Obscurity.Wiki

๐Ÿ“‘ "A Framework for Intelligence and Cortical Function Based on Grid Cells in the Neocortex" by Numenta

#resource/paper

Source

https://www.frontiersin.org/articles/10.3389/fncir.2018.00121/full

๐Ÿ“ Ego Cells ๐Ÿงฉ Spatiotemporal or Network or Social Awareness ๐Ÿงฉ Analogies or Projections

Notes

A spatial-oriented framework of thought

This paper proposes a theoretical framework to understand how the cortical columns in the neocortex might use possible grid cells to generate meaning in a space-oriented manner.

In other words, by using relative spatial positions represented with grid cells, the cortical columns use positions and displacement vectors to reference features and project them onto different spaces (could be relative to different objects).

Grid cells

I'm not entirely sure how do the grid cells work in a low-level sense. I don't know how exactly does a grid cell activate and how does it passes that information onto different cells.

What does make sense is the idea of having a spatial-oriented framework of thought that allows the projection

This represents an opportunity to create a #todo/prototype later on.

Identification of concepts in different hierarchical levels

They propose that recognition of models or objects exist in all levels of hierarchy unlike they're previous assumption that proposed that high-level objects are represented only in the top-most level such as a tea cup.

They share an example where a letter could recognized by the lower V1 level or cells but if zoomed in or seen in a larger scale would be recognized by a higher v-level. This idea resonated with what I learned from fractals. It makes sense that objects have different "zoom levels" since they can be perceived from afar but can also be observered up-close when close. This only reinforces the framework since the experiential evidence backs this.

What it means for me

Reading this paper makes me feel more confident about thinking about orienting thought processing models with the experience that an artificial mind could have in a digital world. It makes me think that in a similar sense, there will have to be some digital spatial rules that apply to the language of thought that artificial minds could have.