I have seen references in your papers (Spatial Pooling and Sequence Learning) that talk a layer consisting of a set of minicolumns. Is there some part of the current HTM theory that models how the different layers interact with each other?
1 Like
We discuss a lot about this here:
Thanks Matt. Is there any built-in support in HTM for this kind of layer modeling?
- Roland
Sure, that is what the Network API allows. But we don’t have any examples of setting up these types of networks at this point yet. It is an active research area.
OK, that makes sense. Thanks.
- Roland