Thursday, October 27, 2005

Proving the MPT theory

Many thanks to Christof Koch for a great article. He saved the best for last. Here's an excerpt from the final paragraph of The Movie in Your Head.

If, in fact, changing coalitions of larger neuron groups are the neuronal correlates of consciousness, our state-of-the-art research techniques are inadequate to follow this process. Our methods either cover large regions of the brain at a crude temporal resolution (such as fMRI, which tracks sluggish power consumption at time-scales of seconds), or we register precisely (within one thousandth of a second) the firing rate of one or a handful of neurons out of billions (microelectrode recording). We need fine-grained instruments that cover all of the brain to get a picture of how widely scattered groups of thousands of neurons work together.

Koch has provided here a critical piece of information for me. Science is not going to be able to prove my theory, even with its most state-of-the-art research techniques. Now I understand why so few in the scientific community are embracing Hawkins. If there's no way for them to prove a theory it's not very interesting. But if science can't prove a theory correct, isn't it destined for the speculation scrap heap? Not in this case.

What's going to prove the MPT theory correct is a working model, a functional AI.
Hawkins and his team at Numenta are even now working feverishly on such a model, as undoubtedly are other labs and individuals. The first proof is going to stun the world. It will be followed by additional proofs. In this way will the frames-in-a-film theory of consciousness be proven. Not by scientists, but by technologists.

Wednesday, October 26, 2005

The Movie in Your Head

It just occurred to me that this article in the latest issue of Scientific American Mind - I saw it about a week ago while magazine surfing at Barnes & Noble - might have been at least partially inspired by my article at Science and Consciousness Review. The author, Christof Koch, is on the Editorial Board at SCR.

Whether or not such inspiration took place, it's significant that a scientist of Koch's caliber is considering the movie theater metaphor as an explanatory device. This is one of my main upgrades to Baars' theater model. The mind is more like a movie theater than a live stage theater - even down to the projection room, which I'll speculate on in the future. Memories are like the frames in film; they're captured as temporal units of perception and stored in sequences for subsequent auto-associative recall. The sequences are held together by gap junctions formed as a consequence of synchronous neural firing (gamma synchrony, Hebbian learning), which always seems to accompany conscious activity. Hawkins' theory, augmented by Antanitus', makes it possible to understand how it all works. That's what my MPT model is about, pointing out this theoretical potential.

In On Intelligence, Hawkins mentions Koch as one of the few scientists willing to tackle consciousness, and one of the team that discovered the Bill Clinton neuron phenomenon, which seems to offer strong evidence for Hawkins' hierarchical regions concept of neocortical memory. However, I haven't read anything Koch has said about Hawkins' ideas anywhere on the Web. In fact, it seems the scientific community is content to wait until Hawkins proves his theory with a functional AI before offering its opinion.

I'll comment on Koch's article itself in a future post.