Neuroinformatics
from the let's-all-work-together dept
I’ve been reading about “neural nets” for years, and remember playing around with a few neural net programs a long time ago. Despite lots of hype, not very much ever came from neural nets – and now some people are trying to change that. The problem with neural nets, some claim, is that most of the work is based on what we knew about the brain many years ago. Current neuroscientists and computer scientists don’t communicate enough. So, now people are talking about “neuroinformatics” to try to update the science behind neural nets.
Comments on “Neuroinformatics”
incorrect..
..”not very much ever came from neural nets”. totally, wildly, and sadly incorrect. enjoy playing the latest PC games? like making relativly clear international phone calls? feel safe and secure that Visa has the technology to recognize spending habits on your card that are fraudulent? shoot – are you glad the police and FBI can match the faces of known criminals with those of Super Bowl attendees? well – then you’ve directly benifited from neural nets. if you were to have said “no conscious machine ever come from neural nets, and that’s their greatest failure” you’d have been far, far more accurate.
Re: incorrect..
Not to mention all the business decisions that are influenced by risk analysis software that use neural nets, and all the cute little robots that can play soccer games, run races, etc.
Neural networks: Yay or nay?
While it’s true that artificial neural networks haven’t ushered in a brave new world a la Spielberg, they are useful tools for many concrete problems, especially pattern recognition and prediction. Profit-oriented companies have spent real money (and received return on) on neural networks to do things like predict which customers are most likely to desert, which products are defective, which companies have “cooked the books” and so forth.