Studying AI as an Undergraduate

rob Uncategorized 0 Comments

One of the most well-known, and perhaps the most exciting, ares of research within computer science is Artificial Intelligence.  From video games like Mass Effect to movies like Terminator to books like I, Robot, AI pervades our popular culture and captivates the imagination.

Probably one of the cooler AI I’ve ever heard about.

What most people don’t realize is that AI research and techniques are already deployed in every day life.  When you google something like “obama  romney” and Google comes back with “new for Obama Romney,” Google has used AI algorithms to determine that “Obama Romney” is something correlated with news, and also to determine what search results qualify as news or not!  Social networking sites and advertising companies are also big on AI technology because of it’s ability to interpret vast amount of data.  Training in this discipline is a hot commodity nowadays.

All of this factored into my decision to take CSCI 460, Introduction to Artificial Intelligence.  Every CS major has some units to fill with “Technical Electives”, which are generally classes that are focused in a specific discipline within the field (for example, another technical elective I’m taking is EE450- Computer Networks). It is up to the student which topics interest them most, and thus, I found myself in an AI course.

 

I have spent a lot of time in the past week on my current AI project, and man, is it a doozy.  We’re working with an AI structure known as a “neural network”.  And no, it’s not a model of a brain.  That would be awesome, and the inspiration for neural networks was derived from the brain, but it’s much different.

 

Basically, it is a bunch of interconnected units called perceptrons, all of which add up a group of input values to determine a single output value.  The input values are multiplied by numbers that are “learned” (changed to improve accuracy, this is what makes it AI!) and the perceptrons can be connected in all sorts of crazy ways to arrive at the correct output.

 

An example of a “one hidden layer, three output layer” neural network. And it get waaay more complicated than even this!

It is definitely fascinating to be working at this level of complexity.  The concepts in this course are by no means simple.  In fact, the professor himself is a bit of a pioneer in the AI field (a perk of attending a research university)!  Read more about him here.

 

While AI is definitely interesting, I’m excited to see what other tech electives I will eventually be taking.  As I mentioned, I’m taking Computer Networks as well, in which we have already gone over how the entire internet works.  I’ve even put knowledge from that class to work on a summer internship application to Google.  Other tech electives range from the super practical (Databases) to the super fun (Robotics).

 

Well, I’ve got to get back to my super-intelligent neural network.  I’d like to apologize now if this baby ends up taking over the world.

Want to learn more? Here's the best place to ask: