(A) In this book "Introduction to Neural Network using Matlab 6.0 – S. N. Sivanandam, S. N Deepa" http://books.google.ca/books?id=jJTN8RPgyXgC&printsec=frontcover#v=onepage&q=genetic&f=false, it is mentioned that Hebb is a supervised learning algorithm in Page21. Later, under Learning Techniques, Chapter 3, Section 3.3 Hebbian is categorized under Unsupervised learning. This is confusing !!
(B) Again in this impelementation NN Hebb Learning, there is a target vector and the weight is updated by cross-multiplying the transpose of the input with the target (Did not quite understand).
(C) In matlab
learnh calculates the weight change dW for a given neuron from the neuron's input P, output A, and learning rate LR according to the Hebb learning rule:
dw = lr*a*p'
This equation shows that Hebb is an unsupervised algorithm
So, my questions are
(Q1) Is Hebb supervised or unsupervised learning algorithm?
(Q2) What is the correct equation?
(Q3) Can somebody provide a simple code for Hebb learning?
(Q4) What are the other supervised learning algorithm?
(Q5) Does evolutionary algorithm for learning like Genetic algorithm, particle swarm fall under supervised?
Thank you
Best Answer
Hebbian learning is unsupervised learning.
There have been multiple interpretation of Hebbian Theory in terms of which equation is given.
There are a lot of supervised learning algorithms! But for artificial neural networks, we mostly do error-backpropagation.
In evolutionary computing the fittest survive - according to some criterion of fitness, which can be defined in either a supervised or an unsupervised manner.