Slashdot Mirror
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."
1. Find complex math problem
2. Build evolutionary chip to solve the problem
3. Invent SkyNet
4. ???
5. PROFIT!
Hi, I'm your tunneling microscope programmer. I'm going to need a few new development tools in order to write your Facebook alternative website ... including a tunneling microscope. Your new site is going to give "head in the clouds" a whole new meaning!
I'm guessing that there are certain problems you can't solve with FSA so you've got to leverage the more exotic computing that happens in the fabric of reality... or at least that makes nice sounding techno-babble.
[signature]
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).
Isn't that how the Matrix supposedly started? Humans invented computers complex and "organic" enough to develop AI?
So should we kill it now before it enslaves us all, or what?
Official Heretic from the "Church of Global Warming". Proven right thanks to whistle blowers. AGW = Flat Earth Theory
Timothy Leary might have prior art on that.
"How do I escape from this lab . . . ?"
Schroedinger's Brexit: The UK is both in and out of the EU at the same time!
Oh, yes.
The easiest way to create AI is to model the neuron inside of a computer, slice up someone's brain into lots of thin slices, and then recreate their brain in the computer. Mapping the inputs is the hard part.
Edward@Tomato - /home/Edward/ man woman
man: no entry for woman in the manual.
"Qua!?"
Sigh... now knowing how to count in 10 bases wont be enough anymore.
$ unzip, strip, touch, finger, grep, mount, fsck, more, yes,fsck,fsck,fsck,umount, sleep
At what point can turning off a brain-like computer be no longer called 'shutdown', but instead be called 'death'?
Goddamnit, that is not how it works. Even if each molecule has four different states, you can easily map them onto a small, finite number of bits - you just represent each molecule with two bits in a computer, and there's your equivalency. You don't get anything out of more states per unit except higher density. Seriously, TFA doesn't make this mistake; why did you have to add some useless speculation to a perfectly reasonable article?
I'm pretty sure that this was how Mother brain got started. I'd think we'd better stop until we hear news of an orphan girl adopted by advanced alien beings.
Intelligently designed to evolve. I love it!
> "The neat part is, approximately 300 molecules talk with each other at a time during information processing," 640k limit anyone?
"My CPU is a neural net processor--a learning computer."
Dont think we know how to measure the point WE become aware yet. Beyond knowing at approximately what age it happens at and how to tell when a tot finally figures out they exist, there's not a lot we know. I figure this'l help us find out more about how our brains work in that regard, seeing as it's made in the likeness of ourselves.
I don't believe in 2.
I'll bet the DoD can't wait to get one of these...
"Goddammit, I'd piss on a spark plug if I thought it'd do any good!"
blowing my mind, man!
How many more years will slashdot have an off-by-one error on your Score in your profile?
Haven't any of these people ever seen "Colossus, the Forbin Project?"
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.
There is more to science than physics!
www.iomalfunction.blogspot.com
I, for one, welcome our new robot overlords
I always love quotes like this... as if this guy (or anyone) knows how neurons actually behave in the brain. So far we're still at the simple model phase... to fire or not to fire, that is the question (apologies to Shakespeare).
Ptolemy thought he understood gravity, then Newton proved him wrong. Newton thought he understood gravity, then Einstein proved him wrong. Einstein thought he understood gravity, but folks like Penrose, Ashtekar, Smolin, and the Loop Quantum Gravity guys are about to overturn Einstein... When it comes to our understanding of how neurons work, we have more in common with Ptolemy than Einstein...
I am going to make a computer with FIVE states!
That will be much better than a quantum computer because it has FIVE, you see.
And it will be an AI because it has FIVE states, and normal computers only have two.
Actually, I will have EIGHT states just to make sure the competition can't catch up to me.
I will be implementing my EIGHT states as 3 binary bits, but that's not important right now.
- For the complete works of Shakespeare: cat
Me, I don't care about the talking molecules - scale it up to a talking dog and then we'll talk :-)
Unfortunately I dont have a link, but an interesting show on NPR talked about a computer that arrived at F = ma from watching a pendulum, and having never been given that as information. This same program was used on a dual pendulum (pendulum on the end of a pendulum) as was able to determine the physics behind this (a previously unsolved physical model) Ultimately it was used to figure out the relationship between different ions in a cell and how the concentrations of each relate to each other.....research that could not be published because they had only the answer, but not the derivation or explanation as to why. I'll search and see if I can find a link to it.
www.RacquetUp.org - Helping Detroit Youth
People have been over hyping AI for as long as it has been around. In this case all the researchers have managed to do is create a 2D cellular automata like structure using a chemical bi-layer. A nice piece of work but they have no idea how to use a 2D cellular automata to create intelligence either! What you use to implement it isn't the issue. Using words like self organising, healing and showing scans of the brain against scans of the chemical layer is simply misleading. AI needs this sort of work - but it can do without this sort of hype http://www.i-programmer.info/news/105-artificial-intelligence/820-brain-like-computing.html
Sure, this was the thing that held AI back for a decade. Frankly AI is pretty much stuck at the very start and has been since the very start. There's been some advances in expert systems but for the most part it's been a huge failure all the way through.
It is no longer uncommon to be uncommon.
Evolution is an algorithm as well as a way biology works. Evolution works quite well as an algorithm and you need not look too deep to find evolutionary algorithms and evolutionary programming or evolutionary design making real progress.
Frankly as far as this thing goes the use of evolution to create and repair the molecular circuits is the most useful part. The rest of it is pretty silly and crap you could model on an a computer and find out it doesn't work in a much shorter order.
It is no longer uncommon to be uncommon.