Interesting A(G)I attempt

[cezar_t]

I just found this Russian group which also approaches AI by making assumptions on the underlying processing in the brain with emphasis on minicolumns functions and role.

Some of their points might be weird but others are really interesting.

Unfortunately most of their material is in Russian.
Their key concepts English page is quite detailed though, here-s just one excerpt:

Hippocampus is not temporary storage of memories, but a dispatcher that assigns identifiers

In order to form a memory, two patterns are needed. One pattern carries the information to store; the other one carries the identifier (descriptor) of this information. We assume that the hippocampus itself does not store any memories. Instead, its responsibility is to generate the unique identifiers of memories. These identifiers propagate to the corresponding areas of the cortex where the identifiers and memorized information interfere. Interference of identifier and information results in forming memory, not in hippocampus, but directly where needed, right in the cortex.

[Bitking]

This is not that far from what I am thinking.

HTM Working Memory

That is a tricky point to answer. The bulk of the cortex learns by “simple” Hebbian process - relatively slow. The fact that patients like HM are capable of learning skills like mirror writing without any memory of having done so reflects that the episodic memory is what is missing when the EC/HC is compromised. Due to the distributed and interlocking nature of the representation, the cortex capacity is enormous. To understand this; there are only a small number of letters in an alphabet but d…

[OS_C]

HM was not only able to learn mirror writing but the spatial layout of rooms where he was at, iirc. Some speculate this was through motor memory in cerebellum, but I doubt that.

He was also able to temporarily learn for prolonged periods a bit of new facts after his accident occured, iirc. Repeated crossword puzzles apparently kept some information in memory refreshed, but later when he stopped dealing with the particulars he seemingly lost the new facts.

What I’ve heard is that metaplasticity can stop neural networks from facing catastrophic forgetting. The amount of persistent activity appears to increase higher in the cortex, iirc. In prefrontal cortex it appears that rather than persistent activity, bursts of activity with intervals of inactivity occur.

I think that what happens is that higher cortex is likely more resistant to synaptic changes from activity due to metaplasticity adaptation(adaptation to the higher persistent levels of activity). The hippocampus is able to generate replays at very high speed, iirc. Fast enough to overcome the metaplasticity adaptation and force storage unto higher cortex of some information.