IBM Devises Software For Its Experimental Brain-Modeling Chips

alphadogg writes "Following up on work commissioned by the U.S. Defense Advanced Research Projects Agency (DARPA), IBM has developed a programming paradigm, and associated simulator and basic software library, for its experimental SyNAPSE processor. The work suggests the processors could be used for extremely low-power yet computationally powerful sensor systems. 'Our end goal is to create a brain in a box,' said Dharmendra Modha, and IBM Research senior manager who is the principal investigator for the project. The work is a continuation of a DARPA project to design a system that replicates the way a human processes information." Also at SlashBI.

Comments

The key to Synapse

[katana]

It's not in the box, it's in the band.

Re:SKYNET T001

[ImdatS]

Actually, the idea is to understand two things, which we still don't know how they work:

1) The brain itself - how does it work, how does it process information
2) More important: what is the relation between brain and "mind" - is "mind" created in the brain or is the brain the mind. More relevant for humans is "consciousness": does the consciousness inform when a complex-enough brain is created or how does it exist?

I'm not sure this project will get closer to answering Q2, but it will probably help in bringing us one step closer to answer Q1.

The goal is to create "a brain in a box."

[sehlat]

They've already done that. They call the boxes cubicles.

Re:First task

[Charliemopps]

They've already found the problem: Doing buisness with the Austrailian Government.

Re:SKYNET T001

[ImdatS]

... and to add: I would *really* like to know if there is a way to actively and consciously use the brain's subconscious ability to perform 0.2 - 2 ExaFLOPS. If we could consciously harness that ability, we wouldn't need any computer at all.

Maybe by understanding how the brain and consciousness/subconsciousness works, we could harness that - but I think, it's just wishful thinking...

Re:Synapses Don't Empower Memory

[glueball]

Synapses most definitely have to do with memory, although the synapse is not where the memory resides. Sensitize or desensitize receptors will have affect memory performance. Read about cocaine studies on short term memory.

Re:Synapses Don't Empower Memory

[Immerman]

Actually, IIRC one of the best guesses as to how the brain stores information is in the pattern and weighting of synapses - neuron A gets energized by some combination of firings of the thousands of incoming synapses, plus its own internal memory/processing, and it fires as appropriate. AFAIK the details are one of those things that keeps getting more complicated the closer we look.