Lab - XOR solving with a MLP
Lab - XOR solving with a MLP
Description
The XOR is not linearly separable, while OR and AND are linearly separable. One way to solve the XOR is to use a Multi Layer Perceptron and to train it to learn the following table of truth:
Files
souce code (xor.cc)
How play with the code ?
1. go to lines 84, 85, 86:
2. change parameters of training such as (maximum number of iterations, minimum Mean Squared Error and learning rate)
3. re-compile
4. run it and plot the file "mse.txt"
Ackowledgment
I would like here to address a special thank to Daniel Collobert. Thanks to him, I discovered Artificial Neural Networks and he provided me with introductory examples such as this one.