Wednesday, November 01, 2006

The Future of Computer Science

Ever since I've been exposed to machine learning, I've started wondering how the field of computer science would look like say 50 years down the road.

Computer science started off as a bastard lovechild of electrical engineering and mathematics, but it's now become so much more. The development of computer science has greatly influenced the development of a number of other fields, such as operations research, information science, computational biology, computational linguistics, and so forth.

One can roughly categorize the majority of machine learning research as glorified computational statistics. Professor Caruana, whose class I'm TAing for, predicted that the fields currently known as machine learning, data mining, and computational statistics, would likely merge into a single big field at some point down the road.

So it begs the question of whether or not that as computer science grows and grows, if its sub-areas will fracture off into their own fields. I highly doubt that computer science will go the way of mathematics, given that most of it is empirically driven. But it probably won't go the way of electrical engineering either, since computational problems are something that's quite different from other kinds of engineering.

Now that we've blazed triumphantly into the information age, our datasets have likewise become gigantic. As such, statistics has been playing a more prominent role. Almost every computational science out there today has become some kind of statistical science.

But then there are those folks who don't deal with experiments as much, but rather more with engineering. These areas include systems, networking, security, graphics, programming languages, compilers, and databases to name a few. While the information age does present other problems of scale that they must address, the actual abundance of data is probably not as critical.

I'm obviously not able to arrive at any conclusions at the moment. The most likely scenario is that future advancements will change the landscape of what we understand so drastically that we're simply unable to predict any trends beyond 5-10 years.

Incidentally, I'm posting about this mostly because I just recently read a NY Times article titled Computing, 2016 where prominent computer scientists, including Cornell's very own Jon Kleinberg, commented on the continuing impact of computer science on other fields. I guess it rekindled my need to ponder aimlessly.

No comments: