What are the hot topics in Computer Science today?

Look here maybe?

If I went back to school, I'd probably try to do something in robotics, but I hate school, so it doesn't really matter.

From a poll around my school's CPSC lab: The Nintendo DS. Everyone's crowded around someone's poking at the screen.

In the field? I have no idea. There's so much to do. My school does a lot of stuff on machine learning and image analysis, that stuff is pretty cool.

network intrusion detection
human/computer interfacing
computer animation
robotics


blah, blah, blah

go to a big CS research school's website (like CMU, MIT, or Stanford), and see what research they are doing, that's basically the tip-top-hottest stuff.

parallel/distributed processing and OSes
stream processing (GPU stuff)

Computational Biology is huge right now, some 70% of the research at my Uni is dedicated to Comp-Bio; things like protein folding, DNA sequencing, I hate it all, but I can't swing a stick without hitting a computational biologist right now :/

Quote:

Original post by Symphonic Computational Biology is huge right now, some 70% of the research at my Uni is dedicated to Comp-Bio; things like protein folding, DNA sequencing, I hate it all, but I can't swing a stick without hitting a computational biologist right now :/

It's the same here...it's because that is where all the money is now.

lane detection: detecting the sides of roads from video images

Quote:

Original post by Lode lane detection: detecting the sides of roads from video images

Hey, I know a phd student that is working in that area...do you know someone who is doing that? Because if so I'm sure he would be interested in that other people are doing.

Quote:

Original post by Symphonic Computational Biology is huge right now, some 70% of the research at my Uni is dedicated to Comp-Bio; things like protein folding, DNA sequencing, I hate it all, but I can't swing a stick without hitting a computational biologist right now :/

Comp Bio mystifies me; it's what I'd call an application of computer science:

Protein folding - Using a computer to model the chemistry, mechanics and thermodynamics of protein.

DNA sequencing - That's a solved problem. Figuring out what the sequence means is what molecular biologist do.

But yeah, thanks for the input.

Quote:

Original post by jsgcdude DNA sequencing - That's a solved problem. Figuring out what the sequence means is what molecular biologist do.

Not quite. We've decoded existing sequences. Creating new sequences to produce proteins with certain geometries is quite difficult.

I am also interested in doing something like this. The problem is that most of the common stuff we hear, like protein folding, is already done. I mean, the guys at Stanford have hundreds of thousands of lines of FORTRAN code, and I got nothing. Would it be possible to compose a list of topics that are hot, but that we won't be idiots to work on. Like something that isn't already being studied to death by the big guns(MIT, CalTech, Stanford, etc).

Quote:

Original post by Sagar_Indurkhya I am also interested in doing something like this. The problem is that most of the common stuff we hear, like protein folding, is already done. I mean, the guys at Stanford have hundreds of thousands of lines of FORTRAN code, and I got nothing. Would it be possible to compose a list of topics that are hot, but that we won't be idiots to work on. Like something that isn't already being studied to death by the big guns(MIT, CalTech, Stanford, etc).

My understanding is most all molecular modeling tools are open source and freely available (google for 'tinker' 'molecular modeling' and throw some 'nih' in there). From what I've seen (tinker) your looking at some hardcore thermodynamic analysis and chemistry

When I contemplate molecular modeling I think about approximating atoms and molecules with a bunch of spheres attached by struts and springs; something you could do with ODE for example. But on the scale of an entire protein it'd be complicated to say the least.

Although I haven't done alot of research, one area I've not seen is modeling 'vicinal water'. Proteins are quite heavy, yet they are suspended in water. To float something must be less dense than its medium. One method of separating protiens is called 'salting out'. This is generally an undergraduate bio chem lab exercise. You increase the concentratin of salt in a protein solution. At certain OH- concentrations specific protiens will fall out of solution. The idea is that proteins are surrounded by crystaline like water structures that increase there volume and lower their density so they float. As salt is added OH ions rob water molecules from the vicinal water bubble around proteins causing them to 'sink'. If this seems unclear then you should remind yourself of the tri-polar nature of water and why ice floats.

My question about all this is how do the vicinal water structures guide the folding of proteins, and indeed, all other enzymatic activity?

So to me this is using a computer to model reality (space invaders, doom 3 etc..). Is it 'computer science'? I don't know. I suppose if researchers concern themselves with finding algorithms that optimize the models then sure, otherwise it seems like a really involved application of known algorithms and data structues.

"Bioinformatics" seems like such a buzzword; used to attract grant money and grad students.