I’ve seen the OgmaNeo demos and they are quiet impressive, and I was wondering, where can I find tutorials where they describe how to build one?
Numenta and the whole HTM community is putting a lot of effort on making the HTM Theory as accesible as posible, for example we have the HTM School videos, where mark Matt does a great job explaining HTM concepts from the ground up, all the papers released by Numenta, there is also the Building HTM Systems blog that adds explanations with code, and even this forum itself. This kind of things has allowed the community to make their own implementations, like the Bare-Bone HTM implementation. Having things like that from about Ogma.ai and Feynman-Machines would be amazing.
I find the Feynman-Machine really interesting, and I will like to learn more about it, how it was implemented, maybe try a simple implementation in Python (a language with which I’m very familiar).
I’m interested in to knowing things like:
- What parts from HTM are kept to build a Feynman-Machine
- What concepts does the Feynman-Machine add to the HTM Theory, like how it integrates the Dynamical Systems Theory to it, and things like that
- Is it posible to build a simple Feynman-Machine using Nupic? If so, how can this be achieved?
- What are the main components of a Feynman-Machine and how they integrate and interact with each other?
- What are some small and simple problems where we can test a simple Feynman-Machine
It will be nice to build a simple working example of a Feynman-Machine implementation just like Bare-Bone HTM, as a way to explain what makes a Feynman-Machine as a good approach to get Intelligent Machines. That’s way I would choose Python (and Nupic if point 3 applies) since, I think, it is more easy to read and understand, over more low level languages like C++ an Java.