Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.programming > #1021

Re: John McCarthy R.I.P.

Cross-posted to 2 groups.

Show all headers | View raw

"Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> wrote:
> On Thu, 10 Nov 2011 11:38:39 +0200, Antti J Ylikoski wrote:
>
> > 10.11.2011 10:45, Dmitry A. Kazakov kirjoitti:
> >> On Wed, 9 Nov 2011 14:04:22 -0800 (PST), casey wrote:
> >>
> >>> On Nov 10, 8:17 am, RichD<r_delaney2...@yahoo.com>  wrote:
> >>>>
> >>>> McCarthy was also one of those responsible for the original
> >>>> 'thinking machines' hype; "we'll have electronic brains
> >>>> any day now".
> >>>
> >>> And the hype continues ...
> >>
> >> Really? I seen no serious works in this field in recent years.
> >>
> >> For long the focus has been shifted to studies of human brain (which
> >> of course would not create AI any time soon), NN hype (ditto) and
> >> practical problems (OCR etc) solvable without general intelligence.
> >
> > In their book "Artificial Intelligence, A Modern Approach", 3rd
> > Edition, Russell-Norvig crudely estimate that the processing power of
> > the human brain as operations/sec is about 10 to the 17th power
> > operations, and the number of memory updates/sec is about 10 to the
> > 14th power updates.
> >   (Russell-Norvig, 2010, p.12).
>
> I love such estimations in the context that nobody knows what does
> general intelligence mean algorithmically.

Yeah, and in the vain, when people post such large estimates, I like to
show how they might be irreverent, and how the amount of processing needed
could equally be amazingly small.

> Continuing in that vein there is actually only one operation needed: the
> operation "THINK". (:-))
>
> > I would like to add that there are some 10 to the 11th power neurons in
> > the human brain, and they are connected in the average to some thousand
> > other neurons.
>
> Without saying how exact a the digital model of a neuron [analogue thing]
> must be [in order to do WHAT?] this number tells nothing.

The number does tell us something valid, it's just that the number contains
a very large error factor.

> And this is only the beginning. I could tell how much atoms of Si one
> Pentium IV crystal possesses. That by no means would make Pentium out of
> 10 grams of river sand.

One thing I like to do, is reverse the question.  I personally think
neurons and nerve fibers are about the most shitty processing devices you
can find.  Only slightly better than marbles dropping though a gate.  Lie
this:

http://www.youtube.com/watch?v=GcDshWmhF4A

Information propagates over nerves at less than 100 m/s and neurons can
output information in the 100 bits per second range - slow as shit.

So what happens when you try to use slow as shit hardware, to perform a
fairly information intensive calculation like visual pattern recognition?
You are forced to use highly parallel hardware to make up in the width what
you could not do in depth.

So yest the brain is big and complex (lots and lots of parts), but how
complex is the calculation it performs?  This we do not know since no one
knows the calculation.

But we do know some other important things, like approximate information
bandwidth of the eyes, and touch, and of the control signals needed to move
our arms and legs.  And these bandwidth based numbers show that the amount
of information the flows thought the brain, is trivial compared to our
modern data processing munchies. An iPhone has more bandwidth power than a
brain.  A HD movie, had more a fare more information than the visual
information our eyes feed us for example, but an iPhone can suck that in
over a wireless connection, uncompresses it, and display it, in real time.

There's just as much reason to argue that a device like an iPhone, when it
displays a video with audio, is doing more computation, than the brain
does.

So back to my turn the question ground idea.

If neurons are such great information processors, lets look at another
question.

How many neurons, does it take to make an iPhone?  Is it EVEN POSSIBLE?
(assuming you had needed senors and effectors to convert signals to never
impulses)?

How do you wire a network of neurons to perform an video uncompress and
audio uncompress algorithm?  And how many neurons - which can only process
information at around 100 bits per second each, would it take?

I don't think that 100 billion neurons could do it.

Or lets try this one, how many neurons would take to duplicate the Google
search engine?  That is, an "intelligence" good enough, to tell you the top
10 web URLs, for any short set of words you give it?  And do it for
millions a people every second?

I think neurons suck at information processing, and the great complexity we
find in the brain, is there MOSTLY not due to how "smart" humans are, but
to how much evolution had to bend over backwards with massive parallelism
to get useful function out of one of the worse information processing
technologies on the planet.

> > So that is an estimate of the complexity that one must to my opinion
> > create to have a genuinely human-level Artificial Intelligence.
>
> The crucial point is computability of general intelligence. Without any
> knowledge about what the intelligence actually does, all such comparisons
> are totally meaningless.

Not totally meaningless.  They give us some bounds.  We can say the
computation needed to duplicate human intelligence is somewhere between a
small hand held battery operated processor, and a few warehouses full of
high end servers.

:)

> Let us consider the task of adding two numbers
> of 6 decimal places. For this task you need X neurons of human brain or a
> calculator of paiir thousand transistors. Does it make 2K transistors
> equivalent to human brain? The answer is YES, for this task, and NO for
> the general intelligence task. Without specifying the TASK, comparisons
> are rubbish.

Exactly.   Without knowing the specifics of the computation, we can't
answer it yet.  We can only give large bounds with some wild ass guesses at
probably distributions across that range.

I strongly believe, that most people believe AI is far harder, and far more
complex than it really is, and that belief causes them to make these
computation estimates WAY too high - they don't want to think of it as
easy, because 1) that shows how great their failures have been at trying to
understand AI and 2) it reduces what they are as a human, to something
insignificant - and most humans don't like to see themselves as something
insignificant.  They believe, like in the movies, we are super-heroes with
magic powers ruling over the world!  The idea that something like an iPhone
has enough power to equal human intelligence would freak a lot of people
out. (and I think human level AI, when it gets here, will freak a lot of
people out).

> > Another matter is that AI as an industry, as a branch of engineering
> > and as a science has been phenomenally successful.  I would like to
> > remark that in order to build aeroplanes it has not been necessary to
> > fly in the same manner as birds.
>
> Absolutely. If AI will ever created, its design won't even resemble human
> brain.

Well, at some level of abstraction, it needs to match the brain, or else it
won't end up acting like a human.  But at all other levels, there need be
no connections.

-- 
Curt Welch                                            http://CurtWelch.Com/
curt@kcwc.com                                        http://NewsReader.Com/

Back to comp.programming | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

John McCarthy R.I.P. RichD <r_delaney2001@yahoo.com> - 2011-11-04 13:17 -0700
  Re: John McCarthy R.I.P. RichD <r_delaney2001@yahoo.com> - 2011-11-09 13:17 -0800
    Re: John McCarthy R.I.P. casey <jgkjcasey@yahoo.com.au> - 2011-11-09 14:04 -0800
      Re: John McCarthy R.I.P. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2011-11-10 09:45 +0100
        Re: John McCarthy R.I.P. casey <jgkjcasey@yahoo.com.au> - 2011-11-10 01:24 -0800
          Re: John McCarthy R.I.P. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2011-11-10 12:14 +0100
        Re: John McCarthy R.I.P. Antti J Ylikoski <antti.ylikoski@aalto.fi> - 2011-11-10 11:38 +0200
          Re: John McCarthy R.I.P. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2011-11-10 11:54 +0100
            Re: John McCarthy R.I.P. Antti J Ylikoski <antti.ylikoski@aalto.fi> - 2011-11-10 14:09 +0200
              Re: John McCarthy R.I.P. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2011-11-10 14:46 +0100
                Re: John McCarthy R.I.P. seeWebInstead@rem.intarweb.org (Robert Maas, http://tinyurl.com/uh3t) - 2011-11-13 18:00 -0800
                Re: John McCarthy R.I.P. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2011-11-14 12:07 +0100
                Towards true A.I. (was: John McCarthy R.I.P.) seeWebInstead@rem.intarweb.org (Robert Maas, http://tinyurl.com/uh3t) - 2012-03-27 12:25 -0700
                Re: Towards true A.I. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2012-03-29 15:21 +0200
                Re: Towards true A.I. curt@kcwc.com (Curt Welch) - 2012-03-30 15:56 +0000
                Re: Towards true A.I. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2012-03-30 18:32 +0200
                Re: Towards true A.I. Daniel Pitts <newsgroup.nospam@virtualinfinity.net> - 2012-03-30 14:20 -0700
                Re: Towards true A.I. Patricia Shanahan <pats@acm.org> - 2012-03-30 15:01 -0700
                Re: Towards true A.I. seeWebInstead@rem.intarweb.org (Robert Maas, http://tinyurl.com/uh3t) - 2012-04-21 15:53 -0700
                Re: Towards true A.I. casey <jgkjcasey@yahoo.com.au> - 2012-04-21 18:47 -0700
                Re: Towards true A.I. "Chris Uppal" <chris.uppal@metagnostic.REMOVE-THIS.org> - 2012-04-22 10:03 +0100
                Re: Towards true A.I. Don Stockbauer <donstockbauer@hotmail.com> - 2012-04-22 04:10 -0700
                Re: Towards true A.I. Gary Forbis <forbisgaryg@msn.com> - 2012-04-02 19:46 -0700
                Re: Towards true A.I. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2012-04-03 09:51 +0200
                Re: Towards true A.I. Gary Forbis <forbisgaryg@gmail.com> - 2012-04-03 05:22 -0700
                Re: Towards true A.I. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2012-04-04 10:31 +0200
                Re: Towards true A.I. curt@kcwc.com (Curt Welch) - 2012-04-05 00:24 +0000
                Re: Towards true A.I. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2012-04-06 10:41 +0200
                Re: Towards true A.I. curt@kcwc.com (Curt Welch) - 2012-04-05 01:00 +0000
                Re: Towards true A.I. seeWebInstead@rem.intarweb.org (Robert Maas, http://tinyurl.com/uh3t) - 2012-04-21 15:44 -0700
                Re: Towards true A.I. seeWebInstead@rem.intarweb.org (Robert Maas, http://tinyurl.com/uh3t) - 2012-04-17 21:57 -0700
                Re: Towards true A.I. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2012-04-21 10:08 +0200
            Re: John McCarthy R.I.P. curt@kcwc.com (Curt Welch) - 2011-11-10 16:03 +0000
              Re: John McCarthy R.I.P. casey <jgkjcasey@yahoo.com.au> - 2011-11-10 12:16 -0800
                Re: John McCarthy R.I.P. "James" <no@spam.invalid> - 2011-11-10 13:00 -0800
              Re: John McCarthy R.I.P. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2011-11-10 21:50 +0100
                Re: John McCarthy R.I.P. curt@kcwc.com (Curt Welch) - 2011-11-10 21:07 +0000
                Re: John McCarthy R.I.P. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2011-11-11 11:43 +0100
                Re: John McCarthy R.I.P. curt@kcwc.com (Curt Welch) - 2011-11-12 22:38 +0000
                Re: John McCarthy R.I.P. casey <jgkjcasey@yahoo.com.au> - 2011-11-13 01:32 -0800
                Re: John McCarthy R.I.P. curt@kcwc.com (Curt Welch) - 2011-11-14 15:28 +0000
                Re: John McCarthy R.I.P. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2011-11-14 16:57 +0100
                Re: John McCarthy R.I.P. curt@kcwc.com (Curt Welch) - 2011-11-17 22:19 +0000
                Re: John McCarthy R.I.P. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2011-11-18 10:51 +0100
                Re: John McCarthy R.I.P. casey <jgkjcasey@yahoo.com.au> - 2011-11-14 11:42 -0800
                Re: John McCarthy R.I.P. "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> - 2011-11-13 12:45 +0100
    Re: John McCarthy R.I.P. RichD <r_delaney2001@yahoo.com> - 2011-12-14 10:28 -0800
  Re: John McCarthy R.I.P. seeWebInstead@rem.intarweb.org (Robert Maas, http://tinyurl.com/uh3t) - 2011-11-13 16:00 -0800

csiph-web