Bidirectional Associative Memory
Bidirectional Associative Memory (BAM) is a type of recurrent neural network. BAM has 2 layers: input and output and information can go in both directions - from input to output and back from output to input. This page presents some demo that can demonsrate learning of BAM.
How to use
Click on any cell if you want it to have different color.
The target output is the numbers 0,1 presented in the binary system with 3 bites but with 0 replaced by -1.
There are 2 input-output pairs for training and one new input to see how it works.
Once you click Submit the BAM will learn the correlation matrix between input and output
and then will associate new input with one of the training inputs.
1. Bidirectional Associative Memory Perl Source Code
2. BAM, Finding the Weight Matrix