Direction for "learning by doing"


After reading “On Intelligence”, the BAMI and a few HTM School videos, I would like to get some more feelings about how an encoder covert data to sparsed binary vectors, how SP learns SDRs from those vectors, and how TP predicts from the SDRs with coding.

I noticed that there’s a workshop repo in Github, HTM Studio, and a NuPIC notebook, but I am not sure where to start from. Can you point a direction for a smoother learning path? Is there another good easy case study that I was missing?



I guess starting from htmpapers in Github could also be a good way.

Hey @JARVIS. Welcome to the forum.

@rhyolight, who among other things maintains this forum, is also developing a series of interactive webpages to help people like you. It is work in progress, but the encoder part is quite advanced, if not finished.

Sometimes he live streams his work and people can actively help out. You can also ask questions in chat.

I’m personally working on a MIDI encoder in Python at the moment, but my skills are limited and I need the occasional shove to keep going.