Hi guys/gals, I’ve been working on my dissertation proposal for 2 years now that proposes building an HTM classifier that recognizes stalling code. Is there any documentation that covers experiments designed to evaluate Temporal pooling algorithms similar to what Purdy did for encoders in his “Encoding Data for HTM Systems” paper? I’ve looked and can’t seem to find any. I’m under the gun to get this turned in by tomorrow to my dissertation chair. Any help today would be huge. Apologize for the last second nature of this request.
Hi Jeff. Sorry, but no. Also, I think you mean “Temporal Memory” not “Temporal Pooling”, yes? Temporal pooling is different. As far as the TM goes, there really is only two algorithms, both of which are in implemented in NuPIC. One is the TM, which is closer to the biology. The other is an older version called “BacktrackingTM” that has other non-biological bells and whistles to make it perform prediction and anomaly detection better.