Military Begins Funding Super-Intelligent Computer Chip

Military Begins Funding Super-Intelligent Computer Chip

www.disinfo.com

The military is funding a project to create neural computing using memristors, a sophisticated circuit component which HP Labs describes as a stepping stone to “computers that can make decisions” and “appliances that learn from experience.”

In a video, HP researcher R. Stanley Williams explains how his team created the first memristor in 2008, while the article also explains how U.C. researchers made an even more startling discover: the memristor “already existed in nature.”

It matches the electrical activity controlling the flux of potassium and sodium ions across a cell mem

(visit the link for the full news article)


Related News Links:

I guess everyone is going to think Terminator; seems like this could revolutionize electronics itself if I understand the article correctly. The original article compares the "memristor" to a neural synapse. It s interesting to think that all these years of electronics and we were missing a 4th crucial element for the systems.

DARPA seems hellbent on artificial intelligences/beings these days. Makes sense. Military campaigns from the west tend to end now when a certain percentage of deaths occur - they even have a name for it. Artilficial soldiers remove this inconveniant media problem.

www.disinfo.com
(visit the link for the full news article)

[edit on 2-4-2010 by liquidself]

Good catch. S&F

I need to think more about this, but bump.

In the meantime, check out this thread:

NSA use of Remote Neural Monitoring

Interface anyone?

Most computational systems that use imprecise data processing like that don't lend themselves to concrete problem solving. It may be that memristor/neuristor elements might be an interesting memory topology but as a hard computational element, it has a lot of the same shortcomings as, say, holographic computing.

It's great for what it's good for, but you wouldn't want to balance your checkbook with it, or design a bridge.

Thus, it might be the ticket for dealing with imprecise situations such as real-world machine vision, yet be useless for computing protein folding. When you hear these guys saying it will be the new computational tech that blows away everything, you gotta see it in the light of "we need a lot of funding, here is some hype for you".

You want some interestingly understated stuff, you ought to go to a Sandia Zettaflop Computing conference where the topic is "Unexpected emergent behaviors", which is a nice way of saying, what happens if you get a Colossus event.

Also MOB and QUEST are interesting DARPA projects that are running now, if you liked SYNAPSE.

The fact that this comes so close to a biological computer kinda scares me. Combine this with the Giant mega super computer (and of course they will) and you could have this : en.wikipedia.org... (by the way ,james cameron says this story sparked the terminator idea) [atsimg]http://files.abovetopsecret.com/images/member/06ad9f411703.jpg[/atsimg]



[edit on 2-4-2010 by bluemooone2]

Fogot to put in the link to the original article the disinfo article is from:

Synapse on a Chip

Those are good points; I will certainly be looking into those references. It seems that if this kind of work is better for "fuzzy" type applications that then maybe it would make sense to combine it with traditional hard computing. I am reminded of Marshal McLuhan s 4 basic terms for analyses of technology from his Laws of Media:

(ENHANCE, REVERSE, RETRIEVE, OBSOLESCE)

could memristors be retrieve?