HTM Hackers' Hangout - Apr 6 2018


#1

3PM TODAY! Sorry for the short notice.

Still brainstorming about topics to discuss, but I will go over work I’ve been doing and some upcoming plans with Jeff.

HTM Hackers’ Hangout is a live monthly Google Hangout held for our online community. Anyone is free to join in the discussion either by connecting directly to the hangout or commenting on the YouTube video during the live stream.

If you have something specific you’d like to discuss, or if you just want to learn more about the HTM Community, please join HTM Forum at https://discourse.numenta.org. We have active discussions about HTM theory, research, implementations, and applications.

More info on all these topics at http://numenta.org.


Oscillatory “Thousand Brains” Mind's Eye For HTM?
#2

Anachronisms from earlier versions of HTM theory and BAMI?
Clarification on what is no longer part of HTM cannon and maybe what has replaced it.


#3

Discussion/Coverage of various projects listed in: Who is currently Building HTM Systems??


#4

Will start in about 30 minutes. I’ll post a link soon. Is anyone waiting to join?


#5

I’m here watching for the link when it pops up.


#6

Yes


#7

Join here.


#8

I just got home from my day job, and with it being 6:00 here on the East
Coast I’ll be listening in on YouTube. I’m now planning what to work on this
weekend.


#9

I will be listening in on YouTube as well.


#10

Thank’s Matt for mentioning my model! I had no microphone connected.

I was expecting it to stream on YouTube, then the “Join here” link brought
me to someplace new to me.


#11

@rhyolight you’re ‘building htm systems’ project looks awesome. Does it exist anywhere yet, or you still developing it locally?


#12

It exists at htmtheory.org, although it is totally incomplete. I just started working seriously on it. You can test it out if you like, but I’m not promoting it right now.


#13

Very nice framework
Are you planning inset graphics panels?
You are more structured than I am; I like the flow of the tool.

I think the finished work product will be excellent.


#14

@Gary_Gaulin Hope to see a short live demo of your model in the next Hangout. I’m keen to port your VB6 code base to Javascript.


#15

Hey - I still code in VB at work. All my test harness are in VB.
Testing IO on embedded system.


#16

You spoke of oscillation!
From one of the old guard - some very smart words on that very topic.
BTW: kinda long but worth the effort.


Slide deck:
http://idies.jhu.edu/wp-content/uploads/2016/09/Sejnowski.JHU16.pdf

Spike timing, signaling, and learning:

Waves (circular and otherwise), and sleep spindles - important bits about the transfer from hippocampus to cortex:


HTM Mini-Columns into Hexagonal Grids!
#17

BASIC (Beginner’s All-purpose Symbolic Instruction Code) is a successful language in my book. Simply put, it was easy to learn and program, which allows the developers to focus on the problems they need to solve. It was the first language I was exposed when I was young. I can say that it’s probably one of the few things that made Microsoft successful commercially, in the “Desktop” or PC era. Javascript may very well be the BASIC for the web. Perhaps, there needs be a new programming paradigm that achieves similar effect as BASIC to accelerate the advancement of HTM framework? NUPIC (and all other frameworks/libraries built by the community) seems to be a step toward that direction. And the cool visualizations from Matt and the upcoming interactive web tool can make HTM more inclusive to broader audience (similar to what Andrew Ng did in Coursera on Machine Learning). Inclusiveness is the key there.


#18

5 posts were merged into an existing topic: Oscillatory “Thousand Brains” Mind’s Eye For HTM?


#20

Yes it was very worth the effort! I appreciate the link, needed that.

A traveling wave involving the entire neocortex takes me back to the very basics of social cell behavior. There is a wave pattern for moving food through the colony, along with other cell to cell signals that allow each to sense what’s going on somewhere else. In case anyone missed this one:

HTM theory accounts for some of this by having many predictive units where each one can like Jeff would say look around through a straw to see the wider picture. One way to do that is move the picture instead of the straw. It’s then predicted that we see waves something like this, in case anyone missed these:

The hippocampi seems to be where the cortical wave is in a way absorbed. It’s possible that by then that internal map/thought was mostly completed by short cuts direct from the cortical surface to inner structures, but I’m now speculating. In either case as this applies to my model there are two complete pictures of what is being sensed where the other kinda swirls around while the hippocampus related areas pieces that together into a stable mental vector map/image of what we sense in the environment or imagine ourselves to be navigating.