Sorry, I have read the white paper several times. Is there an animation or a diagram that fully demonstrates the network’s build and learning process?
I want to implement an HTM algorithm, have some questions:
1.How to generate (or map?) sparse representations from input data?
2.What does the topology of the HTM cell network look like?
3.There are some variables in the pseudo code of the white paper. Which parts of the HTML network are they adjusting?
4.From the build to the learning, how does the HTML network change?