This site is not maintained. Click here for the new website of Richard Dawkins.

← Computer Based on Insights From The Brain Moves Closer to Reality

Computer Based on Insights From The Brain Moves Closer to Reality - Comments

AMreasonedthinker's Avatar Comment 1 by AMreasonedthinker

IBM = Intelligent Bloody Machines - no matter how smart they get, something tells me in 30 years time I will still be bombarding my computer with expletives and coffee spray for its failure to perform a simple function.

Thu, 19 Nov 2009 13:13:00 UTC | #414761

NewEnglandBob's Avatar Comment 2 by NewEnglandBob

Reading this hurts MY brain. I will try it again later in the day.

Thu, 19 Nov 2009 13:22:00 UTC | #414763

glid's Avatar Comment 3 by glid

Amazing!
Sad though that the mostly crazy-christian military is paying for the research. So the technology will primarily be used for their crusades.

Thu, 19 Nov 2009 13:54:00 UTC | #414768

Nunbeliever's Avatar Comment 4 by Nunbeliever

Well, I did not even understand half o the technical information they presented... But it sounds cool :-)

Thu, 19 Nov 2009 15:34:00 UTC | #414788

DeusExNihilum's Avatar Comment 5 by DeusExNihilum

Will IBM be the Real life Cyberdyne (as opposed to the actual real life cyberdyne)?

Either way, It would amaze me to no end if they can build something smaller than a brain and that is as powerful as a brain in my life time. Next stop, Holodecks; Come on IBM get me those holodecks!

Thu, 19 Nov 2009 16:07:00 UTC | #414794

TIKI AL's Avatar Comment 6 by TIKI AL

If they run into problems mapping the wiring diagram of the brain, I highly recommend the guy that did our hot tub.

Thu, 19 Nov 2009 16:15:00 UTC | #414796

Left Ahead's Avatar Comment 7 by Left Ahead

The mind boggles--literally. I wonder if this is the first step to the next step: silicon-based evolution.

Oh, and by the way, I'm lining up for life after death. Upload me, baby.

Thu, 19 Nov 2009 16:24:00 UTC | #414798

Left Ahead's Avatar Comment 8 by Left Ahead

Say, do you think that if mindsets like Bananaman's ever made the jump to droidworld, they'd spend their time denying links to our carbonaceous forebears? Didn't Dr. Dawkins mention something like this in The Blind Watchmaker?

Thu, 19 Nov 2009 16:31:00 UTC | #414801

ridelo's Avatar Comment 9 by ridelo

Interesting question: if they manage to upload my brain functions (in fact: me) in such a system with all my memories and my ego, does this mean I will achieve immortality even if my biological body will perish? I remember the SF stories about matter transmission: you were annihilated in one place and reassembled in another, maybe a few galaxies away. Could it be something like that but now in the form of thinking and feeling computer software?

Thu, 19 Nov 2009 17:10:00 UTC | #414816

God fearing Atheist's Avatar Comment 10 by God fearing Atheist

Don't get your hopes up too far. All they have so far is a lump of silicon neurons slightly larger than a cat brain running in near real time.

While every computational neuroscientist on the planet is going to ask Father Christmas (aka grant body) for one, it will be to use as a workbench to play with their simulations that have so far been run on ordinary PCs, clusters, and occasionally super computers. It will be years before "cat brain" software is running on the "cat brain" hardware. Then it will need scaling up by 50-100 in number of neurons, and 500-1000 in synapses to get an emulation of a human brain running in real time. In terms of Moores Law that's only 15 years of hardware development. I'm going to guess that the software to run a human brain is no more than 3 times as complex as that of a cat. However, the cat software is FSM knows how far away; but then scaling up the existing models has been impossible without this kind of hardware, so who knows!

Thu, 19 Nov 2009 17:52:00 UTC | #414824

robotaholic's Avatar Comment 11 by robotaholic

10 bux says Ray Kurtzweil will go nuts on this and make some "prediction" based on exponential increase-

Thu, 19 Nov 2009 18:00:00 UTC | #414825

Cents's Avatar Comment 12 by Cents

Re: Robotaholic
The machine intelligence future is out there.
I'm currently reading his latest book on "Spiritual Machines" Isn't it enough for him to say "I told you so".

Thu, 19 Nov 2009 18:29:00 UTC | #414832

Peter Grant's Avatar Comment 13 by Peter Grant

Glad they're doing more research into analogue computing again, digital will only take us so far.

Thu, 19 Nov 2009 18:29:00 UTC | #414833

Carl Sai Baba's Avatar Comment 14 by Carl Sai Baba

It won't be long now... IBM is going to replace Dan Dennett's brain with a computer named Yorick. And then where will he be?

Thu, 19 Nov 2009 21:24:00 UTC | #414888

alexo's Avatar Comment 15 by alexo

"Well I don't think there is any question about it. It can only be attributable to human error. This sort of thing has cropped up before, and it has always been due to human error." HAL 9000, "2001 A Space Odyssey"

Will we ever achieve a "Hal"? - I hope so, but obviousley without the psychosis!

Thu, 19 Nov 2009 21:32:00 UTC | #414889

Crazycharlie's Avatar Comment 16 by Crazycharlie

The Singularity is near.

Thu, 19 Nov 2009 22:11:00 UTC | #414902

the4thNeutralNuclide's Avatar Comment 17 by the4thNeutralNuclide

I like the idea that they are moving toward low power consumption:
:

[IBM]...announced significant progress toward creating a computer system that simulates and emulates the brain's abilities for sensation, perception, action, interaction and cognition, while rivaling the brain's low power and energy consumption and compact size


The (lower power version of) Blue Gene computer is planned to draw 6 megawatts. Humans get by on 100 watts. So 60,000 humans are 'equivalent' to one carefully-controlled simulation using 1.6 million processors occupying 250 sq meters! We're safe for a while!

BTW on the subject of downloaded memories transferred into new bodies (hormonally-flooded teenage ones), Peter Hamilton writes very good stories in a future based around the possibilities for evolution to the silicon level.

Tom

Thu, 19 Nov 2009 22:24:00 UTC | #414908

alexo's Avatar Comment 18 by alexo

No 9000 Alexo has ever made a mistake or distorted information, apart from how to spell "obviously"

Thu, 19 Nov 2009 22:27:00 UTC | #414910

Henri Bergson's Avatar Comment 19 by Henri Bergson

It'll be interesting to see where this leads; but of course consciousness cannot be explained purely by the brain, due to the 'explanatory gap'.

Thu, 19 Nov 2009 22:53:00 UTC | #414921

Steve Zara's Avatar Comment 20 by Steve Zara

Comment #433214 by God fearing Atheist

I agree that this over hyped.

I think it might be reasonable to get excited once the electrical and chemical activity of even a single synapse has been accurately modelled. We are years away from that.

Thu, 19 Nov 2009 22:59:00 UTC | #414925

Crazycharlie's Avatar Comment 21 by Crazycharlie

Reporter Martin Amer interviews Hal.

Amer--
'Hal, you have an enormous responsibility on this mission, in many ways perhaps the greatest responsibility of any single mission element, does this ever cause you any lack of confidence?'

Hal--
'Let me put it this way mister Amer. The nine thousand series is the most reliable computer ever made, no nine thousand computer has ever made a mistake or distorted information, we are all, by any practical definition of the words, fool proof and incapable of error....'

Thu, 19 Nov 2009 23:31:00 UTC | #414936

Crazycharlie's Avatar Comment 22 by Crazycharlie

".... We're safe for a while."

".... We are years away from that."

".... All they have so far is a lump of silicon neurons slightly larger than a cats brain running in near real time...."


Ah yes, Humans, keep reassuring yourselves.

Year after year Science, continually, will chip away at your conceits.

Thu, 19 Nov 2009 23:54:00 UTC | #414940

Oberon's Avatar Comment 23 by Oberon

Comment #433315 by Steve Zara

The electrical characteristics of biological neural networks can be modeled with a high degree of fidelity using both synthetic models (http://www.izhikevich.org/human_brain_simulation/Blue_Brain.htm) and more biologically realistic models.

As far as simulations are concerned, it depends if you only want to simulate the computational aspects of neural networks (i.e., their spiking behaviour) or their rather messy biology (for simulating neurological diseases, etc).

In the former case, I would argue that we have a few suitable models for this task.

Fri, 20 Nov 2009 00:05:00 UTC | #414943

Steve Zara's Avatar Comment 24 by Steve Zara

Comment #433333 by Oberon

I am familiar with the Blue Brain project.

But I also feel that messy biology is also part of the normal functioning of neural tissue.

I think that simulation of biological neural networks is important and exciting research.

But, I would like to suggest that there are some ethical concerns.

If, like me, you consider that our minds and experiences are nothing more than the activity of neurones, then we should be careful about simulating the activity of neurones.

If we get to the stage where we can simulate most of the brain of a mammal, then don't we have a responsibility to take care of that simulated brain? If we really are materialists, then we surely can't approve of simulating pain in a model brain.

Perhaps we are only years, rather than decades, away from this situation.

Fri, 20 Nov 2009 00:17:00 UTC | #414946

Oberon's Avatar Comment 25 by Oberon

Comment #433336 by Steve Zara

I agree whole heartedly that there are big ethical issues with simulating mammalian like intelligence within a computer (whether using simulated neural nets or more traditional AI techniques). If a computer simulation behaves and responds in the same way as a human then it out ethical responsibility to accord the simulation the same rights we would accord to a real human. I suspect our ethics in this sphere might lag behind the technology though as some people might have difficulty seeing computers as something more.

For me, one of the exciting things about this research is not being able to simulate mammalian like brains but rather unraveling the computational paradigm of these neural networks so that the principles can be incorporated into constructing more capable software/hardware that behave in more intelligent ways.

Fri, 20 Nov 2009 00:32:00 UTC | #414951

Crazycharlie's Avatar Comment 26 by Crazycharlie

It was said, with arrogant confidence, a computer would never beat a human at chess.

Deep Blue did it, beating world champion Garry Kasparov over ten years ago in 1997.

I watch, with keen interest, news of computer development. If only because it exposes human arrogance.

Fri, 20 Nov 2009 00:34:00 UTC | #414952

Bluff_King_Hal's Avatar Comment 27 by Bluff_King_Hal

"It was said, with arrogant confidence, a computer would never beat a human at chess.

Deep Blue did it, beating world champion Garry Kasparov over ten years ago in 1997. "

Er, said by who? I dont recall anyone ever saying that. Not anyone whose opinion was of merit in any case. Also it didnt really win computationally. Of its two "wins" Kasparov first resigned in a drawn position, and in the second he fell into a book trap which the beast simply looked up in a database - not a flop of analysis was required for it. Long ago former world champion Bronstein pointed out that winning by just looking up a move in a database hardly counts as technical brilliance. Deep Blue's designer speculated, probably correctly imo, that Kasparov knew of the trap and gambled it wasnt in Blue's database. In any event it was also known long before that "western" chess was not the greatest test of intelligence even of similar board games, as its high dependence on material, rather than positional strength, made it easier to program than Japanese chess Shogi and especially Go. Even more notably, the game Of Arimaa can be played with a conventional chess set and board, has rules simple enough for a 5 year old to understand, yet, last time I checked, s/w running on the now v powerful commercial PCs can barely lay a glove on competent human players, even though there are very few of them to provide a Gazza. Imho Kasparov or Kramnik playing Arimaa would probably nail a Deep Blue strength machine.

In the Wikipedia entry, this bit is the most apt:

http://en.wikipedia.org/wiki/Arimaa#Computer_ineptitude

Interestingly on Horizon a woman from NASA claimed that human exploration of Mars was essential as robots couldnot be made smart enough for over 30 years. Not sure I believe it. For years people were saying robot exploration made sense and human exploration was simply political to gwet the public interest to justify the funding - and given the rapid rate of advance, logically that ought to be even more true now. I donthink its ethical to risk human lives and spend a huge amount of money when bots can achieve as much or better for a great deal less.

Fri, 20 Nov 2009 01:21:00 UTC | #414959

Crazycharlie's Avatar Comment 28 by Crazycharlie

It didn't win computationally? Really ..I'll have to take your word on that. Chess is a complete mystery to me. Never made a claim to be a chess expert. But.... I do know it was reported in the news at the time, and if you look it up on Wiki, that Kasparov lost to Deep Blue in a 1997 match.

Wiki,--- "... In May of 1997, an updated version of Deep Blue defeated Kasparov 3 1/2- 2 1/2 in a highly publicised six-game match...."--"This was the first time a computer had ever defeated a world champion in match play."

Computers are getting more powerful all the time and I love watching people with bruised egos squirm because of it.

Fri, 20 Nov 2009 02:05:00 UTC | #414962

Crazycharlie's Avatar Comment 29 by Crazycharlie

As for my comment about it being said that computers would never beat a person at chess, I thought I had heard that somewhere. I should have been clear and said.... 'I'm sure there were many who used to think or say computers would never beat a chess champion.--- I'd say that's true given the many human conceits that have been demolished in the past.

P.S. And I'd bet it was said with arrogant confidence too.

Fri, 20 Nov 2009 02:20:00 UTC | #414966

Bluff_King_Hal's Avatar Comment 30 by Bluff_King_Hal

"It didn't win computationally? Really ..I'll have to take your word on that."

You don't. I clearly (to me anyway) never disputed but actually confirmed that Deep Blue won the match. TO repeat, of its two won games, one Kasparov resigned in a drawn position and the latter fell into a book trap that might have been been in chess opening books since before the publication of "Origin of Species". All that Deep Blue had to do was look up the move in its gigantic opening database with no more sophistication than a bank server looks up your balance when you use an ATM.

You dont have to take my word as if you continue your researches on the match you will almost certainly find confirmation. I found out these facts in a book about the Deep Blue story written by its designer, who is hardly likely to play down his own machines victory, which unfortunately I no longer have, or I'd quote it.

I guess I have to admit though that the "experts" are often full of it. With regard to the 'false resignation' game many chess experts declared that Deep Blue made a move "no computer would ever make" and the Deep Blue team must have cheated. Well, yours truly plugged the position into chess s/w on his humble PC and left it munching overnight and lo and behold, it eventually found the move "no computer would ever make". I compared the ratio of analysis speed of Deep Blue to my PC s/w and found that my PC found the move after considering almost precisely the same # of positions as Deep Blue did - which means, even not considering the likely superior programming of Blue, it was perfectly capable of finding the "impossible move" in the time it used, and indeed one would expect it to.

This is also only one of the bullsh1t comments about the supposed inability of puters to understand chess that have been made, if you go back far enough, so I guess you were right about that. :)

Fri, 20 Nov 2009 03:18:00 UTC | #414973