Calculus and Computer Science

A place to discuss the science of computers and programs, from algorithms to computability.

Formal proofs preferred.

Moderators: phlip, Moderators General, Prelates

Posts: 4
Joined: Wed Mar 27, 2013 7:36 pm UTC

Calculus and Computer Science

Postby edsteen » Fri Dec 05, 2014 7:57 pm UTC

I am a High School Senior, planning on going to college to study Computer Science and Neuroscience. I want to study AI. I haven't done a ton of super complicated coding, just enough to get my robotics team by with a nicely driving robot, and enough to know I really enjoy it. Unfortunately, I am currently taking Differential Calculus in school and absolutely hating it. I have studied my butt off for tests, and failed them. I am trying as hard as I can, and just not doing anywhere near what I need. So my question is, how much calculus will I need for AI? If I will need a lot, I want to start considering alternate careers now, before I get too far in.

User avatar
Posts: 1710
Joined: Wed Apr 21, 2010 12:29 pm UTC
Location: Netherlands

Re: Calculus and Computer Science

Postby Jplus » Sat Dec 06, 2014 2:56 pm UTC

In my experience (AI Master) you can study AI without touching any calculus. AI has many aspects and calculus isn't central to any of them, although in some cases it is relevant (see below). Study programs differ, though, so you should probably check with the specific AI program that you are considering. Usually there is considerable room to make your own choices within the program.

Note that calculus can be really useful in robotics. For example, if you have a robot arm that needs to make smooth, precise movements, you can use both integrals and differentials to damp vibrations.

Also, the fact that calculus appears to not be your cup of tea right now does not have to mean that you'll hate it and fail at it forever. Perhaps it just isn't taught in a way that works for you. Perhaps you first need a convincing use case that helps you understand what it is all about.

At any rate, please don't let yourself be discouraged from AI just because of calculus. I wish you much luck!
"There are only two hard problems in computer science: cache coherence, naming things, and off-by-one errors." (Phil Karlton and Leon Bambrick)

coding and xkcd combined


Posts: 527
Joined: Thu Feb 21, 2008 12:16 am UTC

Re: Calculus and Computer Science

Postby wumpus » Sun Dec 07, 2014 6:09 pm UTC

Generally speaking, math is the language of the sciences. There are two issues here:
1. Getting by the gateway that colleges put freshmen through for a background for future study.
2. Actually doing AI.

My guess is that any reasonably hard-science BS degree will require taking calculus, and then plenty of courses that use calc as a pre-req. This may or may not require much calc (typically in the form of differential equations. Nature *loves* differential equations). You might also need to ask someone in Neuroscience, because I can't imagine that it ignores a math background.

I'd also second looking into learning differential calculus any other way. My memory was that differential was easy (simple memorization of a few rules) while integral calculus was the hard part (recognizing all those rules already applied, then breaking them up and applying them).

As far a robotics goes, I don't think you can spell control theory without learning the calculus behind it (all the robotics depend on several derivatives worth of motion all working together). On the other hand, I'm not sure you often have to bang out the differential equations in practice. In similar things (based on the same theories), DSP programming can simply use FFTs all day long (which are based on a nasty looking integral equation) without ever manually doing the conversions (which would be looked up anyway, not solved via the integral).

User avatar
Posts: 1378
Joined: Wed Mar 26, 2008 1:16 pm UTC

Re: Calculus and Computer Science

Postby Cleverbeans » Sun Dec 07, 2014 7:35 pm UTC

I have only taken one machine learning course but we definitely used calculus for the gradient descent algorithm. That being said it was only one topic in the course and most of the other problems we solved primarily used statistics and linear algebra. You won't need calculus that much but I expect will be required from time to time. In the end you can customize your learning to avoid it, but I don't think you'll get out of doing at least a years worth of undergraduate calculus.
"Labor is prior to, and independent of, capital. Capital is only the fruit of labor, and could never have existed if labor had not first existed. Labor is the superior of capital, and deserves much the higher consideration." - Abraham Lincoln

Posts: 244
Joined: Fri Jun 13, 2014 1:33 am UTC

Re: Calculus and Computer Science

Postby schapel » Wed Dec 10, 2014 7:27 pm UTC

For most applications, you need to know only the basics of calculus, if that. So I wouldn't worry about failing calculus.

But lots of math is needed in many areas of computer science. You'll need to have a good grasp of linear algebra to do some AI work. And working with probabilities is going to be important if you specialize in AI. If you study computer science (even not working towards AI), you'll need to take classes in discrete mathematics.

I think the bigger issue is whether you're having trouble just with calculus, or with math in general, or with continuous mathematics as opposed to discrete mathematics. If the problem is just calculus, it's probably not a huge issue. If you're going to have trouble in all your advanced math classes, you can still be a decent programmer, but going to college and learning AI programming will be very difficult.

Return to “Computer Science”

Who is online

Users browsing this forum: No registered users and 4 guests