Researchers Build Evolving Brain Computer?

destinyland writes "'We have mimicked how neurons behave in the brain,' announces an international research team from Japan and Michigan Tech. They've built an 'evolutionary circuit' in a molecular computer that evolves to solve complex problems, and the molecular computer also exhibits brain-like massive parallel processing. 'The neat part is, approximately 300 molecules talk with each other at a time during information processing,' says physicist Ranjit Pati of Michigan Tech. When viewed with a scanning tunneling microscope, the evolving patterns bear an uncanny resemblance to the human brain as seen by a Functional MRI. Using the electrically charged tip of a tunneling microscope, they've individually set molecules to a desired state, essentially writing data to the system. And while conventional computers are typically built using two-state (0, 1) transistors, the molecular layer is built using a hexagonal molecule, and can switch among four conducting states — 0, 1, 2 and 3, suggesting it may ultimately have more AI potential than quantum computing."

Star Trek predicted this one....

[Dr_Barnowl]

can switch among four conducting states

Hmm, maybe that's why all the memory units in Star Trek are "quads"..... (I've heard it retconned as "quadrillion bits" - but really this fits better).

Re:IMHO

[Anonymous Coward]

"What we really need for these kind of processes is a computer made out of very simple, small and fast elements that do exactly the task you want them to do and that are all connected."

aka an analog computer...

Re:Goddamnit, no.

[somersault]

Apparently this is obligatory [xkcd.com], so I'd better post it

Re:IMHO

[RockoTDF]

Most computational neuroscientists are not interested in whole brain simulation. So while those that are have some good stabs at modeling various invertebrates and insects, it isn't fair to talk about modeling neurons in terms of a complete nervous system. Most models are very small scale (one or two neurons) for learning about ion channels, neurochemistry, etc or large scale, such as the visual system. With these models, you can make predictions about the effects of drugs (at the synapse) or about large scale brain damage (visual or memory systems). Granted, most people in comp neuro do not consider themselves to be AI researchers, but more like theoretical physicists: build the models, be they mathematical or neural, and use them for experimental predictions.

Re:How is ths different than neural networks

[RockoTDF]

The neural nets from the 60s (perceptrons) were 2 layered networks that lacked inhibition, sparse connectivity, self-organization, and were prone to catastrophic interference. They were rubbish compared to what we have now, but a necessary step along the way. One of my biggest resentments towards AI researchers is that they use arguments from the 60s that were solved/refuted by psychologists and neuroscientists in the 80s.

Re:IMHO

[blincoln]

What we really need for these kind of processes is a computer made out of very simple, small and fast elements that do exactly the task you want them to do and that are all connected.

I believe Thinking Machines beat you to it, but almost no one was interested in writing software for the architecture.

Re:Goddamnit, no.

[Nadaka]

Funny thing is. That joke works for every base.

There are 10 states in base 10.
There are 10 states in base 2.
There are 10 states in base 367.