to say going grocery shopping, then it become a lot more difficult: what products to buy, at which supermarket(better quality, closer location, cheaper prices), how to carry them home(on car or by foot, limiting the amount of groceries you are able to buy). This becomes such a complex task that we have to individually dedicate our time to thinking about this, it doesn't happen automatically(although sometime it does, such as a simple milk run). Most of what AI does is try to solve these similar problems using our modern computers. They have an input of the different variables and elements, such as possible groceries, possible stores, store information, etc; and then they calculate the most efficient and best way to shop. The more variables that are present the better the choice becomes but at the same time the more difficult it is to calculate. This kind of AI is used in many Voice recognition systems, they have a set amount of words and phrases that they know and then they try to listen for when you tell them one of them.

AI has an incredibly wide application in our modern world, it could be used to model efficient food transportation routes, it could provide the least congested city street using smart traffic lights, it could provide incredible realistic gaming scenarios and it already provide fairly good voice recognition.

This is just a little peek at what AI is, if you are interested read the book called 'Thinking as Computation' it covers the examples I have here as well as many others.

[email protected]
Toronto, Canada

PS If you are by any change looking for a developer starting May 2015 please don't hesitate to contact me. I am open to working anywhere in the world!

comments powered by Disqus