What Makes Intelligence Work? Jeff Hawkins Theory of 1000 brains

drawn image of calendar icon

May 23, 2024

hand drawn writing icon

Anastassia Lauterbach

In my updates, I mention publications and research that might be a couple of years, if not decades, back in time. I promise to bring to your attention only those sources that brought the AI and Robotics industry forward. In the end, ‘Romy and Roby’ isn’t a place for fast food and thoughts.

Despite all neuroscience advances, we’ve yet to progress on the most critical question: How do simple cells in the brain create ‘thinking’ and ‘intelligence’?

Before the Covid pandemic, Jeff Hawkins and his team discovered that the brain uses map-like structures to build a model of the world. In fact, it’s not just one model. Those are thousands of models of everything we know.

This discovery allowed Hawking to answer an essential question about how we perceive the world, why we have a sense of self and the origin of high-level thought or obstruction.

Hawkins started to ponder how brains work decades before AI and Robotics became fashionable in the general public. In 1979, he read an essay by the DNA pioneer Francis Crick, who thought about the mysteries of the brain.

Hawkins was highly frustrated about neuroscience’s slow progress in the following years. He took a detour from science and volunteered in entrepreneurship. He created Palm Computing. He left the organization in 2005 to establish Numenta Labs, a brain research company and institution in California.

Let me explain the basics of Hawkins’s theory.

Our brains have two major sources.

The old reptilian brain comprises several differentiated pieces, such as the medulla or the cerebellum. Its structure is heterogeneous.

But there is something in the brain that exists in all mammals and only mammals. This is the neocortex.

This is a place where our advanced intelligence sits and is being produced.

The neocortex is structurally homogeneous: Any section looks the same as any other piece.

The neocortex is highly dense. 100,000 neurons are crammed into a space as ample as a grain of rice.

They are connected at half a billion synapses by several kilometres of cabling called axons and dendrites.

Many different activities are going on inside various parts of the neural cortex. What differentiates them is not our structure but their connections. When we learn something, those connections are strengthened; when we forget something, they are weakened.

What are the basic units in the narrow cortex? Those are so-called cortical columns. Each column contains upwards of 100 neurons.

They have between one and 2 million of them in the neural cortex.

Here comes the critical insight into Hawkins’s theory.

What are those cortical columns? How do they attach so-called reference frames to objects in the world? How do they create abstract concepts?

There are ‘what’ columns that attach frames to external objects and ‘where’ columns that attach frames to your body.

This enables the brain to understand where it is in the world and, therefore, can vacate the world.

So, in the neural cortex, the neurons produce and update models of the world.

Hawking says there is no central control room in our brains; instead, perception is a consensus that the columns reach by voting.

So, predictions are created within the columns, within the neurons.

Depending on how successful those predictions are, the neurons will vote for their ‘version of events.’

The model that emerges from this voting is the aggregated strengths of all those predictions.

Our brains are pretty similar to what happens in some democratic institutions. Our cells vote, provide feedback, and pick the most strong assessment.

I admire Hawkins’s work because he does not shy away from asking profound questions. He could have ‘just’ build deep learning models behind practical applications and services. Just building, however, does not move the AI field forward.

I wonder, what further parts of the brain contribute to mammal intelligence, the formation of thinking, and the shaping of what some scientists describe as ‘awareness’ or ‘consciousness.’ In one of the upcoming podcast episodes, I will talk to Ludwig Huber, a Professor of the University of Vienna, and researcher of animal cognition. Stay tuned for the most amazing processes behind ‘thinking.’ As an example, while listening to Ludwig, I got rid of one of the myths behind my own perception of the energy efficiency of brains.

romy and roby and the secrets of sleep book cover

Book 1

Romy & Roby And the Secrets Of Sleep.

0 Comments

Submit a Comment

Your email address will not be published. Required fields are marked *

Select your currency