1000 Brains vs 100 Brains


I think the 1000 Brains Theory essentially posits that the same objects (concepts?) are processed simultaneously in all columns and then consensus is reached by a vote among all (many) of the columns. In one of Numenta’s papers it was stated that:

“The primary benefit of multiple columns is to dramatically reduce the number of sensations needed to recognize objects. A network with one column is like looking at the world through a straw; it can be done, but slowly and with difficulty.”

But does this mean that a smaller “100-column” brain (neocortex) could eventually learn complex concepts (“democracy”)? Or perhaps is this where the traditional hierarchy weighs in and thus the formulation of “deeper” concepts benefits from more columns? Something else? I guess the core question is why does a big neocortex matter from the 1000 Brain’s perspective?


1 Like

More models means more space for intricate details. For example, take V1, the largest sensory area of the neocortex. The more cortical columns it contains, the more the field of view is split amongst them, and the more details each one can get with its particular input.

We’ll probably be able to do a lot with just 100 cortical columns, but adding more will add resolution to sensory input and enable a richer (deeper * wider) hierarchy for more abstract concepts.