Computational Science and Engineering Track Overview Video

The following is a transcript of the video clip "Computational Science and Engineering Track Overview Video":

Daniel Sutton (Senior): So I'm DJ Sutton, and ...

Richard Byrd (Professor): Richard Byrd.

Sutton: Talking about Computational Science and Engineering. So first off could you tell me, in general, what makes this track different from something like General Computing?

Byrd: I think Computational Science and Engineering, it's a subject that's driven by certain types of applications, and these are applications where computing comes into contact with the physical world, and it's greatest end of computation, applications involving floating point numbers. So basically it's used in computing to solve problems that tend to be mathematical in nature -- but not necessarily, you know, graphics is a big part of it, so but computing interacting with the physical world and serving those kind of applications.

Sutton: Could you tell us a little about the specialized foundation courses that you take here, Computational Science and Engineering.

Byrd: Yeah. So actually it's kind of interesting; there's a couple of things. The things you see in the foundation courses, these are things that basically you need to know to do this effectively. Numerical computing is computing with the floating point numbers, that's very important. One thing Computational Science and Engineering is you're going to be pushing computers to the frontiers of what they can do. So you're going to be using the biggest computers, the fastest computers, and that's why we have High-Performance Scientific Computing together. Serious scientific computing is done very often on very large computers, a lot of computers working together to make that happen. You're talking about very large software -- need to make that "go".

If you look at this group, the core group, it's interesting, they're kind of divided into two groups. Some of these things are things that help you work more effectively in Computational Science and Engineering. I would say that's Operating Systems, User-Centered Design -- you're going to be designing very likely software to do hard things, and so making it easy to use is very important. Databases, yeah you're going to be writing software to handle a lot of data. Performance modeling, you want to make sure that the computer is doing the right thing, doing what you think it's doing to solve your problem. Some of these other things are more applications. Graphics, you're going to be doing things, you're going to be describing the real world, you're presenting the real world, and so presenting a bunch of numbers is not as helpful as presenting a good picture. So computer graphics is very important.

However, some of these other things are things that you can do with Computational Science and Engineering. For example, game programming. Games, basically they present a picture of the real world that has to be sort of realistic; to make that happen right, you need to do essentially numerical programming, graphics, same thing for animation. Chaotic dynamics is part of the real world that operates in kind of an unpredictable, chaotic fashion. Machine learning very intensively uses numerical computing, because basically machine learning is essentially optimization now, at the highest level. For example, machine translation, is basically making choices by optimizing over a lot of numerical values. So some of these things are to help you do your job better, and some of these courses will show you what Computational Science and Engineering can do.

Sutton: What would you say are some of the interesting current research topics?

Byrd: So one of the most interesting ones is machine learning, I'd say right now. It's amazing, people used to think machine learning was sort of coming up with a verbal description of the world and manipulating that, and now we're seeing it's really all about interconnections and numerical weights, and that's how we translate things now; I find that very interesting. Modeling, physical modeling of the world, people at NCAR are modeling the atmosphere. OK so modeling in the real world, in all aspects, so like at NCAR, controlling spacecrafts is a very important one. Computational biology is another one.

Sutton: And you mentioned math is important for this track. What other sciences or areas of engineering studies would you say people in this track should really have experience in?

Byrd: Math is absolutely essential, and you should understand, it would be good to be very good with some part of science because you're going to be interacting with some part of science or engineering. So certainly physics is very natural. You're very likely, if you work in this area, you're going to be interacting with engineers and scientists and you got to be able to talk to these people, talk their language. So physics is very important; chemistry, in many cases, is very important; geology would be very important, there are a lot of jobs, looking for oil, for example, geophysical modeling.

Sutton: OK and that's probably my last question, what kind of jobs could someone who does this track expect to get in the industry?

Byrd: So I would say this is basically, the jobs are primarily in science and engineering. So for example, high-tech companies like, or a lot of people are working at places like Ball Aerospace, Lockheed Martin, the national laboratories, NCAR here in Boulder, Sandia laboratories, Argonne laboratories, people that are out working on the frontiers of science. Also the oil companies, they all use, a lot of them, scientific computing.

Sutton: Alright, anything else you want to add, or any questions from the audience? Alright, thank you.

Byrd: OK.

Transcript provided by Erik Silkensen.