Here we'll talk about all the concepts that ultimately influence the design of SEMAK, and maybe I'll start a forum or something where people can make queries and throw ideas back and forth. I'll try and make things as accessible and pedagogical as possible, so that about anyone with some programming familiarity can get something out of this.
After all this, hopefully you get some of the stuff we're dealing with, and maybe understand a bit about how this approach contrasts to (dare I say) any AI method in use today. So much academic work and industry has been focused on looking at AI through the lens of the perceptron, taking these kind of meaningless things and letting them organize themselves over a tremendous amount of trial and error to eventually become competent enough at a task. As a society we've looked at how to form complex systems using low level perceptrons, it's hard to develop or modify, it's expensive, and it doesn't scale particularly well.This project is all about looking at how the brain works at a higher level. The typical brain of a species evolves with the species, and high level features of brain function become distinguishable among the populus as their brains evolve to become better suited for particular tasks. It is by noticing these higher level features (and details like what parts of those features are necessary or just sufficient to accomplish a function) that we can begin to think about what an equivalent software system should or must look like, both in terms of specific features and cumulatively as something that performs functions isomorphic to all those in a biological brain (which is just a matter of knowing how to compose the systems that carry out their isomorphic features into a final system). Indeed, many questions can be answered by asking "What would brain do?", and trying to rewrite that answer in C++ (preferably under the lens of a COSM as described in the representation talk)