Neural Nets

Created by Jijith Nadumuri at 28 Apr 2010 17:16 and updated at 29 Apr 2010 05:50

Neural Nets are a promising way of programming that complements the conventional way of programming. There is a lot of hype around neural net programming (not to be confused with the neural nets in our brain, which is a biological system). Especially programmers who have not worked with neural nets in their life time consider it as sort of a magic box that can solve any complex problem. They also feel that these kinds of programs are artificial intelligence programs that can become somehow 'self aware'! These are far fetched. Neural nets do promise solutions where conventional programming struggle but wrong application of neural nets will cause much frustration. If you are to use neural net approach, first you have to study the problem and understand if the problem is suitable to be tackled by a neural net solution. From my experience with different types of neural nets, what I concluded is that every neural network design try to smoothen the binary-ness (YES OR NO) of the conventional computing and allow computers to experience 'grayness' where things are not a simple yes or no. For this, techniques like signal summation, integration over time etc are incorporated into neural net systems.

In the articles listed below, I tried to list down some of the concepts of neural nets. However it is to be understood that this subject is still evolving.

  1. The Hype Over Neural Nets - Yea, They are not yet self aware!
  2. Introduction To Neural Nets - Keep your expectations low
  3. Human Brain and Artificial Neural Nets - We copied the design from the brain
  4. Neurons - The building block of any neural net, biological or artificial
  5. The Processing Element - The new avatar of a neuron
  6. Working of an Artificial Neural Net - How it works?
  7. Training a Neural Net - Yea, they need training and coaching!
  8. Neural Net and Traditional Computing - What make it special?

Share:- Facebook

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License