Author: Adrian Perez

A Fun Event

Last Friday, the AI team had a fantastic time at the social (those Fruit Roll-Ups helped out a lot). Our team was able to get to work on coding the different projects with new faces to come by and ask us the hard hitting questions. For this blog, I thought it would be good to answer the question we got asked several times at the event; “Where do you start in AI?”

AI can seem pretty daunting at first. We all hear whispers of the difficulty and abstractness of the field and for most, it’s this elusive idea that only the brightest can access. I’m here to say that really isn’t the case. There is a somewhat steep learning curve, yes, but it’s not so steep that everyone will fall off the curve. These are my top 3 tips for getting your foot in the door.

Tip #1: Don’t Be Scared

Admittedly, AI is not a particularly easy field at first. Much of it contains jargon and formulae that will not make sense at first. I remember reading my first paper and after the first page, feeling like my brain had melted. The best thing to do is to not let it discourage you and take it at one step at a time. You are not any less intelligent if you don’t know what “categorical cross-entropy” is for example; you just need to spend the time to figure out the lay of the land. The same thing with those absurdly long equations. At first glance it’s a wall of incoherent symbols, but really, it’s just some mathematicians representation of an easy concept. Don’t let these things scare you away! If AI is really something you want to pursue, just take the extra time to understand the language the field speaks.

Tip #2: Find Somebody Who Knows Something

Something that I’ve found is that, even though there are lots of resources available online, there are always those questions that are never answered online. For example, there was no clear answer on StackOverflow for when I needed to understand how exactly data is transformed between the layers of a neural network. I understood the theory, but the actual implementation was hard to come by. Luckily for me, I was able to befriend people who knew way more than me and ask them those nitty gritty technical questions after I read a really complicated paper. Seek out people on campus or via email and ask them questions; you’ll find yourself getting a better understanding when it comes from a peer.

Tip #3: Just Do It

The best thing you can do, in my opinion, is to just dive in. Too many people say, “I’ll get to it at some point” and never do. The fastest way to learn is to just start with a project. It doesn’t need to be extravagant either. Just something to get you familiar with the field. Once you do a few of those, then I’d say increase the difficulty of the problem; otherwise, you might defeat yourself before you even have the chance to succeed. TowardsDataScience, Medium, and Hackernoon are all great places to find hands-on projects if you’re not in something like SparkDev.

Until Next Time

Go out there and do something cool in AI! If you ever have questions, the AI team is always happy to answer. Cheers!