Tuesday, February 15, 2005

Neuroinformatics

I've been reading about bioinformatics and there's an interesting subfield I've discovered: Neuroinformatics (related to computational neuroscience)http://www.neurovia.umn.edu/INC/Research/NI/Portal/ni_resources.php

This field is basically the bioinformatics of the brain and neurons. That is, applying computer science techniques (of which AI is just a part) tothe huge amount of brain data being generated. To link it, formalize it,make sense of it. A big part of neuroinformatics involves "The Human BrainProject."http://www.nimh.nih.gov/neuroinformatics/index.cfm

Think the human genome project but... harder.

Right now I'm doing bioinformatics/AI/Cognitive Science. That means, in mycase, that I'm primarily trying to make contributions to cognitive science, secondarily to AI, and third to biology. I am trying to avoid doing things with my AIs that are not innovative AI techniques, even if they might be best to solve the biological problem at hand, which is, inmy current case, protein folding. Further, I'm not all that interested in new AI techniques unless they are cognitively plausible either.However I really do care about solving this folding problem, even though it's third on my list. What makes me excited about neuroinformatics,though, is that, unlike protein folding, neuroscience is actually kind of a part of cognitive science. My AIs and theories would be contributing to our understanding of the mind while what those AIs think about is trying to contribute to our understanding of the brain. I think this could be good (or better, anyway) for my career, partially because
I'm more interested in Neuroscience than protein structure. And who knows how the knowledge I'd gain from having the brain as my subject matter will helpmy cogsci and AI level theories?I would be making AI that understands neuroscience. How cool is that?

Certainly this isn't something I'm going to start during this postdoc, but I want to think about my long-term plan. I like the idea of having a real domain, and bioinformatics is very hot. Perhaps I can segue into neuro once I take a faculty position. I want to keep this in mind as I look at schools to apply to next fall-- who is doing human brain project stuff with whom I could potentially collaborate? Also, if I fail to get a sufficiently good faculty offer, I could potentially look for aneuroinformatics postdoc.

No comments: