In part 2, I used htm.core as a single order sequence memory by allowing only one cell per mini-column. In this post I’ll finally have a first look at the high order sequence memory.
Before we do that, I want to show you one last single order memory example however.
Single Order Sequence Memory Recap
As you might remember from the last post, these were the settings for our htm.core temporal memory (aka sequence memory).
columns = 8
inputSDR = SDR( columns )
cellsPerColumn = 1
tm = TM(columnDimensions = (inputSDR.size,),
cellsPerColumn = cellsPerColumn,
To allow the htm.core temporal memory to learn sequences effectively, it is import to understand the impact of the different parameters in more detail.
In this part I will introduce
Temporal Memory – Previously on this blog…
Part 1 just covered enough basics of htm.core to get us started, and we actually saw how the single order memory got trained.
A cycle of encoded increasing numbers from 0 to 9 was very easy to predict, as there was always just one specific value that could follow the … more
I came across the concept of Hierarchical Temporal Memory (HTM) and its implementation a while ago, and am still very fascinated about this approach to artificial intelligence.
When, about one year ago, the active development shifted towards the community fork named htm.core, which supports Python 3, it became finally time to have a closer look and try it out by myself.
BTW: According to this forum post, there are no plans to upgrade the older NuPIC library to Python 3.
A lot of documentation about the theory of HTM is available at numenta.org… more