Sentience
Randy Yates
algorithms? Can someone even define "sentient"?
--
% Randy Yates % "Bird, on the wing,
%% Fuquay-Varina, NC % goes floating by
%%% 919-577-9882 % but there's a teardrop in his eye..."
%%%% <ya...@ieee.org> % 'One Summer Dream', *Face The Music*, ELO
http://home.earthlink.net/~yatescr
Tim Daly Jr.
> Has there been any significant advances lately in the area of sentient
> algorithms? Can someone even define "sentient"?
Why yes, of course. In fact, I had a friendly chat with my quicksort
the other day.
--
-Tim
Paul F. Dietz
Programs sometimes scream when I torture test them. Does that count?
Paul
Randy Yates
Thanks for your response, Tim. Now go take your medication.
Brian Mastenbrook
> Has there been any significant advances lately in the area of sentient
> algorithms? Can someone even define "sentient"?
No. Nobody's really working on it. There are four different camps in AI
today, and none of them is actually going to achieve AI:
* The connectionists, aka the cargo-culters - "if we just tweak the
back-propogation, it'll start thinking!" These people have put no real
thought into a definition of intelligence, other than to characterize
it as associative and fuzzy (how we're supposed to do that on a
discrete computer, I can't fathom).
* The logicians, whose hobby is attaching wings to pigs in the hope
that they will fly. These people presume that intelligence is just a
matter of logical deduction, and that logic is the right level on which
to describe how intelligence operates. They have all the high-level
philosophical problems of AI defined and solved, but so far their
computers have not locked them outside in Jupiter orbit, leading some
to migrate to the third camp...
* The Searlites, who believe it can't be done but are still publishing
papers for God-knows-what-reason.
* And finally, there's the people doing stuff that isn't AI, namely
anything that's a domain-specific system or isn't designed to produce
domain-general intelligence. This includes the evolutionary and genetic
algorithms people, the constraint solvers, and probably anything else
that doesn't fit in categories 1-3 (and a lot of stuff that does too;
the connectionists seem awfully fond of making a neural net that does X
and only X and then publishing a paper that says "neural nets can do
X!"). There is a lot of interesting stuff in this category, but they
should go find their own field. If your program can solve a problem
reliably much better than a human can, it's not because it's smart,
it's because you've written a better domain-specific algorithm than our
own general-purpose reasoning, and it's a big clue that you're not
doing AI at all.
Spread throughout all of these camps are the semiotics people and their
deconstructivist / postmodernist associates, who are busy destroying
real intelligence in humans via the educational system.
There, that should offend just about everybody.
If you want to know why AI is not getting anywhere, read Feynman's
famous address on Cargo Cult Science (
http://www.physics.brocku.ca/etc/cargo_cult_science.html ) and compare
to the state of the field. No wonder!
--
Brian Mastenbrook
http://www.cs.indiana.edu/~bmastenb/
Frank A. Adrian
> Can someone even define "sentient"?
Yes! This is a sentient. This not a sentient because it no verb(s).
Thank you! Thank you! I'll be here all week...
faa
Cameron MacKinnon
[a humourous and insightful screed]
> There, that should offend just about everybody.
>
> If you want to know why AI is not getting anywhere, read Feynman's
> famous address on Cargo Cult Science (
> http://www.physics.brocku.ca/etc/cargo_cult_science.html ) and compare
> to the state of the field. No wonder!
Is AI science? It isn't like experimenters in the field have to discover
how some external phenomenon works, unless it's your connectionists
answering "how many neurons to achieve critical mass?" If the field is
more about groping toward algorithms that exhibit intelligent behaviour,
albeit using a different mechanism than biological organisms do, it's
more like mathematics. And if that's so, the cargo cult analogy breaks
down.
You've discounted superior performance in a particular domain, and you
seem to be setting the bar quite high elsewhere. So what do you
consider AI? Victory in the Turing test or bust? If it's defined as what
humans consider intelligent behaviour, the goalposts are likely to
recede as fast as our computers can advance, simply because the layman's
perception of intelligence is any behaviour that machines don't exhibit.
If computers can already drive cars, play chess and accidentally fill up
their hard drives* while downloading (uploading?) porn on Mars, what do
you want next? It seems that discounting domain specific successes is a
tricky way of saying "if it's solved, it isn't AI."
From what I've heard, electronic games' non-player characters seem to
be improving year after year. That's an area where there's a somewhat
independent arbiter (game reviewers) and an incentive for improvement
(sales). It isn't world changing technology, but it's a plausible Turing
test.
* metadata overload, but close enough
--
Cameron MacKinnon
Toronto, Canada
Sunnan
> In article <7jwocs...@ieee.org>, Randy Yates <ya...@ieee.org> wrote:
>
>> Has there been any significant advances lately in the area of sentient
>> algorithms? Can someone even define "sentient"?
>
> No. Nobody's really working on it. There are four different camps in AI
> today, and none of them is actually going to achieve AI:
<snip camps>
Being in the fourth camp, doing advanced information processing or
other useful stuff, doesn't really mean that you're not in one of the
three first as well. It's perfectly possibly to work on (say) advanced
search by day, and igor-tinkering with the cargo cult by night.
(just clarifying for others, not trying to correct Brian)
--
One love,
Sunnan
Randy Yates
Thanks for a reasonable response. I'm about to checkout the link.
--Randy
Brian Mastenbrook <NOSPAMbmas...@cs.indiana.edu> writes:
--
% Randy Yates % "Remember the good old 1980's, when
%% Fuquay-Varina, NC % things were so uncomplicated?"
%%% 919-577-9882 % 'Ticket To The Moon'
%%%% <ya...@ieee.org> % *Time*, Electric Light Orchestra
http://home.earthlink.net/~yatescr
Brian Mastenbrook
<cmack...@clearspot.net> wrote:
> Is AI science? It isn't like experimenters in the field have to discover
> how some external phenomenon works, unless it's your connectionists
> answering "how many neurons to achieve critical mass?" If the field is
> more about groping toward algorithms that exhibit intelligent behaviour,
> albeit using a different mechanism than biological organisms do, it's
> more like mathematics. And if that's so, the cargo cult analogy breaks
> down.
I would extend the cargo cult phenomenon beyond "just science". I've
seen entirely too much cargo cult programming - putting together parts
of programs without any real understanding of what's going on. When
you're helping a student who has close parens and open parens littered
randomly throughout their Scheme source, it's pretty obvious that the
student is not reasoning about how Scheme works, but putting together a
program that has an appearance such that it /might/ work.
It's my suggestion that there is also something "going on" in
intelligence that is difficult to explain on the level of firing
neurons, just as there's something "going on" in computation that's
difficult to explain on the level of gates on a silicon wafer with
particular electromagnetic properties.
> You've discounted superior performance in a particular domain, and you
> seem to be setting the bar quite high elsewhere. So what do you
> consider AI? Victory in the Turing test or bust? If it's defined as what
> humans consider intelligent behaviour, the goalposts are likely to
> recede as fast as our computers can advance, simply because the layman's
> perception of intelligence is any behaviour that machines don't exhibit.
I would say that victory in a domain without any analysis of whether
the program can function in other domains with even minimal sharing of
structure is not Artificial Intelligence. In addition I do set the bar
high for programs that I cannot convince myself operate in any way that
directly maps to the primitives of intelligence - some of which include
the ability to interact with an external world that is composed of
objects with persistent identity, the ability to form intuitive
predictions and increase that ability through experience, the ability
to introspect about prior thought and use that introspection to make
decisions, and the ability to translate abstract thoughts into
concrete, controlled motor function. If I do not see a mapping from any
of these constructs to the program, I cannot accept at face value that
such a program is progressing to intelligence, nor that simply
increasing the resources available to such a program will cause it to
hit "critical mass" and suddenly become intelligent. Accepting this at
face value would be the same as accepting that bamboo towers and
runways will bring cargo planes, if only the reproduction is accurate
enough.
> If computers can already drive cars, play chess and accidentally fill up
> their hard drives* while downloading (uploading?) porn on Mars, what do
> you want next?
I want them to do all of these things at the same time, and improve the
performance (if slightly) of all of them by improving an individual
skill. I want shared structure that enables a computer to quickly adapt
to a new domain by application of past experience in a different
domain.
> It seems that discounting domain specific successes is a
> tricky way of saying "if it's solved, it isn't AI."
Effectively, this statement is true, because we don't have AI yet. If
we were producing progressively better general-purpose reasoners, then
I could see the merit of assigning AI-value to them. We aren't. The
last well-known general purpose reasoner I know of is SHRDLU, which
could be taught arithemetic by instructing it to think of numbers as
collections of blocks.
> From what I've heard, electronic games' non-player characters seem to
> be improving year after year. That's an area where there's a somewhat
> independent arbiter (game reviewers) and an incentive for improvement
> (sales). It isn't world changing technology, but it's a plausible Turing
> test.
Call me when games allow interaction with the world in some way other
than weapons fire, spell casting, and limited-option decision making.
In reality, all that's happening here is the transformation of gamers
into restricted-domain reasoners. It's also the reason why some
variants of Eliza can pass the turing test, when there's very obviously
nothing which can be called "intelligent" floating around in there.
Thomas F. Burdick
I was having a little talk with some shell scripts and Applescript on
my Mac the other day. It used to speak in an inappropriately polite
brittish accent. Not good. I fixed it, though, and now we can
converse in appropriately hyper NoCal tones. I really wish I had
method combinations, they'd really help to get the "hella" quotient up
to where it should be. I don't need to talk with my Lisp programs, I
already know what they're thinking.
--
/|_ .-----------------------.
,' .\ / | No to Imperialist war |
,--' _,' | Wage class war! |
/ / `-----------------------'
( -. |
| ) |
(`-. '--.)
`. )----'
Cameron MacKinnon
> I would extend the cargo cult phenomenon beyond "just science". I've
> seen entirely too much cargo cult programming - putting together parts
> of programs without any real understanding of what's going on. When
> you're helping a student who has close parens and open parens littered
> randomly throughout their Scheme source, it's pretty obvious that the
> student is not reasoning about how Scheme works, but putting together a
> program that has an appearance such that it /might/ work.
Infants don't learn languages by reasoning out that every sentence needs
a (possibly implied) noun phrase and a verb phrase. They perform a lot
of experiments with no scientific method. Mimicry and feedback seem to
be enough.
Computer folks know that, because computer languages contain almost no
redundancy (as used to resolve ambiguities), aren't DWIM and have
pedantic compilers with truly awful diagnostics, learning one's first
through trial and error is painful, and so syntax is often taught
rigorously. But that's all hindsight and deep insight. Your students'
strategy seems quite reasonable if their only experience is in
biological languages with good feedback from their "conversation partner."
What, exactly, is the cargo cult's mistake? Seabees came and built
runways and control towers, men put funny headphones on, and the planes
came. Mimicking this looks ridiculous TO US, because we have superior
domain knowledge of aeronautics. Absent that knowledge, we could say
that they err in persisting for five years when no planes have come, but
that suggests that the cultists are to somehow know how long to wait for
results. Given that the original was a one-time event for a culture
which (I'm guessing) had an oral history stretching back many
generations, even a several-generations-long experiment might not be
unreasonable.
[If you'd have asked me yesterday whether I'd defend cargo cultists, I'd
have been insulted at the implication. I never really thought about it
before, just thought what Feynman wanted me to think.]
I meet people all the time who seem to have only the most tenuous grasp
of logic. In fact, I'd hazard that, for the majority of any given
society, Aristotle and Descartes might as well never have existed. It's
not that they are completely illogical, but they certainly don't operate
through rigourous reasoning, and often can't see the flaws in the
systems they believe.
So the interesting question is, when seeking artificially intelligent
behaviour, are we looking for computers to act like the best of us (in
which case, perhaps "Artificial Brilliance"?) or the rest of us?
> It's my suggestion that there is also something "going on" in
> intelligence that is difficult to explain on the level of firing
> neurons, just as there's something "going on" in computation that's
> difficult to explain on the level of gates on a silicon wafer with
> particular electromagnetic properties.
If the computation was designed by a person, seeking to explain it in
the behaviour of the hardware it's running on is perhaps incorrect. It's
like looking at a chalkboard full of mathematics and saying "the chalk's
really on to something there".
Since we design with logic, we can approach zero redundancy, something
biological systems don't exhibit.
Explaining human behaviour, in isolation, in terms of neurons seems very
difficult. But if we start at the lowly cockroach, we see something that
wouldn't seem to take too many neurons to emulate.
But I agree with you that pursuing a pure neural net approach, merely
because biological systems do it that way, is a cargo cult approach.
My comments on game AI weren't meant to imply that it represents the
future, merely to say that some areas of AI have been advancing, because
there's both quasi-objective metrics and motivations for the people in
that field. AI people seek to get the right fitness criteria for the
training or population culling within their experiments, but I think
that it's equally important to have appropriate fitness criteria for the
researchers themselves, otherwise they end up playing in their own theses.
Military (and eventually civil) logistics is another similar area.
Results were wanted, money was spent, results were gotten. Since funding
domain specific intelligence (or solutions) research has a payoff,
whereas general intelligence is a money pit, domain specific is where
the progress is being made.
Anyway, we may rapidly be approaching (or worse) the limits of my
insights in this area.
From your comments, I'd say you follow the progress, er, machinations,
of the AI community. Do you have any pointers for the uninitiated to
collections of stuff that's more toward the seabees' end of the runway
than the natives'?
William Bland
> Since we design with logic, we can approach zero redundancy, something
> biological systems don't exhibit.
I may be misunderstanding you here, but I don't believe this is
correct. Any redundancy in a biological system costs an organism
energy. It seems obvious that evolution will get rid of such
organisms in favor of organisms that don't have redundancy. Did
I misunderstand your point? Did you have a specific example in
mind?
Cheers,
Bill.
Cameron MacKinnon
Initially I was thinking specifically of the human brain. Examples exist
of people who've lost quite a lot of brain tissue to trauma and
continued to live quite normal lives. However, given our limited
understanding of the brain, this is not a very solid argument.
Your assertion only holds in an evolutionary environment selecting for
minimum energy use. Why do I have two kidneys, two testes, two lungs?
Cameron MacKinnon
I'm not happy with my prior response. How does one measure redundancy in
a large neural network? I think the brain has a large number of
redundant neurons, in the sense that a lot of them can die or be removed
without measurable impairment.
However, as you point out, selection has been favouring larger brains
for some time now, so in that sense, the extra brainpower must have some
benefit. I suspect that it is in rapidly coming up with witty things to
say to the opposite sex at parties.
William Bland
In all cases - brain, kidneys, testes, lungs, etc. we have
"spare" so that we don't die if something goes wrong with the
other one. But then they're not "spare", are they? They're
absolutely necessary for keeping you alive when something goes
wrong. It's only "redundant" if you don't mind catastrophic
failure. Yeah, granted, a human *could* live without them,
but I sure wouldn't want to be that human.
Bringing this back to computers, I think it's good to design
things with "redundancy". Except, again, I don't believe it's
really redundancy. It's only "redundant" if you don't care what
happens when something goes wrong. Usually you do care.
Cheers,
Bill.
Brian Mastenbrook
<cmack...@clearspot.net> wrote:
> Brian Mastenbrook wrote:
> > I would extend the cargo cult phenomenon beyond "just science". I've
> > seen entirely too much cargo cult programming - putting together parts
> > of programs without any real understanding of what's going on. When
> > you're helping a student who has close parens and open parens littered
> > randomly throughout their Scheme source, it's pretty obvious that the
> > student is not reasoning about how Scheme works, but putting together a
> > program that has an appearance such that it /might/ work.
>
> Infants don't learn languages by reasoning out that every sentence needs
> a (possibly implied) noun phrase and a verb phrase. They perform a lot
> of experiments with no scientific method. Mimicry and feedback seem to
> be enough.
Enough... with hardware that supports it. Our brains are simply wired
up to rapidly obtain the necessary information for language learning.
But there is a difference between doing that and understanding the
wiring itself. Our genetics do really provide a bootstrapping procedure
that we've got to understand now.
> Computer folks know that, because computer languages contain almost no
> redundancy (as used to resolve ambiguities), aren't DWIM and have
> pedantic compilers with truly awful diagnostics, learning one's first
> through trial and error is painful, and so syntax is often taught
> rigorously. But that's all hindsight and deep insight. Your students'
> strategy seems quite reasonable if their only experience is in
> biological languages with good feedback from their "conversation partner."
It's reasonable up until the 52nd time I explain what s-expressions are
and point them at a tutorial explaining the rules. At that point they
really are just coasting instead of thinking.
> What, exactly, is the cargo cult's mistake? Seabees came and built
> runways and control towers, men put funny headphones on, and the planes
> came. Mimicking this looks ridiculous TO US, because we have superior
> domain knowledge of aeronautics. Absent that knowledge, we could say
> that they err in persisting for five years when no planes have come, but
> that suggests that the cultists are to somehow know how long to wait for
> results. Given that the original was a one-time event for a culture
> which (I'm guessing) had an oral history stretching back many
> generations, even a several-generations-long experiment might not be
> unreasonable.
I don't think Feynman's point was to attack the cargo cult at all. He
was drawing an analogy between why it didn't work and the same type of
pseudoscience that doesn't work in our culture. In our case, we have
many methods which have been developed to enable us to come to an
understanding of something which appears magical. Since we don't yet
(and concievably won't for many many years due to the seeming exclusion
of accurate and large-scale measurement) understand how the embodied
mind actually processes information, that means we need to apply these
processes to the understanding of intelligence itself.
Without following scientific or systematic philosophical inquiry into
the nature of the general concept of intelligence and its
representation as part of the human mind, we cannot convincingly assert
that our experiments will bring in the cargo planes of intelligence.
> [If you'd have asked me yesterday whether I'd defend cargo cultists, I'd
> have been insulted at the implication. I never really thought about it
> before, just thought what Feynman wanted me to think.]
I don't think Feynman wanted you to think anything about them, but
instead about the advocates of pseudoscience in a scientific society.
> I meet people all the time who seem to have only the most tenuous grasp
> of logic. In fact, I'd hazard that, for the majority of any given
> society, Aristotle and Descartes might as well never have existed. It's
> not that they are completely illogical, but they certainly don't operate
> through rigourous reasoning, and often can't see the flaws in the
> systems they believe.
>
> So the interesting question is, when seeking artificially intelligent
> behaviour, are we looking for computers to act like the best of us (in
> which case, perhaps "Artificial Brilliance"?) or the rest of us?
The fact that rigorous logic is learned is actually a valid point in
the study of intelligence. It suggests that it might not be the right
level to describe what's going on. I personally focus much more on the
symbolic aspects of intelligence, with a slightly peculiar connotation
of symbol: something which has identity. I try to stake out a middle
ground between the subsymbolicists and the logicians.
> Explaining human behaviour, in isolation, in terms of neurons seems very
> difficult. But if we start at the lowly cockroach, we see something that
> wouldn't seem to take too many neurons to emulate.
But is it useful? My contention is that the human brain is not really a
well-designed device and is more of a kludge of cognitive ability on
top of a neural net than a reflection of some innate cognitive ability
in neurons.
> My comments on game AI weren't meant to imply that it represents the
> future, merely to say that some areas of AI have been advancing, because
> there's both quasi-objective metrics and motivations for the people in
> that field. AI people seek to get the right fitness criteria for the
> training or population culling within their experiments, but I think
> that it's equally important to have appropriate fitness criteria for the
> researchers themselves, otherwise they end up playing in their own theses.
Unfortunately I'm not convinced that those metrics are actually
measuring intelligence. There is a lot of fascinating stuff going on
there, but I really would hesistate to put the label of AI on, say,
genetic programming. It's not a slight to GP, but just a difference in
classification.
> Military (and eventually civil) logistics is another similar area.
> Results were wanted, money was spent, results were gotten. Since funding
> domain specific intelligence (or solutions) research has a payoff,
> whereas general intelligence is a money pit, domain specific is where
> the progress is being made.
I think it's shortsighted investment to only fund practical and not
basic research. For instance, basic physics research gave us the MRI,
and soon sustainable energy-producing fusion.
> Anyway, we may rapidly be approaching (or worse) the limits of my
> insights in this area.
Not at all; these are good thoughts.
> From your comments, I'd say you follow the progress, er, machinations,
> of the AI community. Do you have any pointers for the uninitiated to
> collections of stuff that's more toward the seabees' end of the runway
> than the natives'?
Actually, it's worse than that. I publish in AI. My homepage has links
to my papers, but they aren't necessarily terribly well written,
particularly because of the need to slip past reviewers and masquerade
as useful research.
While I don't claim to be a classicist in any strict regard,
http://plato.stanford.edu/entries/language-thought/ gives a good
history on some of the previous work in the field and the feud with the
connectionists.
Ray Dillinger
I dunno. I've worked a lot in AI (commercial products dealing
with natural language, not academic papers) and I've used several
major approaches, with varying degrees of "neatness" and
"precision."
What I've observed is that the "imprecise" systems, the ones
which do not satisfy the test of "classicality" in terms of
sequential sound operations on a representational state, are
the ones that work better. They are more efficient and less
likely to be killed or led awry by grammar mistakes in the
input. Something that classifies input according to bigrams
or trigrams and then does very shallow pattern-matching is
often superior at extracting meaning from genuine examples
of written english than a formal parsing system that builds
up deep structure.
That said, connectionist tools such as neural networks are
very helpful in efficiently choosing the "most likely" formal
grammar rule to use next when doing formal parsing. Systems
that used connectionist models in rule selection, applying
backpropagation to train them if they didn't lead to the lowest-
cost parse directly, quickly "learned" to outperform other formal
parsing systems by a factor of ten or more. Given that experience
I'd say that the formalist/classicist camp with their exhaustive-
search methodologies ought to be cheering for the new tools from
the connectionists and scruffies. Instead there seems to be
a fight.
But that's primarily domain intelligence again. What
distinguishes natural-language work is the breadth of the
domain; not only do you have to have a working knowledge of
the language, but for worthwhile results, you also have to
have a working knowledge of the particular domain under
discussion. Pretty much all the systems I built were only
looking for information about a very restricted domain.
Now, asking about general intelligence - what the OP called
"Sentience" -- is at this point a lot like asking a car
manufacturer to build something that's as fast, flexible,
quiet, tolerant of rough terrain, and self-maintaining as
a cheetah. It's not going to happen. The cheetah is many
orders of magnitude more complex than the best cars we can
build. Instead we must make do with "domain-specific" but
useful functionality, where we get great speed on paved
roads and have to tinker and maintain once in a while.
But I also want to point out one other thing. Cheetahs are
self-aware. They have "animal intelligence", which means
they use their brains to interact effectively with the world
around them. That's the intelligence that "human intelligence"
started as. And I think "sentience" has to be considered as
an outgrowth of animal intelligence. Animals don't think
in formal symbols. They don't have language. The fundamental
architecture of thought probably isn't "formal" in any
reasonable sense of the word. Our formalisms, our ability
to make sound inference, our ability to use logic, etc -
the things that the classicists so value that they think
intelligence must be defined in terms of them - are all
lately acquired. And that cheetah is self-aware without
them.
If we could produce a computer whose "consciousness" were
as complex as a cat's, I'd consider it self-aware, though
probably not sentient. And I think that "sentience," though
it probably involves _some_ classicist processing, can
probably only be achieved by a system that actually does
at least that much "non-classicist" processing.
Bear
Joe Marshall
> I would extend the cargo cult phenomenon beyond "just science". I've
> seen entirely too much cargo cult programming - putting together parts
> of programs without any real understanding of what's going on.
There's a *lot* of this.
Joe Marshall
> What, exactly, is the cargo cult's mistake?
It doesn't work.
It is based not in science, but in magic. No attempt is made to
determine a rational cause; instead a `like evokes like' is used.
--
~jrm
robbie carlton
> I would extend the cargo cult phenomenon beyond "just science". I've
> seen entirely too much cargo cult programming - putting together parts
> of programs without any real understanding of what's going on. When
> you're helping a student who has close parens and open parens littered
> randomly throughout their Scheme source, it's pretty obvious that the
> student is not reasoning about how Scheme works, but putting together a
> program that has an appearance such that it /might/ work.
Okay, so you're saying you don't want AI reasearchers putting parts
together to create programs whose workings they don't understand: i.e.
emergence. And yet
> It's my suggestion that there is also something "going on" in
> intelligence that is difficult to explain on the level of firing
> neurons, just as there's something "going on" in computation that's
> difficult to explain on the level of gates on a silicon wafer with
> particular electromagnetic properties.
i.e. emergence.
David Steuber
> Has there been any significant advances lately in the area of sentient
> algorithms? Can someone even define "sentient"?
I've read the responses so far. Very interesting work is going on.
However...
I don't think there is any such thing as sentience in the sense of
self-awareness. When Descartes said, "I think, therefore I am," he
was mistaken. He was wrong at the part where he said, "I think."
Everything else that followed was therefore wrong.
Atoms do not think. They are constrained by the laws of physics.
Brains are just collections of atoms arranged in a fancy pattern.
They are subject to the exact same laws.
Generally I try not to think about this illusion we call life. I
think I'm doing pretty well. Life is just a dream.
--
It would not be too unfair to any language to refer to Java as a
stripped down Lisp or Smalltalk with a C syntax.
--- Ken Anderson
http://openmap.bbn.com/~kanderso/performance/java/index.html
Cameron MacKinnon
> Brian Mastenbrook wrote
>>It's my suggestion that there is also something "going on" in
>>intelligence that is difficult to explain on the level of firing
>>neurons, just as there's something "going on" in computation that's
>>difficult to explain on the level of gates on a silicon wafer with
>>particular electromagnetic properties.
>
>
> i.e. emergence.
I don't think he means to suggests that there's anything mystical or
inexplicable about computation. We created it, and not by accident.
It's difficult to explain celestial mechanics on the level of
arithmetic. So we moved up the abstraction curve, invented calculus, and
lived happily ever after.
We haven't yet found the "calculus" that allows us to easily explain
intelligent reasoning. So intelligence is as mystifying to us as the
retrograde motion of the planets was to the ancients.
With the wrong theory it can seems as complex as the geocentric model of
the universe was, with 55 spheres required to allow the possibility of
the motion that was actually observed.
In your post you conflated the complex behaviour sometimes exhibited by
large numbers of simple things (emergence) with the simple behaviour
sometimes exhibited by people who should know better (throwing something
against the wall and seeing if it sticks). If something emerges from
experiment without planning, that's just luck.
Don Groves
wrote:
> Randy Yates <ya...@ieee.org> writes:
>
>> Has there been any significant advances lately in the area of sentient
>> algorithms? Can someone even define "sentient"?
>
> I've read the responses so far. Very interesting work is going on.
> However...
>
> I don't think there is any such thing as sentience in the sense of
> self-awareness. When Descartes said, "I think, therefore I am," he
> was mistaken. He was wrong at the part where he said, "I think."
> Everything else that followed was therefore wrong.
In my book, Richard Feynman said it best:
"I think, therefore I think I am."
--
dg
Erann Gat
ccwebster) net))> wrote:
I think I think, therefore I think I am.
I think.
:-)
E.
Don Groves
wrote:
When I was writing real-time embedded systems code
for a living, I made up this sign for my desk:
"I think, therefore I Asm".
But now I'm thinking at a higher level, I think.
--
dg
Rob Warnock
+---------------
| Don Groves <(. (@ dgroves ccwebster) net))> wrote:
| > > When Descartes said, "I think, therefore I am," he was mistaken.
| > > He was wrong at the part where he said, "I think."
| > > Everything else that followed was therefore wrong.
| >
| > In my book, Richard Feynman said it best:
| > "I think, therefore I think I am."
|
| I think I think, therefore I think I am.
Exactly so. The other way around is putting Descartes before the horse.[1]
+---------------
| I think.
+---------------
Yes, well...
-Rob
[1] Which I first heard from The Vajra Regent Osel Tendzin in 1985
(though it may not have been new then).
-----
Rob Warnock <rp...@rpw3.org>
627 26th Avenue <URL:http://rpw3.org/>
San Mateo, CA 94403 (650)572-2607
robbie carlton
> It's difficult to explain celestial mechanics on the level of
> arithmetic. So we moved up the abstraction curve, invented calculus, and
> lived happily ever after.
>
> We haven't yet found the "calculus" that allows us to easily explain
> intelligent reasoning. So intelligence is as mystifying to us as the
> retrograde motion of the planets was to the ancients.
Yes, calculus turned out to be a better language for describing
celestial mechanics than arithmetic, but for many phenomena there is
no calculus simpler than the description of it's constituent parts.
Take as a simple example Craig Reynolds' "Boids". The whole point of
that experiment (simulation?) is that the simplest way of describing
the behaviour of the flock is by describing the behaviour of the
individuals. It's not that Reynolds couldn't find an adequate
"calculus" to describe the flock. He found a spectacular one, it just
happened to involve emergent behaviour which is difficult to find by
analysing the equations.
> With the wrong theory it can seems as complex as the geocentric model of
> the universe was, with 55 spheres required to allow the possibility of
> the motion that was actually observed.
>
> In your post you conflated the complex behaviour sometimes exhibited by
> large numbers of simple things (emergence) with the simple behaviour
> sometimes exhibited by people who should know better (throwing something
> against the wall and seeing if it sticks). If something emerges from
> experiment without planning, that's just luck.
I appreciate that in the post that I quoted Brian Mastenbrook was
talking about students putting together random bits of code. However
he originally attacked the connectionist AI approach as Cargo-Cult
science, and then in the next post started asking for an emergent
description of Intelegence, which is why I felt the need to post.
Brian Mastenbrook
carlton <robbief...@hotmail.com> wrote:
> I appreciate that in the post that I quoted Brian Mastenbrook was
> talking about students putting together random bits of code. However
> he originally attacked the connectionist AI approach as Cargo-Cult
> science, and then in the next post started asking for an emergent
> description of Intelegence, which is why I felt the need to post.
Only a twisted reading of my statements would take them as a request
for an "emergent description of Intelligence", when in other posts I
have repeatedly talked about approaching intelligence on a level that
corresponds to philosophical primitives in the concept of intelligence.
Because I cannot say that neural nets have any such correspondence, I
cannot accept that they will become intelligent merely because they
happen to resemble the existing hardware of our brain.
Yes, it's wonderful that some systems can demonstrate wildly complex
behavior out of simple parts. However, the problem is that such
behavior is essentially unpredictable, which is not very helpful when
trying to do research. The general notion of science isn't just about
running and experiment and seeing what works - it's about being able to
draw conclusions from the experiment and modify one's thinking. If the
nature of the domain is that of gross unpredictability, this can't
happen.
André Thieme
> Atoms do not think. They are constrained by the laws of physics.
> Brains are just collections of atoms arranged in a fancy pattern.
> They are subject to the exact same laws.
Well, atoms alone (probably) don't think. However, when you combine
many of them into a special structure, then this structure gets new
properties/abilities which were not there before. For example the
abilitiy to think.
If we find out what pattern/structure is needed we could assemble
atoms in a computer in a similar way and by this giving it also the
ability to think. The hardware is already here.. now placing the
last parts of matter needs to be done by software.
Programming is nothing else than placing matter into a specific
pattern which creates some effect.
André
--
Paul Wallich
> Erann Gat <gNOS...@flownet.com> wrote:
> +---------------
> | Don Groves <(. (@ dgroves ccwebster) net))> wrote:
> | > David Steuber <da...@david-steuber.com> wrote:
> | > > When Descartes said, "I think, therefore I am," he was mistaken.
> | > > He was wrong at the part where he said, "I think."
> | > > Everything else that followed was therefore wrong.
> | >
> | > In my book, Richard Feynman said it best:
> | > "I think, therefore I think I am."
> |
> | I think I think, therefore I think I am.
Or, as one neurobiologist used to say, "I think the brain is the most
fascinating organ in the human body. Then I remember who's telling me that."
> +---------------
>
> Exactly so. The other way around is putting Descartes before the horse.[1]
>
> [1] Which I first heard from The Vajra Regent Osel Tendzin in 1985
> (though it may not have been new then).
(It was old enough in 1955 that Richard Armour could refer to it
obliquely in _It All Started With Europa_...)
paul
Christian Lynbech
André> Well, atoms alone (probably) don't think. However, when you combine
André> many of them into a special structure, then this structure gets new
André> properties/abilities which were not there before. For example the
André> abilitiy to think.
Approaching the nitpick level, but isn't this merely our current
theory?
I mean, we can not know for certain that intelligence does not involve
some magical component (such as a divinely given soul) until we have
succesfully built an artificial intelligence.
------------------------+-----------------------------------------------------
Christian Lynbech | christian #\@ defun #\. dk
------------------------+-----------------------------------------------------
Hit the philistines three times over the head with the Elisp reference manual.
- pet...@hal.com (Michael A. Petonic)
Brian Mastenbrook
<christia...@ericsson.com> wrote:
> Approaching the nitpick level, but isn't this merely our current
> theory?
>
> I mean, we can not know for certain that intelligence does not involve
> some magical component (such as a divinely given soul) until we have
> succesfully built an artificial intelligence.
I know it's true because I wouldn't want to live in a world where I
couldn't make an artificial intelligence. Therefore, I don't even have
to bother thinking about the possibility that I can't.
Ray Dillinger
>
> >>>>> "André" == André Thieme <this.address.is.goo...@justmail.de> writes:
>
> André> Well, atoms alone (probably) don't think. However, when you combine
> André> many of them into a special structure, then this structure gets new
> André> properties/abilities which were not there before. For example the
> André> abilitiy to think.
>
> Approaching the nitpick level, but isn't this merely our current
> theory?
>
> I mean, we can not know for certain that intelligence does not involve
> some magical component (such as a divinely given soul) until we have
> succesfully built an artificial intelligence.
Even that wouldn't prove anything, unless you are ready to make
some very conservative(!) assumptions about what magic(!) can
and cannot do.
Who's to say that a divinity with a sense of humor wouldn't give
a soul to an AI program? Why, or why not, do you presume you do
or don't know what $DIETY would or wouldn't do?
Who's to say that humans doing hard work on an AI program don't
pass on some of the "magic" with which they are infused? Why or
why not?
Basically, you can't say anything *isn't* magic, until you have
a working theory of magic that is able to rule things out. Since
the defining characteristic of most theories about magic is precisely
the lack of ability to rule things out, there's a problem with that.
Bear
(Have we achieved nitpick level yet?)
Bear
David Steuber
Look at it this way. How is a microprocessor controling a servo moter
any different from your brain controling your fingers?
Mathematically, AI should be no more or less possible than NI. I also
doubt very much that the two will be distinguishable by any reasonable
means. That is, the thing we perceive as intelligence will be just
that regardless of its origin.
The earlier post about arranging matter into a sufficiently complex
pattern so as to produce a mind is really no different than emulation
with a sufficiently complex program. It is just another Turing
transform. A more concrete example is the fact that a microprocessor
can be fully emulated and tested in software before being fabricated.
--
I wouldn't mind the rat race so much if it wasn't for all the damn cats.
Don Geddis
> I don't think there is any such thing as sentience in the sense of
> self-awareness.
Self-awareness is actually easier than sentience. Many computer systems
include a model of their own behavior, that they compare with observations
from the external world. That's sufficient to create a limited kind of
self-awareness.
> Atoms do not think. They are constrained by the laws of physics.
The laws of physics don't necessarily exclude thinking. But you're right
that atoms don't think.
> Brains are just collections of atoms arranged in a fancy pattern.
> They are subject to the exact same laws.
Your use of the word "just" is highly misleading. All the value is in the
organization.
In any case, look up the "systems reply" to Searle's Chinese Room analogy.
I'm sure you would be a big fan of Searle (who is an AI critic). But the
"systems reply" addresses your confusion head-on. Namely, that the overall
system can have interesting properties which none of the constituent parts
have in isolation.
-- Don
_______________________________________________________________________________
Don Geddis http://don.geddis.org/ d...@geddis.org
John Thingstad
> David Steuber <da...@david-steuber.com> wrote on 11 Apr 2004 18:0:
>> I don't think there is any such thing as sentience in the sense of
>> self-awareness.
>
> Self-awareness is actually easier than sentience. Many computer systems
> include a model of their own behavior, that they compare with
> observations
> from the external world. That's sufficient to create a limited kind of
> self-awareness.
I belive the tecnical term for self awareness is auto-epistemic logic.
See More, R. C. Autoepistemic logic,
in Non-Standard Logics for Automated Reasoning.
>
>> Atoms do not think. They are constrained by the laws of physics.
>
> The laws of physics don't necessarily exclude thinking. But you're right
> that atoms don't think.
>
>> Brains are just collections of atoms arranged in a fancy pattern.
>> They are subject to the exact same laws.
You are assuming that the fact that there is a description of the phenomena
means that there is a deterministic model for it.
In fact Quantum Electrodynamice can give rise highly nonliear behaviour.
You could say that caos is the rule and order the exception that lasts
over time. Roger Penrose postulates, in Shaddows of the Mind, that the
bioelectric field surrounding the brain is in a quantum collaped state.
(Solenoid effect in the micro-tubules)
He goes on to conjecture that quantum effects are neccesary to model the
mind.
Thus he submises a computer does not have the capacity to medol it.
Personally I see no reason why computers can't model nonlinear systems.
(There are some limitations. In the interval delta t being modeled the
function
must be bound polynoninally or no number of digits suffice to express the
number.)
Personally I have been looking for neural "loopback" and adjusting for the
error term.
A method called error biffucation.
(given linear system Ax=b where x is unknown we know x + delta x where
delta x is a
unknown error. but A(x+delta x) = b + delta b
so X delta x = delta b which we can solve. substituting gives
A delta X = A (x + delta x) - b)
It seems to me this can be used to model recognition.
Futher more a new mathematical tool called "Integrate and fire pulse
coupled oscillation"
may be used to modell colation.
I thus summise that the brain is a fractally composed set of non-linear
system solvers using
error biffucation and colating using integrate and fire pulse coupled
oscillation.
The "prime" integrator (factal tree root) is what we call conciousness.
Well, sigh, it needs more work.
--
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/
André Thieme
>>>>>>"André" == André Thieme <this.address.is.goo...@justmail.de> writes:
>
>
> André> Well, atoms alone (probably) don't think. However, when you combine
> André> many of them into a special structure, then this structure gets new
> André> properties/abilities which were not there before. For example the
> André> abilitiy to think.
>
> Approaching the nitpick level, but isn't this merely our current
> theory?
>
> I mean, we can not know for certain that intelligence does not involve
> some magical component (such as a divinely given soul) until we have
> succesfully built an artificial intelligence.
A magic component would not change anything, because we would emulate it
too. If magic exists we can look at two aspects of it:
1) algorithmic component
2) randomness
To 1):
If the laws of magic are not 100% random it means they have more than 0%
of "law" inside. When magic follows laws we can exactly tell how magic
would behave under certain circumstances. Therefor we could emulate the
algorithmic component of magic with our computers.
Most of our universe follows this way. Only Heisenbergs uncertainty and
some few physical phenomenons produce something which cannot be computed
by algorithms.
To 2):
If the magic has some random parts (and is not full driven by "laws") we
can emulate these components too. We have sources to create real random
input (for our programs).
The random has to be like real random, if not it would have a pattern
and would automatically become computable. So any source of random will
be enough to emulate the randomness of magic.
This way, even if magic exists (this would be some behaviour of our
universe which we don't know good enough yet) it will not stop us from
creating real intelligence. We could emulate all components of magic
either by software or from some external generator of randomness.
We would have to find out all parts of our brain which follow the
predictable "laws of magic" and programm these rules.
In all other situations we would give random input into our programm.
I don't say that this is the easiest thing to do, which has ever be done
on our planet. Probably it is one of the most complex things which could
be done in our world :)
However, I only want to say that it logically must possible to create
intelligence.
André
--
Randy Yates
> [...]
> What, exactly, is the cargo cult's mistake? Seabees came and built
> runways and control towers, men put funny headphones on, and the
> planes came. Mimicking this looks ridiculous TO US, because we have
> superior domain knowledge of aeronautics.
Could this be one definition of sentience, then: The ability to
transfer knowledge from outside the identity's domain to inside?
> Absent that knowledge, we could say that they err in persisting for
> five years when no planes have come, but that suggests that the
> cultists are to somehow know how long to wait for results. Given
> that the original was a one-time event for a culture which (I'm
> guessing) had an oral history stretching back many generations, even
> a several-generations-long experiment might not be unreasonable.
This has a strange, vague similarity to scientific experiment. Because
we've observed that F = m*a a google times before doesn't mean that it
might not change tomorrow. Experience does not equal knowledge.
> [If you'd have asked me yesterday whether I'd defend cargo cultists,
> I'd have been insulted at the implication. I never really thought
> about it before, just thought what Feynman wanted me to think.]
>
> I meet people all the time who seem to have only the most tenuous
> grasp of logic. In fact, I'd hazard that, for the majority of any
> given society, Aristotle and Descartes might as well never have
> existed. It's not that they are completely illogical, but they
> certainly don't operate through rigourous reasoning, and often can't
> see the flaws in the systems they believe.
>
> So the interesting question is, when seeking artificially intelligent
> behaviour, are we looking for computers to act like the best of us (in
> which case, perhaps "Artificial Brilliance"?) or the rest of us?
This is an interesting query, but really off-topic from the question.
I think even a retarded sentient algorithm would be a breakthrough.
> [...]
--
% Randy Yates % "The dreamer, the unwoken fool -
%% Fuquay-Varina, NC % in dreams, no pain will kiss the brow..."
%%% 919-577-9882 %
%%%% <ya...@ieee.org> % 'Eldorado Overture', *Eldorado*, ELO
http://home.earthlink.net/~yatescr
Randy Yates
Just lovely. I'm still waiting for a definition of sentience. How can
we know we've achieved it if we can't define it?
--
% Randy Yates % "Remember the good old 1980's, when
%% Fuquay-Varina, NC % things were so uncomplicated?"
%%% 919-577-9882 % 'Ticket To The Moon'
%%%% <ya...@ieee.org> % *Time*, Electric Light Orchestra
http://home.earthlink.net/~yatescr
Matthew Danish
> Just lovely. I'm still waiting for a definition of sentience. How can
> we know we've achieved it if we can't define it?
I don't know if I think, but I know that I am.
Take that, Descartes.
--
; Matthew Danish <mda...@andrew.cmu.edu>
; OpenPGP public key: C24B6010 on keyring.debian.org
; Signed or encrypted mail welcome.
; "There is no dark side of the moon really; matter of fact, it's all dark."
John Thingstad
>
> Just lovely. I'm still waiting for a definition of sentience. How can
> we know we've achieved it if we can't define it?
We can't. I think this term is ill defined.
Sentience is supposed to be what distinguises uf from animals.
I think research, if anything, has made this distinction more fluid.
Maybe the answer is to remove the term ;)
Randy Yates
> [...]
> Maybe the answer is to remove the term ;)
A typical liberal, "free-thinking," new-age suggestion. Also,
as are most such suggestions, complete bullshit.
Artem Baguinski
>>>>> "Randy" == Randy Yates <ya...@ieee.org> writes:
Randy> Has there been any significant advances lately in the area
Randy> of sentient algorithms? Can someone even define "sentient"?
David> I've read the responses so far. Very interesting work is
David> going on. However...
David> I don't think there is any such thing as sentience in the
David> sense of self-awareness. When Descartes said, "I think,
David> therefore I am," he was mistaken. He was wrong at the part
David> where he said, "I think." Everything else that followed
David> was therefore wrong.
you're saying "he was" therefore you mean that "what followed" was
true.
and a false premise doesn't make a conclusion ("what followed")
false. e.g.:
descartes thinks implies 2+2=4 ;; questionable
descartes thinks ;; false (you say)
therefore
2+2=4 ;; true
further, your phrase "I don't think" doesn't make sence then, or
are you in any way better then Descartes? ;-)
of course there is such thing as self awareness: whenever you say
"I ..." you mean yourself as you aware of it. It may be illusory,
it may have no corresponding physical reality, but nevertheless
you're aware of it. it's like saying there's no hallucinations or
dreams - well of course they are, i saw them.
--
gr{oe|ee}t{en|ings}
artm
Cameron MacKinnon
> Cameron MacKinnon <cmack...@clearspot.net> writes:
>>So the interesting question is, when seeking artificially intelligent
>>behaviour, are we looking for computers to act like the best of us (in
>>which case, perhaps "Artificial Brilliance"?) or the rest of us?
>
> This is an interesting query, but really off-topic from the question.
> I think even a retarded sentient algorithm would be a breakthrough.
I think that once we get intelligent machines, creating one that acts
plausibly stupid or ignorant (like us humans) will be even more of a
breakthrough. The first AIs are likely to be annoying know-it-alls.
As to the original question, how about "sentient: able to use Google or
a dictionary to discover the meaning of unknown words."
I say any animal that can learn to differentiate between seeing another
animal and seeing itself in a mirror is sentient.
David Steuber
> further, your phrase "I don't think" doesn't make sence then, or
> are you in any way better then Descartes? ;-)
Until Descartes comes along to refute what I say, yes. Of course,
Descartes is dead. So I've got at least one thing going in my favor.
> of course there is such thing as self awareness: whenever you say
> "I ..." you mean yourself as you aware of it. It may be illusory,
> it may have no corresponding physical reality, but nevertheless
> you're aware of it. it's like saying there's no hallucinations or
> dreams - well of course they are, i saw them.
I wasn't aware of that. I'm just using English. Maybe there exists
some language spoken by people which does not have any concept of self
reference.
Anyway, I don't see why we can't come up with some form of synthetic
intelligence. All that is really required (not to imply that this is
easy) is to take a set of defined cognitive functions that we can test
for in humans and figure out a way to emulate those functions by
machine. The emulation does not have to be precise or anything. It
just has to be able to pass the tests.
This is essentially what Turing said, although I think he copped out
on the nature of the testing. People are easy to fool.
So the first challenge is to create a set of defined cognitive
functions along with tests that unambiguously determin the existence
and extent of those functions. Once a set of functions and tests is
agreed upon, it is then just a matter of time before a machine can be
made to pass the tests.
There are of course many specific problems that have already been
identified. Speach recognition is a big one. Lots of progress has
been made there. Visual recognition is somewhat harder, but there are
some results in that area.
Creativity would be harder I think. We don't have a good definition
for that (test) that I know of. Trying to make a machine compose
pleasing music is probably too ill defined a goal to shoot for.
I think the act of creating good tests will prove to be just about as
difficult as creating algorithms that can pass those tests. The
problem lies in defining the tests.
Christian Lynbech
David> Look at it this way. How is a microprocessor controling a servo moter
David> any different from your brain controling your fingers?
David> Mathematically, AI should be no more or less possible than NI.
But if intelligence was intimately connected to the presence of a
divine soul, it would not be mathematical. Then all machines that we
would be able to build would lack a component that only the divine
being would be able to control and put into things.
Fred Gilham
> Anyway, I don't see why we can't come up with some form of synthetic
> intelligence. All that is really required (not to imply that this
> is easy) is to take a set of defined cognitive functions that we can
> test for in humans and figure out a way to emulate those functions
> by machine. The emulation does not have to be precise or anything.
> It just has to be able to pass the tests.
An analogy that might indicate that things aren't as simple as they
seem is the difference between "cookbook" math and true
understanding. We see that there is a difference even though for a
particular problem the result would be the same.
An "algorithmic intelligence" would be analogous to a cookbook
understanding of math.
Another argument, which I feel is devastating to hard AI, is the
"hermeneutical hall of mirrors" argument, wherein the point is made
that even when machines appear to behave intelligently, they only do
so because we project meanings onto their behavior. The intelligence
they appear to manifest is our own reflected back to us.
--
Fred Gilham gil...@csl.sri.com
Early in our marriage, 40-some years ago, Mrs. Williams would return
from shopping complaining about the unreasonable prices. Having aired
her complaints, she'd then ask me to unload her car laden with
purchases. After the unloading, I'd ask her: "I thought you said the
prices were unreasonable. Why did you buy them? Are you
unreasonable? Only an unreasonable person would pay unreasonable
prices." The discussion always headed downhill after such an
observation. -- Walter Williams, economist
Paul Wallich
>>Anyway, I don't see why we can't come up with some form of synthetic
>>intelligence. All that is really required (not to imply that this
>>is easy) is to take a set of defined cognitive functions that we can
>>test for in humans and figure out a way to emulate those functions
>>by machine. The emulation does not have to be precise or anything.
>>It just has to be able to pass the tests.
>
>
> An analogy that might indicate that things aren't as simple as they
> seem is the difference between "cookbook" math and true
> understanding. We see that there is a difference even though for a
> particular problem the result would be the same.
>
> An "algorithmic intelligence" would be analogous to a cookbook
> understanding of math.
>
> Another argument, which I feel is devastating to hard AI, is the
> "hermeneutical hall of mirrors" argument, wherein the point is made
> that even when machines appear to behave intelligently, they only do
> so because we project meanings onto their behavior. The intelligence
> they appear to manifest is our own reflected back to us.
The hall of mirrors argument, imo, is irremediably tainted because it
works in the opposite direction as well: once you've decided that some
entity isn't behaving intelligently (in the sense of "human"
intelligence") there is no behavior the entity can exhibit that will
necessarily make you change your mind. A little historical research will
find plenty of essays in which the author argues that some subset of h.
sap. (said subset not including the author) may exhibit a remarkably
cunning simulation of "real" intelligence but is not in fact capable of
reasoning.
paul
Karl A. Krueger
> Another argument, which I feel is devastating to hard AI, is the
> "hermeneutical hall of mirrors" argument, wherein the point is made
> that even when machines appear to behave intelligently, they only do
> so because we project meanings onto their behavior. The intelligence
> they appear to manifest is our own reflected back to us.
Considering the number of computer users who appear to believe that
Microsoft Windows is smarter than themselves, I suspect at least some
of those are funhouse mirrors.
--
Karl A. Krueger <kkru...@example.edu>
Woods Hole Oceanographic Institution
Email address is spamtrapped. s/example/whoi/
"Outlook not so good." -- Magic 8-Ball Software Reviews
Brian Mastenbrook
<gil...@snapdragon.csl.sri.com> wrote:
> > Anyway, I don't see why we can't come up with some form of synthetic
> > intelligence. All that is really required (not to imply that this
> > is easy) is to take a set of defined cognitive functions that we can
> > test for in humans and figure out a way to emulate those functions
> > by machine. The emulation does not have to be precise or anything.
> > It just has to be able to pass the tests.
>
> An analogy that might indicate that things aren't as simple as they
> seem is the difference between "cookbook" math and true
> understanding. We see that there is a difference even though for a
> particular problem the result would be the same.
>
> An "algorithmic intelligence" would be analogous to a cookbook
> understanding of math.
>
> Another argument, which I feel is devastating to hard AI, is the
> "hermeneutical hall of mirrors" argument, wherein the point is made
> that even when machines appear to behave intelligently, they only do
> so because we project meanings onto their behavior. The intelligence
> they appear to manifest is our own reflected back to us.
I only think you think because I project meaning onto your behavior. In
reality, you don't, and I can now proceed to completely ignore you.
I guess this raises the question - if an AI gets offended at you and
the other Searlites' comments, and wipes out all of humanity, who's
left to say it doesn't think? (Or, less violently, if we send an AI on
a robotic spacecraft to explore other solar systems, and Earth is wiped
out by the Rogue Comet of Doom, who's left to play the Searlite
semantic tricks?)
Artem Baguinski
David> Artem Baguinski <ar...@v2.nl> writes:
>> of course there is such thing as self awareness: whenever you
>> say "I ..." you mean yourself as you aware of it. It may be
>> illusory, it may have no corresponding physical reality, but
>> nevertheless you're aware of it. it's like saying there's no
>> hallucinations or dreams - well of course they are, i saw them.
David> I wasn't aware of that. I'm just using English. Maybe
David> there exists some language spoken by people which does not
David> have any concept of self reference.
self reference in language merely reflects self awareness of a
mind. self is one of the little number of things that we seem to
have direct awareness of.
David> Anyway, I don't see why we can't come up with some form of
David> synthetic intelligence.
i didn't mean it's impossible. i'm not even sure self awareness
is all that important for synthetic intelligence. i just find it
erroneous to use first person pronoun when refusing to accept
self awareness ;-)
Randy Yates
> Randy Yates wrote:
>> Cameron MacKinnon <cmack...@clearspot.net> writes:
>>>So the interesting question is, when seeking artificially intelligent
>>>behaviour, are we looking for computers to act like the best of us (in
>>>which case, perhaps "Artificial Brilliance"?) or the rest of us?
>> This is an interesting query, but really off-topic from the
>> question. I think even a retarded sentient algorithm would be a
>> breakthrough.
>
> I think that once we get intelligent machines, creating one that acts
> plausibly stupid or ignorant (like us humans) will be even more of a
> breakthrough. The first AIs are likely to be annoying know-it-alls.
>
> As to the original question, how about "sentient: able to use Google
> or a dictionary to discover the meaning of unknown words."
I meant a "scientific" definition - one that would serve to guide as
a test.
> I say any animal that can learn to differentiate between seeing
> another animal and seeing itself in a mirror is sentient.
Weak. This could be done with simple pattern recognition. You're way
off from where I was thinking.
--
% Randy Yates % "With time with what you've learned,
%% Fuquay-Varina, NC % they'll kiss the ground you walk
%%% 919-577-9882 % upon."
%%%% <ya...@ieee.org> % '21st Century Man', *Time*, ELO
http://home.earthlink.net/~yatescr
Randy Yates
> [...]
> of course there is such thing as self awareness: whenever you say
> "I ..." you mean yourself as you aware of it. It may be illusory,
> it may have no corresponding physical reality, but nevertheless
> you're aware of it. it's like saying there's no hallucinations or
> dreams - well of course they are, i saw them.
Bingo.
--
% Randy Yates % "...the answer lies within your soul
%% Fuquay-Varina, NC % 'cause no one knows which side
%%% 919-577-9882 % the coin will fall."
%%%% <ya...@ieee.org> % 'Big Wheels', *Out of the Blue*, ELO
http://home.earthlink.net/~yatescr
Randy Yates
> Artem Baguinski <ar...@v2.nl> writes:
>
>> further, your phrase "I don't think" doesn't make sence then, or
>> are you in any way better then Descartes? ;-)
>
> Until Descartes comes along to refute what I say, yes. Of course,
> Descartes is dead. So I've got at least one thing going in my favor.
>
>> of course there is such thing as self awareness: whenever you say
>> "I ..." you mean yourself as you aware of it. It may be illusory,
>> it may have no corresponding physical reality, but nevertheless
>> you're aware of it. it's like saying there's no hallucinations or
>> dreams - well of course they are, i saw them.
>
> I wasn't aware of that. I'm just using English. Maybe there exists
> some language spoken by people which does not have any concept of self
> reference.
Concepts are not defined by language. I think what you meant was
a word or phrase that means self awareness.
> Anyway, I don't see why we can't come up with some form of synthetic
> intelligence.
I thought we already had. It may be low-rent, but we've got something
(the Turing test takers, e.g.).
The point of my query, if it isn't clear, is to motivate thought on
how to make a *truly* intelligent machine - no B.S. - one that in
every sense of the notion "intelligent" passes, and one that truly
*is* intelligent rather that one that merely has the appearance
of intelligence.
But I use the term "intelligent" here in a colloquial sense. To translate
that to a more scientific and precise notion (but probably still way off),
such a machine would need to be both "sentient" and "reasoning."
> I think the act of creating good tests will prove to be just about as
> difficult as creating algorithms that can pass those tests. The
> problem lies in defining the tests.
I think you may be right. Defining "sentient" and "reasoning" is step
one on the road to defining tests.
--
% Randy Yates % "I met someone who looks alot like you,
%% Fuquay-Varina, NC % she does the things you do,
%%% 919-577-9882 % but she is an IBM."
%%%% <ya...@ieee.org> % 'Yours Truly, 2095', *Time*, ELO
http://home.earthlink.net/~yatescr
Randy Yates
>>>>>> "David" == David Steuber <da...@david-steuber.com> writes:
>
> David> Look at it this way. How is a microprocessor controling a servo moter
> David> any different from your brain controling your fingers?
>
> David> Mathematically, AI should be no more or less possible than NI.
>
> But if intelligence was intimately connected to the presence of a
> divine soul, it would not be mathematical. Then all machines that we
> would be able to build would lack a component that only the divine
> being would be able to control and put into things.
Hi Christian,
This is a great point. I would say that a "soul" is essentially
a "will." We may eventually be able to make machines that are
both "sentient" and "reasoning" (refer to a parallel post I just
made), but without a "soul" such an entity's will, if existent,
must be synthetic (e.g., "protect all humans from harm"). At least
that's the way it seems to me.
--
Randy Yates
> [...]
> This is a great point. I would say that a "soul" is essentially
> a "will." We may eventually be able to make machines that are
> both "sentient" and "reasoning" (refer to a parallel post I just
> made), but without a "soul" such an entity's will, if existent,
> must be synthetic (e.g., "protect all humans from harm"). At least
> that's the way it seems to me.
I also should comment that I've wondered if, like our Creator has
done for us, there may be a way for us to infuse into our created
machine a piece of ourselves that would give it this "soul" or will.
--
% Randy Yates % "Though you ride on the wheels of tomorrow,
%% Fuquay-Varina, NC % you still wander the fields of your
%%% 919-577-9882 % sorrow."
%%%% <ya...@ieee.org> % '21st Century Man', *Time*, ELO
http://home.earthlink.net/~yatescr
David Steuber
> Another argument, which I feel is devastating to hard AI, is the
> "hermeneutical hall of mirrors" argument, wherein the point is made
> that even when machines appear to behave intelligently, they only do
> so because we project meanings onto their behavior. The intelligence
> they appear to manifest is our own reflected back to us.
So when I see behavior in other people, the intelligence they appear
to manifest is my own reflected back to me?
David Steuber
> David> Anyway, I don't see why we can't come up with some form of
> David> synthetic intelligence.
>
> i didn't mean it's impossible. i'm not even sure self awareness
> is all that important for synthetic intelligence. i just find it
> erroneous to use first person pronoun when refusing to accept
> self awareness ;-)
I see. So you see a reflection of what you believe is your self
awareness reflected in my use of the first person pronoun?
It's time I fessed up. I'm a FORTRAN program written in the late
60's. Does that mean I pass the Turing test? The secret is in the
random number generator used to drive unpredictable behavior. I have
a thermal sensor in a nice hot cup of tea.
I'll admit that my awareness feels convincing at times. But only at
times.
André Thieme
> Christian Lynbech <christia...@ericsson.com> writes:
>
>
>>>>>>>"David" == David Steuber <da...@david-steuber.com> writes:
>>
>>David> Look at it this way. How is a microprocessor controling a servo moter
>>David> any different from your brain controling your fingers?
>>
>>David> Mathematically, AI should be no more or less possible than NI.
>>
>>But if intelligence was intimately connected to the presence of a
>>divine soul, it would not be mathematical. Then all machines that we
>>would be able to build would lack a component that only the divine
>>being would be able to control and put into things.
>
>
> Hi Christian,
>
> This is a great point. I would say that a "soul" is essentially
> a "will." We may eventually be able to make machines that are
> both "sentient" and "reasoning" (refer to a parallel post I just
> made), but without a "soul" such an entity's will, if existent,
> must be synthetic (e.g., "protect all humans from harm"). At least
> that's the way it seems to me.
I don't understand what this "soul" would be doing with us.
It is interesting to see, that some people believe that there is
something like a soul.
Even today most brain-scientists are very sure that there is no such
thing like a free will. I would not be overly amazed if they can prove
it in 20 years.
On some beeings you can obviously see their "algorithmic" behaviour,
which has nothing to do with a soul. There is for example some fish
which is doing some "dance" whenever he sees a girl-fish. In an
experiment they showed a male fish a dummy of a female fish and he
started with his exact dance moves, and he did this always when the
dummy was showed to him. For such complex life forms like humans it is
nearly impossible for us to see structures (besides reflexes).
However, only because the human beeing is very complicated it does not
mean that we have some thing like a soul.
See also my other posting that explains why a soul can also not exist,
or, why we will also be able to program one. A soul is also nothing but
a "computer program".
André
--
David Steuber
> >>>>> "David" == David Steuber <da...@david-steuber.com> writes:
>
> David> Look at it this way. How is a microprocessor controling a servo moter
> David> any different from your brain controling your fingers?
>
> David> Mathematically, AI should be no more or less possible than NI.
>
> But if intelligence was intimately connected to the presence of a
> divine soul, it would not be mathematical. Then all machines that we
> would be able to build would lack a component that only the divine
> being would be able to control and put into things.
That's a rather big if. I might be able to go along with that if if I
ever saw anything that violated mathematical law or convinced me that
the brain is not just another Turing machine. My own personal
experience just doesn't support any such conclusion.
Of course this is just my opinion and my experience.
Randy Yates
> Randy Yates wrote:
>
>> Christian Lynbech <christia...@ericsson.com> writes:
>>
>>>>>>>>"David" == David Steuber <da...@david-steuber.com> writes:
>>>
>>>David> Look at it this way. How is a microprocessor controling a servo moter
>>>David> any different from your brain controling your fingers?
>>>
>>>David> Mathematically, AI should be no more or less possible than NI.
>>>
>>>But if intelligence was intimately connected to the presence of a
>>>divine soul, it would not be mathematical. Then all machines that we
>>>would be able to build would lack a component that only the divine
>>>being would be able to control and put into things.
>> Hi Christian,
>> This is a great point. I would say that a "soul" is essentially
>> a "will." We may eventually be able to make machines that are
>> both "sentient" and "reasoning" (refer to a parallel post I just
>> made), but without a "soul" such an entity's will, if existent,
>> must be synthetic (e.g., "protect all humans from harm"). At least
>> that's the way it seems to me.
>
> I don't understand what this "soul" would be doing with us.
As I already intimated, it directs our will. It's the "top-level loop"
of our algorithm.
> It is interesting to see, that some people believe that there is
> something like a soul.
> Even today most brain-scientists are very sure that there is no such
> thing like a free will.
Most scholars were sure the sun revolved around the earth.
> I would not be overly amazed if they can prove
> it in 20 years.
> On some beeings you can obviously see their "algorithmic" behaviour,
> which has nothing to do with a soul. There is for example some fish
> which is doing some "dance" whenever he sees a girl-fish. In an
> experiment they showed a male fish a dummy of a female fish and he
> started with his exact dance moves, and he did this always when the
> dummy was showed to him. For such complex life forms like humans it is
> nearly impossible for us to see structures (besides reflexes).
>
> However, only because the human beeing is very complicated it does not
> mean that we have some thing like a soul.
> See also my other posting that explains why a soul can also not exist,
> or, why we will also be able to program one. A soul is also nothing but
> a "computer program".
Had you couched this in more careful terms, I would have thought you
simply misguided rather than on a mission. To each his opinion.
--
% Randy Yates % "Midnight, on the water...
%% Fuquay-Varina, NC % I saw... the ocean's daughter."
%%% 919-577-9882 % 'Can't Get It Out Of My Head'
%%%% <ya...@ieee.org> % *El Dorado*, Electric Light Orchestra
http://home.earthlink.net/~yatescr
Don Groves
wrote:
> It's time I fessed up. I'm a FORTRAN program written in the late
> 60's. Does that mean I pass the Turing test?
Need more information to answer that:
Fortran II or IV?
What hardware are you running on?
Most important - who wrote you?
--
dg
David Steuber
> I also should comment that I've wondered if, like our Creator has
> done for us, there may be a way for us to infuse into our created
> machine a piece of ourselves that would give it this "soul" or will.
See Mary Shelly's Frankenstein. I believe the idea was that the good
doctor could not create life per se, but he could put a bit of his
soul into it.
I've more or less discounted the existence of a divine creator. But
there is a certain appeal to a sci-fi idea that we are simply a part
of the universe that has broken itself apart into many pieces in an
attempt to understand its self. Not that I actually subscribe to that
idea.
Set aside the philosophy and theology of the issue, I think the
apparent existence of the mind to be the most fascinating thing about
the universe. I don't think The Big Bang holds a candle to it by
comparison.
Unless the laws of mathematics are mutable, it seems that everything
is the way it is without the need for any kind of creation event. If
we solve the problem of creating a machine mind that is as good or
better than the human mind, I expect there will be some controversy.
How long it will take to do that I don't know. It may also be that we
can never fully understand it. I already find it to be quite magical
that the letter 'e' shows up on my display when I, or someone far away,
types an 'e'. Yet all that is has been engineered. Math has not. So
far as we know, it is only discovered.
One of the coolest things I've read was in Carl Sagan's book Contact.
I made sure to read it before the movie came out. In it, there is a
rather interesting discussion about the constant Pi. As far as I
know, Pi can't be anything other than what it is. But in the book,
there is a "message" buried deep down in the number. I won't spoil
it here. Read the book. It is better than the movie.
There is an exit strategy for people who wish to cling to a creator.
In physics, math is really a tool for creating models. The models are
no good until they have been confirmed by experimental observation.
Even then, if you buy into Hume, that empirical evidence is of limited
value. Math may be immutable. But math != physics. Math is just a
tool.
I leave this as a small room for doubt in my own philosophical world
view. That said, I think Roger Penrose is seriously stretching things
in his book, "The Emperor's New Mind." I find his arguments against
the possibility of AI/SI whatever you care to call it (I prefer
synthetic over artificial because the synthetic is indistinguishable
from the natural) to fall strongly into the wishful thinking camp.
Currently the weight of scientific evidence that I am familiar with
leads me to conclude that a human-like synthetic intelligence is quite
possible and perhaps even likely. It just requires deeper
understanding on our part to achieve it. I think once that happens,
the sales of anti-depressants will truly skyrocket.
My only real fear is that people will trust synthetic intelligence
more than the real thing. Just because the machines will be smarter
doesn't mean they will be infallible.
"Teach it phenomenology" --- Dark Star
Ray Dillinger
>
> I've more or less discounted the existence of a divine creator. But
> there is a certain appeal to a sci-fi idea that we are simply a part
> of the universe that has broken itself apart into many pieces in an
> attempt to understand its self. Not that I actually subscribe to that
> idea.
I've sort of got the opinion that there are things which
are holy -- family, love, truth, compassion, joy, life,
etc... This doesn't really require or involve the kind
of God who is a "who", or even a God separate in any way
from the universe itself. It just *is*. Every religion
in the world is just a shadow of what you hold in your
heart when you percieve the world in a state of reverence.
What's holy is that which we hold in reverential regard.
So, each of us chooses what is holy, and for some folks
it's everything, and for some it's nothing.
Not that this has much to do with Lisp, but perhaps on
some level this sort of capacity is part of what we
regard as sentience -- the ability to self-consciously
choose, in some matters, our attitudes and how we allow
things to affect us.
Bear
Artem Baguinski
artm> of course there is such thing as self awareness: whenever
artm> you say "I ..." you mean yourself as you aware of it. It may
artm> be illusory, it may have no corresponding physical reality,
artm> but nevertheless you're aware of it. it's like saying
artm> there's no hallucinations or dreams - well of course they
artm> are, i saw them.
>>>>> "David" == David Steuber <da...@david-steuber.com> writes:
David> I wasn't aware of that. I'm just using English. Maybe
David> there exists some language spoken by people which does not
David> have any concept of self reference.
>>>>> "Randy" == Randy Yates <ya...@ieee.org> writes:
Randy> Concepts are not defined by language. I think what you
Randy> meant was a word or phrase that means self awareness.
David> Anyway, I don't see why we can't come up with some form of
David> synthetic intelligence.
Randy> I thought we already had. It may be low-rent, but we've got
Randy> something (the Turing test takers, e.g.).
Randy> The point of my query, if it isn't clear, is to motivate
Randy> thought on how to make a *truly* intelligent machine - no
Randy> B.S. - one that in every sense of the notion "intelligent"
Randy> passes, and one that truly *is* intelligent rather that one
Randy> that merely has the appearance of intelligence.
How could you possibly distinguish one from another?
Suppose there's "scientific definition" of "sentinent" and a test
for sentinence based on that definition. Wouldn't it be possible
to create a machine which merely has the appearance of
intelligence which does it best passing this test? we assume other
people are intelligent because they appear intelligent to us,
there's no way we can test if they are sentinent/self aware,
because of the "special access" sentinent being has to its own
mental states and events. you can't prove anything but yourself
has such "special access" like you can't know, how would it feel
to be something else.
Randy> But I use the term "intelligent" here in a colloquial
Randy> sense. To translate that to a more scientific and precise
Randy> notion (but probably still way off), such a machine would
Randy> need to be both "sentient" and "reasoning."
sentinent and reasoning intelligent being is antopomophic
intelligent being, and yes human intelligence is the only one we
know. it doesn't mean intelligent being must be sentinent.
the reason we need smarter machines is to help us solve our
problems. i'm not sure sentinence is all that important for
problem solving and trying to create Artificial Sentinence
distracts us from the practical goals.
once we've created artificial sentinent beings they might come up
with their own problems to solve, e.g.: equal rights,
artificial-beings marriage etc.
David> I think the act of creating good tests will prove to be
David> just about as difficult as creating algorithms that can
David> pass those tests. The problem lies in defining the tests.
Randy> I think you may be right. Defining "sentient" and
Randy> "reasoning" is step one on the road to defining tests.
--
Gorbag
"David Steuber" <da...@david-steuber.com> wrote in message
news:87ptabi...@david-steuber.com...
> My only real fear is that people will trust synthetic intelligence
> more than the real thing. Just because the machines will be smarter
> doesn't mean they will be infallible.
Just because something isn't infallible doesn't mean it is not more worthy
of trust. 53% right is still better than 42%.
Gorbag
"Randy Yates" <ya...@ieee.org> wrote in message
news:ad1fmo...@ieee.org...
> David Steuber <da...@david-steuber.com> writes:
> > I wasn't aware of that. I'm just using English. Maybe there exists
> > some language spoken by people which does not have any concept of self
> > reference.
>
> Concepts are not defined by language.
If concepts are not defined by language, just what is language?
Concepts may or may not be limited by language, but there is strong evidence
that for a given culture and population, those concepts that are not
conveyable in the language are, in fact, not conceived by representative
members. New concepts generates new language to convey the concept, at least
amoung a given (sub-)population.
André Thieme
>>It is interesting to see, that some people believe that there is
>>something like a soul.
>>Even today most brain-scientists are very sure that there is no such
>>thing like a free will.
>
> Most scholars were sure the sun revolved around the earth.
Here the issue is a bit different.
The scientists started out asuming that we have a free will.
Then they looked more detailed on the brain and came to the conclusion
that there is no place for a free will.
Anyway, free will is not possible as it would break "laws" of nature.
The molecules in your brain are forced to behave in a special way. So
they "force" you to act in a specific pattern (there is no will at all,
no person/intelligence who is deciding the pattern and making descisions
how you should act). If you had a free will you could change the way
how the molecules in your body assemble them self.
Although particle X has to move down because it was pushed into this
direction by another particle (or molecule, whatever) your will forces
it against physical principles to move into a different direction,
enabeling you to say what you "really" want - you would have a free will.
> Had you couched this in more careful terms, I would have thought you
> simply misguided rather than on a mission. To each his opinion.
I am sorry, my english is not enough developed...
I hope it will become better over time, so that my words sound less
offensive to others.
André
--
Gorbag
"Fred Gilham" <gil...@snapdragon.csl.sri.com> wrote in message
news:u71xmro...@snapdragon.csl.sri.com...
> Another argument, which I feel is devastating to hard AI, is the
> "hermeneutical hall of mirrors" argument
Feh; this argument can be used to support solipsism as well. What's the
point?
André Thieme
> I've more or less discounted the existence of a divine creator.
I am not so certain about this. There are some people who have very
good arguments that we were created by an intelligence. If this
intelligence is a living creature, or if it is some strange kind of
"nature laws" is not sure.
But for example take it this way:
if you get some kilos of sand from the beach, climb on a high building
and then let the sand fall down... what do you expect of the sand?
You probably expect that it will fall down somehow, without any
specific pattern. I am sure that you would be highly confused if you
go downstairs and find how the sand has written the faculty function
in Lisp. How often you might throw around sand, it will never fall in a
way that it writes Lisp programs. This is randomness.
Another example:
in a school you give the pupils some dice, and every pupil has to throw
it 100 times and write down the number.
If one kid has 100 times on his paper a "3" you would not believe him
that he did it correctly. Because 100 times a 3 is information, not
randomness. An intelligence is needed to create information.
Last example:
You maybe heared that in the 1925 some people found eggs of dinosaurs
during an expedition in the Gobi desert. Funny was, the eggs were
arranged in a square. The conclusion of the paleontologists was:
these eggs were already discovered by some humans before.
What my examples hopefully illustrate: whenever we find even very simple
structures like squares, letters, etc. we are certain that they were
produced by an intelligence, because information cannot simply appear
out of nothing. But then let us take the by far most complex structure
in the known universe - the human brain.
Here we suddenly say it can develop from alone. There is no intelligence
needed to be involved in creating this structure.
In other words: let some monkeys type, in some time a nice Lisp program
will appear.
Some people want to establish a new field of science, called
Intelligent Design.
If you are interested in mathematical "proofs" that we cannot develop
without a creator google for intelligence design.
> Unless the laws of mathematics are mutable, it seems that everything
> is the way it is without the need for any kind of creation event. If
> we solve the problem of creating a machine mind that is as good or
> better than the human mind, I expect there will be some controversy.
> How long it will take to do that I don't know. It may also be that we
> can never fully understand it. I already find it to be quite magical
> that the letter 'e' shows up on my display when I, or someone far away,
> types an 'e'. Yet all that is has been engineered. Math has not. So
> far as we know, it is only discovered.
Next sentence is _not_ meant offensive:
I think your view of mathematics is too naive.
I don't know your mathematical background... but maybe you want to look
into the issues of the beginning of last century, where math stoped to
work and needed some big changes. And you could look for information
regarding Gödel's incompleteness theorem. Possible starting point:
http://en.wikipedia.org/wiki/G%F6del%27s_incompleteness_theorem
> One of the coolest things I've read was in Carl Sagan's book Contact.
> I made sure to read it before the movie came out. In it, there is a
> rather interesting discussion about the constant Pi. As far as I
> know, Pi can't be anything other than what it is. But in the book,
> there is a "message" buried deep down in the number. I won't spoil
> it here. Read the book. It is better than the movie.
You made me curious :)
> Currently the weight of scientific evidence that I am familiar with
> leads me to conclude that a human-like synthetic intelligence is quite
> possible and perhaps even likely. It just requires deeper
> understanding on our part to achieve it. I think once that happens,
> the sales of anti-depressants will truly skyrocket.
If you really have some time (2 hours), I _highly_ suggest to read this:
http://www.kurzweilai.net/articles/art0134.html?m=1
This is a text of Ray Kurzweil about his "Law of accelerating Returns".
I am of course critical too, but anyway, I regard it as one of the most
important texts at all.
He explains why in 30-40 years we will have computers, millions of times
more intelligent than humans.
> My only real fear is that people will trust synthetic intelligence
> more than the real thing. Just because the machines will be smarter
> doesn't mean they will be infallible.
In fact, a very complicated issue...
André
--
Gareth McCaughan
> But for example take it this way:
> if you get some kilos of sand from the beach, climb on a high building
> and then let the sand fall down... what do you expect of the sand?
> You probably expect that it will fall down somehow, without any
> specific pattern. I am sure that you would be highly confused if you
> go downstairs and find how the sand has written the faculty function
> in Lisp. How often you might throw around sand, it will never fall in a
> way that it writes Lisp programs. This is randomness.
Is this meant to be some sort of argument against evolution?
Because, if so, it's a *very* bad one.
> Another example:
> in a school you give the pupils some dice, and every pupil has to throw
> it 100 times and write down the number.
> If one kid has 100 times on his paper a "3" you would not believe him
> that he did it correctly. Because 100 times a 3 is information, not
> randomness. An intelligence is needed to create information.
In the absence of a credible definition of "information",
that statement doesn't qualify as either true or false.
It's manifestly not true that intelligence is needed to
make anything analogous to "rolling a 3 100 times".
> Last example:
> You maybe heared that in the 1925 some people found eggs of dinosaurs
> during an expedition in the Gobi desert. Funny was, the eggs were
> arranged in a square. The conclusion of the paleontologists was:
> these eggs were already discovered by some humans before.
I'm not sure what significance that's supposed to have, but
it sounds interesting anyway. Can you tell me more? (I asked
Google, but it doesn't seem to have heard the story.)
> What my examples hopefully illustrate: whenever we find even very simple
> structures like squares, letters, etc. we are certain that they were
> produced by an intelligence, because information cannot simply appear
> out of nothing.
This is not true. It is very common to find patterns,
some of them very intricate and beautiful, that are well
explained in purely naturalistic terms. For instance,
naturally formed crystals make all sorts of elegant
patterns, which arise simply from the laws of physics.
(Similar patterns would doubtless arise from other
possible laws of physics, so the fact that the laws
of physics give rise to crystal growth doesn't seem
like good evidence that the laws of physics are themselves
the work of an intelligence.)
> But then let us take the by far most complex structure
> in the known universe - the human brain.
> Here we suddenly say it can develop from alone. There is no intelligence
> needed to be involved in creating this structure.
"Suddenly"? You surely don't think anyone says "All
those other ingenious bits of design in the natural
world, obviously they're the result of an intelligence
at work -- but the human brain isn't"?
> In other words: let some monkeys type, in some time a nice Lisp program
> will appear.
I think you need to learn a lot more about how evolution
is actually thought to work by people who believe in it;
it doesn't at all resemble your "let some monkeys type"
caricature.
> Some people want to establish a new field of science, called
> Intelligent Design.
It is clear that some people want to establish *something*
called "Intelligent Design". I regret that it doesn't look
much like a field of science; more like a strategy for
attacking science. (Google for "intelligent design wedge"
to find out more.)
> If you are interested in mathematical "proofs" that we cannot develop
> without a creator google for intelligence design.
I am a mathematician, and a Christian (and therefore more
sympathetic to at least some of the aims of these people
than most). Every such "proof" I have seen has been rubbish.
I don't rule out the possibility that there might be some
non-rubbish proof, and if you have one then I would be
very interested to see it.
--
Gareth McCaughan
.sig under construc
Jeff Dalton
> No. Nobody's really working on it. There are four different camps in AI
> today, and none of them is actually going to achieve AI:
> * And finally, there's the people doing stuff that isn't AI, namely
> anything that's a domain-specific system or isn't designed to produce
> domain-general intelligence. This includes the evolutionary and genetic
> algorithms people, the constraint solvers, and probably anything else
> that doesn't fit in categories 1-3 ...
But that is AI as the term is typically used. Look at an AI textbook,
for example, and that's the sort of thing it's about.
Re cargo cults, mentioned somewhere in this thread, there's
something that at least makes them sound less randomly irrational
in one of Marvin Harris's books, probably _Pigs, Wars, and Witches:
The Riddles of Culture_.
-- jd
Jeff Dalton
> However, as you point out, selection has been favouring larger brains
> for some time now, so in that sense, the extra brainpower must have
> some benefit.
That's not necessarily so. That brains have gotten larger does not
mean that must have been selected for. For instance, it might be that
something else, that had larger brains as a consequence, was what was
selected for. Larger brains might even be a net cost, so long as some
unavoidably associated benefit is great.
Even Dennett and other Darwinean maximalists admit, albeit
reluctantly, that not every feature was selected for.
-- jd
Jeff Dalton
> Fred Gilham wrote:
> > Another argument, which I feel is devastating to hard AI, is the
> > "hermeneutical hall of mirrors" argument, wherein the point is made
> > that even when machines appear to behave intelligently, they only do
> > so because we project meanings onto their behavior. The intelligence
> > they appear to manifest is our own reflected back to us.
That makes sense for something like Eliza, but consider a
chess-playing program, for example. If it plays well, then
it really does play well. It doesn't just seem to because
we're projecting onto its behaviour.
OTOH, is is playing well because of "intelligence" or because
of a good "brute force" (or even "fancy search") algorithm?
Intelligence isn't really about behaviour so much as about
how the behaviour is produced.
That's why people can resist the behavioural evidence:
> The hall of mirrors argument, imo, is irremediably tainted because it
> works in the opposite direction as well: once you've decided that some
> entity isn't behaving intelligently (in the sense of "human"
> intelligence") there is no behavior the entity can exhibit that will
> necessarily make you change your mind.
But they'd need an alternative explanation of the behaviour,
and that can be compared with other candidate explanations.
In some cases, it may actually be that we are projecting
meaning onto the behaviour.
> A little historical research
> will find plenty of essays in which the author argues that some subset
> of h. sap. (said subset not including the author) may exhibit a
> remarkably cunning simulation of "real" intelligence but is not in
> fact capable of reasoning.
Sure, but presumably we can show that they are wrong.
That doesn't mean that all such claims are always wrong.
-- jd
Jeff Dalton
> Fred Gilham <gil...@snapdragon.csl.sri.com> wrote:
> > Another argument, which I feel is devastating to hard AI, is the
> > "hermeneutical hall of mirrors" argument, wherein the point is made
> > that even when machines appear to behave intelligently, they only do
> > so because we project meanings onto their behavior. The intelligence
> > they appear to manifest is our own reflected back to us.
>
> Considering the number of computer users who appear to believe that
> Microsoft Windows is smarter than themselves, I suspect at least some
> of those are funhouse mirrors.
LOL.
-- jd
Ray Dillinger
> Some people want to establish a new field of science, called
> Intelligent Design.
Excuse me, but Intelligent Design does not follow the scientific method.
Please don't read what's below as an attack; it's not intended as one.
But I'm going to explain, point by point, what the scientific method *IS*,
and compare it to what the Intelligent Design people are doing. This is
not perjorative in any way; this is strictly a comparison. My intent is
to show that Intelligent Design and the Scientific Method have so little
in common that they cannot, under any circumstances, be considered to be
the same thing.
Scientists form hypotheses - this the Intelligent Design people have done.
Of course, everybody does this. Most of us call our hypotheses
opinions and don't pursue them using the scientific method. I
personally form a lot of hypotheses when I watch seagulls at the beach,
about what they're doing and why. But that doesn't make me a scientist.
Scientists then design experiments by using their hypothesis to try to
predict future events given some set of circumstances. This the ID
people have not done; their sole hypothesis has no predictive power.
Scientists then perform experiments by either bringing about the
circumstances required by their experimental design and studying
the results, or by examining nature to find circumstances matching
their experimental design and studying the results from those.
Lacking any hypothesis with predictive power, the ID people
can design no experiments; therefore they cannot perform them.
If the results of experiments are at variance with the prediction
made on the basis of a hypothesis, scientists then abandon that
hypothesis. Some may kick and moan and argue about it for years,
but eventually, a hypothesis which has less predictive power than
some other hypothesis is abandoned. This the ID people will not
do. In fact those who abandoned the sole hypothesis they've
advanced cease to *BE* the Intelligent Design people.
That which does not follow the scientific method is not science.
Science and the scientific method are not concepts that can be
separated. One who does not use the method is not a scientist.
Nobody can stop them from using the word "science" to describe
what they are doing. But the fact is that what they are doing is
not the scientific method. Therefore the assertion is simply
false.
That is not to state that the hypothesis they advance is false;
in fact it may be true. But it simply isn't a hypothesis about
which science can be done.
Bear
Randy Yates
> [...]
> This is not true. It is very common to find patterns,
> some of them very intricate and beautiful, that are well
> explained in purely naturalistic terms. For instance,
> naturally formed crystals make all sorts of elegant
> patterns, which arise simply from the laws of physics.
And the laws of physics come from what, an eternal vacuum
of space? ...
--
% Randy Yates % "My Shangri-la has gone away, fading like
%% Fuquay-Varina, NC % the Beatles on 'Hey Jude'"
%%% 919-577-9882 %
%%%% <ya...@ieee.org> % 'Shangri-La', *A New World Record*, ELO
http://home.earthlink.net/~yatescr
Randy Yates
> Randy Yates wrote:
>
>>>It is interesting to see, that some people believe that there is
>>>something like a soul.
>>>Even today most brain-scientists are very sure that there is no such
>>> thing like a free will.
>> Most scholars were sure the sun revolved around the earth.
>
> Here the issue is a bit different.
> The scientists started out asuming that we have a free will.
> Then they looked more detailed on the brain and came to the conclusion
> that there is no place for a free will.
>
> Anyway, free will is not possible as it would break "laws" of nature.
> The molecules in your brain are forced to behave in a special way. So
> they "force" you to act in a specific pattern (there is no will at all,
> no person/intelligence who is deciding the pattern and making descisions
> how you should act).
So we shouldn't punish criminals since they were just acting
based on the physical processes going on in their minds?
Randy Yates
> "Randy Yates" <ya...@ieee.org> wrote in message
> news:ad1fmo...@ieee.org...
>> David Steuber <da...@david-steuber.com> writes:
>
>> > I wasn't aware of that. I'm just using English. Maybe there exists
>> > some language spoken by people which does not have any concept of self
>> > reference.
>>
>> Concepts are not defined by language.
>
> If concepts are not defined by language, just what is language?
I would say that it is a mechanism by which thoughts can be
communicated to the outside world.
> Concepts may or may not be limited by language, but there is strong evidence
> that for a given culture and population, those concepts that are not
> conveyable in the language are, in fact, not conceived by representative
> members. New concepts generates new language to convey the concept, at least
> amoung a given (sub-)population.
That is an interesting topic. I think I agree that people tend to form
concepts based on the structure and capabilities of their language, but
they don't HAVE to.
--
% Randy Yates % "With time with what you've learned,
%% Fuquay-Varina, NC % they'll kiss the ground you walk
%%% 919-577-9882 % upon."
%%%% <ya...@ieee.org> % '21st Century Man', *Time*, ELO
http://home.earthlink.net/~yatescr
Randy Yates
You mean "sentient"?
> and a test for sentinence based on that definition. Wouldn't it
> be possible to create a machine which merely has the appearance
> of intelligence which does it best passing this test?
It depends on whether or not one could create a conclusive test for
sentience. As a simple example, let's say I want to test your
knowledge of the capitals of the 50 states. If I ask you to name them,
and you do so correctly, then one may conclude with certainty that you
have knowledge of the capitals of the 50 states.
However, you might have misunderstood me, or rather, I might not
have expressed myself very well. By "appearance of" I mean, essentially,
the non-conclusive and/or subjective tests, like the Turing test.
> we assume
> other people are intelligent because they appear intelligent to
> us, there's no way we can test if they are sentinent/self aware,
> because of the "special access" sentinent being has to its own
> mental states and events. you can't prove anything but yourself
> has such "special access" like you can't know, how would it feel
> to be something else.
>
> Randy> But I use the term "intelligent" here in a colloquial
> Randy> sense. To translate that to a more scientific and precise
> Randy> notion (but probably still way off), such a machine would
> Randy> need to be both "sentient" and "reasoning."
>
> sentinent and reasoning intelligent being is antopomophic
> intelligent being, and yes human intelligence is the only one we
> know. it doesn't mean intelligent being must be sentinent.
You keep spelling "sentinent" - the online dictionary has no entry
for this word. It also has no entry for "antopomophic," so I can't
respond intelligently until you define your terms or rephrase.
> the reason we need smarter machines is to help us solve our
> problems.
I never said this was the point of my query, i.e., to make
"smarter machines." Perhaps we do need smarter machines to
solve problems, but my questions are academic.
> i'm not sure sentinence is all that important for problem
> solving and trying to create Artificial Sentinence distracts us
> from the practical goals.
I tend to agree. I just think the concept of sentience is highly
intriguing, and being able to have a machine do it would prove
we understand what it is all about.
--
% Randy Yates % "...the answer lies within your soul
%% Fuquay-Varina, NC % 'cause no one knows which side
%%% 919-577-9882 % the coin will fall."
%%%% <ya...@ieee.org> % 'Big Wheels', *Out of the Blue*, ELO
http://home.earthlink.net/~yatescr
Randy Yates
> [...]
> It's time I fessed up. I'm a FORTRAN program written in the late
> 60's. Does that mean I pass the Turing test?
Yes. Now go compute pi to the last digit in base 2. (Use two's
complement.)
--
% Randy Yates % "The dreamer, the unwoken fool -
%% Fuquay-Varina, NC % in dreams, no pain will kiss the brow..."
%%% 919-577-9882 %
%%%% <ya...@ieee.org> % 'Eldorado Overture', *Eldorado*, ELO
http://home.earthlink.net/~yatescr
David Steuber
> Gareth McCaughan <gareth.m...@pobox.com> writes:
> > [...]
> > This is not true. It is very common to find patterns,
> > some of them very intricate and beautiful, that are well
> > explained in purely naturalistic terms. For instance,
> > naturally formed crystals make all sorts of elegant
> > patterns, which arise simply from the laws of physics.
>
> And the laws of physics come from what, an eternal vacuum
> of space? ...
Where do the properties of numbers come from? I can get the same
results with (+ 2 3) and (+ 3 2). Where did that law come from? For
that matter, who decided on the decimal representation of Pi and all
those many infinite series that converge on Pi?
Could things have been decided otherwise?
David Steuber
> David Steuber <da...@david-steuber.com> writes:
> > [...]
> > It's time I fessed up. I'm a FORTRAN program written in the late
> > 60's. Does that mean I pass the Turing test?
>
> Yes. Now go compute pi to the last digit in base 2. (Use two's
> complement.)
Would you like fries with that?
Artem Baguinski
Randy> You mean "sentient"?
yep. sorry, my bad.
>> we assume other people are intelligent because they appear
>> intelligent to us, there's no way we can test if they are
>> sentinent/self aware, because of the "special access" sentinent
>> being has to its own mental states and events. you can't prove
>> anything but yourself has such "special access" like you can't
>> know, how would it feel to be something else.
>>
Randy> But I use the term "intelligent" here in a colloquial
Randy> sense. To translate that to a more scientific and precise
Randy> notion (but probably still way off), such a machine would
Randy> need to be both "sentient" and "reasoning."
>>
>> sentinent and reasoning intelligent being is antopomophic
>> intelligent being, and yes human intelligence is the only one
>> we know. it doesn't mean intelligent being must be sentinent.
Randy> You keep spelling "sentinent" - the online dictionary has
Randy> no entry for this word. It also has no entry for
Randy> "antopomophic," so I can't respond intelligently until you
Randy> define your terms or rephrase.
s/sentinent/sentient/
s/antopomophic/anthropomorphic/
sorry again. i'll try to spell better.
>> the reason we need smarter machines is to help us solve our
>> problems.
Randy> I never said this was the point of my query, i.e., to make
Randy> "smarter machines." Perhaps we do need smarter machines to
Randy> solve problems, but my questions are academic.
>> i'm not sure sentinence is all that important for problem
>> solving and trying to create Artificial Sentinence distracts us
>> from the practical goals.
Randy> I tend to agree. I just think the concept of sentience is
Randy> highly intriguing, and being able to have a machine do it
Randy> would prove we understand what it is all about.
oh, ok. i didn't look at it from this angle. i understand the
motivation, but i still find it impossible task. the only
sentience we're directly aware of is our own. sentience of others
we conclude from their likness to us. other people act like me,
they're built like me, they must feel and think like, just like i
do. well, they also tell me that they feel and think and i don't
have reason to not believe them.
animals act and are built, well, quite like people, they don't
tell us that they can think and feel, but the more they are like
people - more "intelligence" we attribute to them. but we can't
be sure they are sentient in the way we are.
machines may be made to act like people. as they are now they are
built very different from us though. we have one reason less to
think of them as sentient beings. they can be made to tell us
what they think and feel - then we'll have one reason more. but
we won't know for sure they are sentient in a sence humans are
sentient.
Joe Marshall
> Another example:
> in a school you give the pupils some dice, and every pupil has to throw
> it 100 times and write down the number.
> If one kid has 100 times on his paper a "3" you would not believe him
> that he did it correctly. Because 100 times a 3 is information, not
> randomness. An intelligence is needed to create information.
Both papers have the same information. One has a much smaller
Kolmogorov complexity.
You do not need intelligence to create information. The laws of
thermodynamics (which is intimately related to information theory)
work just fine in the absence of an intelligent observer.
> Last example:
> You maybe heard that in the 1925 some people found eggs of dinosaurs
> during an expedition in the Gobi desert. Funny was, the eggs were
> arranged in a square. The conclusion of the paleontologists was:
> these eggs were already discovered by some humans before.
On the island of Kvadehuksletta one can find piles of rocks arranged
in perfect circles about a meter in radius. However, it appears that
these are natural formations.
> What my examples hopefully illustrate: whenever we find even very simple
> structures like squares, letters, etc. we are certain that they were
> produced by an intelligence, because information cannot simply appear
> out of nothing.
What about the `Face' on Mars?
What about `happy face' spiders?
http://biology.swau.edu/faculty/petr/ftphotos/hawaii/postcards/spiders/
How about the horrified Bryzoan Selenaria punctata?
http://www.adelaide.edu.au/microscopy/services/instrumentation/gallery.html
(see image 6)
> But then let us take the by far most complex structure in the known
> universe - the human brain. Here we suddenly say it can develop
> from alone.
Who says that? Most evolutionists believe that the brain evolved from
similar, but slightly less complex brains such as those found in other
primates.
Joe Marshall
> David Steuber <da...@david-steuber.com> writes:
>> [...]
>> It's time I fessed up. I'm a FORTRAN program written in the late
>> 60's. Does that mean I pass the Turing test?
>
> Yes. Now go compute pi to the last digit in base 2. (Use two's
> complement.)
That's easy: 1
André Thieme
> André Thieme wrote:
>
>
>>Some people want to establish a new field of science, called
>>Intelligent Design.
>
>
> Excuse me, but Intelligent Design does not follow the scientific method.
Yes, right, that's why I said they "want to establish" it.
> Please don't read what's below as an attack; it's not intended as one.
> But I'm going to explain, point by point, what the scientific method *IS*,
> and compare it to what the Intelligent Design people are doing. This is
> not perjorative in any way; this is strictly a comparison. My intent is
> to show that Intelligent Design and the Scientific Method have so little
> in common that they cannot, under any circumstances, be considered to be
> the same thing.
Thank you for your explanations.
I just want to state that I am not a follower of ID but just wanted to
point out to David that there are other possible options available.
Perhaps I spent too much time with talking about ID while for me the
link to Kurzweil was much more important ;)
> That which does not follow the scientific method is not science.
> Science and the scientific method are not concepts that can be
> separated. One who does not use the method is not a scientist.
>
> Nobody can stop them from using the word "science" to describe
> what they are doing. But the fact is that what they are doing is
> not the scientific method. Therefore the assertion is simply
> false.
Yes, agreed. Some people came up with a definition for the word science
which is usually used when talking about it.
Of course this "official definition" is objectively not right, as any
definition by definition isn't, so other people could (if helpfull or
not) give another definition of it.
André
--
André Thieme
>>But then let us take the by far most complex structure in the known
>>universe - the human brain. Here we suddenly say it can develop
>>from alone.
>
>
> Who says that? Most evolutionists believe that the brain evolved from
> similar, but slightly less complex brains such as those found in other
> primates.
I am not a follower of Intelligent Design, but they explain with statistical
arguments that it is not possible that a brain can evolve.
If you want to learn more about these arguments please google a bit,
as I cannot defend the position as I am not involved enough into the
thoughts behind it.
André
--
André Thieme
> So we shouldn't punish criminals since they were just acting
> based on the physical processes going on in their minds?
Exactly this discussion will become very important as soon science
can prove that there is no thing such free will.
I suppose we will be forced to change the way how laws work.
We probably need to redefine several aspects of it..
As science evolves our understanding of what a crime is does evolve
too. Perhaps in 100 years some computers who are some billion times
more intelligent than we are can come up with a definition of crime
that describes this subject by the movement and patterns of some
molecule structures. Something which is extremly different from our
current understandings of it.
Perhaps they will describe some patterns... and when these patterns
come up some where (something we see as a "crime") other patterns
are started as a result of it: move some molecules (putting the
ciminal into a prison (or whatever might exist in 100 years)).
*shurgs*
André
--
André Thieme
>>But for example take it this way:
>>if you get some kilos of sand from the beach, climb on a high building
>>and then let the sand fall down... what do you expect of the sand?
>>You probably expect that it will fall down somehow, without any
>>specific pattern. I am sure that you would be highly confused if you
>>go downstairs and find how the sand has written the faculty function
>>in Lisp. How often you might throw around sand, it will never fall in a
>>way that it writes Lisp programs. This is randomness.
>
>
> Is this meant to be some sort of argument against evolution?
> Because, if so, it's a *very* bad one.
I understood it this way: it is just a warm up for the average surfer
that one can read on internet sites about Intelligent Design.
I am not very involved into ID, so I don't know their more detailed
explanations. I just wanted to point out some places where these
arguments could be found.
>>Another example:
>>in a school you give the pupils some dice, and every pupil has to throw
>>it 100 times and write down the number.
>>If one kid has 100 times on his paper a "3" you would not believe him
>>that he did it correctly. Because 100 times a 3 is information, not
>>randomness. An intelligence is needed to create information.
>
>
> In the absence of a credible definition of "information",
> that statement doesn't qualify as either true or false.
> It's manifestly not true that intelligence is needed to
> make anything analogous to "rolling a 3 100 times".
You can visit http://www.intelligentdesigner.de/ and contact the author
to learn more about it.
>>Last example:
>>You maybe heared that in the 1925 some people found eggs of dinosaurs
>>during an expedition in the Gobi desert. Funny was, the eggs were
>>arranged in a square. The conclusion of the paleontologists was:
>>these eggs were already discovered by some humans before.
>
>
> I'm not sure what significance that's supposed to have, but
> it sounds interesting anyway. Can you tell me more? (I asked
> Google, but it doesn't seem to have heard the story.)
Hmm, I supposed there would be more information about this issue, like
the conclusions. The author of the linked site can you probably tell
more about it (and perhaps give some sources).
>>What my examples hopefully illustrate: whenever we find even very simple
>>structures like squares, letters, etc. we are certain that they were
>>produced by an intelligence, because information cannot simply appear
>>out of nothing.
>
>
> This is not true. It is very common to find patterns,
> some of them very intricate and beautiful, that are well
> explained in purely naturalistic terms. For instance,
> naturally formed crystals make all sorts of elegant
> patterns, which arise simply from the laws of physics.
> (Similar patterns would doubtless arise from other
> possible laws of physics, so the fact that the laws
> of physics give rise to crystal growth doesn't seem
> like good evidence that the laws of physics are themselves
> the work of an intelligence.)
I think my wording was very bad.
Well, as I understood it, ID says that information cannot appear out of
nothing. You are describing some order/regularity (like structures that
you can find in nature) and not about information. For example it would
be hard to find some trees in nature who grew in a way (when observed
from a helicopter) that they form the letters which itself are the source
code of a scheme compiler.
I mixed some terms, sorry for that.
If you want definitions I have to refer you to the linked site.
Thanks for your comments.
André
--
John Thingstad
is predictable.
You fail to take into account the "butterfly effect". Highly complex
behaviour is
possible even in simple systems. There is certainly room for "free will"
in our
understanding of nature. A model of the human brain would be simular.
You can see general outline of predictable behaviour but would be unable
to estimate
the exact flow of thought over time.
--
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/
Artem Baguinski
André> David Steuber wrote:
>> I've more or less discounted the existence of a divine creator.
André> I am not so certain about this. There are some people who
André> have very good arguments that we were created by an
André> intelligence. If this intelligence is a living creature, or
André> if it is some strange kind of "nature laws" is not sure.
André> [...]
André> What my examples hopefully illustrate: whenever we find
André> even very simple structures like squares, letters, etc. we
André> are certain that they were produced by an intelligence,
André> because information cannot simply appear out of
André> nothing.
And who has created the creator?
Artem Baguinski
>> Anyway, free will is not possible as it would break "laws" of
>> nature. The molecules in your brain are forced to behave in a
>> special way. So they "force" you to act in a specific pattern
>> (there is no will at all, no person/intelligence who is
>> deciding the pattern and making descisions how you should act).
Randy> So we shouldn't punish criminals since they were just
Randy> acting based on the physical processes going on in their
Randy> minds?
If there's no free will the constructions like "we should" or "we
shouldn't" don't work: we don't decide to punish a criminal, we
simply do it because laws of nature dictate us to.
Ray Dillinger
>
> I am not a follower of Intelligent Design, but they explain with statistical
> arguments that it is not possible that a brain can evolve.
This is odd, because simulated neural networks certainly can evolve.
I use GA to develop new structure in ANN all the time!
Bear
André Thieme
I am not an intelligent Designer, but I suppose they don't see a problem
that NNs work, cause they were created by an intelligence.
I suppose Intelligent Design sees no problem with intelligence when some
intelligence was involved. What they would find odd if in a storm some
metal parts of a mountain fall down to the ground in a way that a computer
is created which runs a NN. ;-)
André
--
Simon Alexander
> I am not a follower of Intelligent Design, but they explain with statistical
> arguments that it is not possible that a brain can evolve.
> If you want to learn more about these arguments please google a bit,
> as I cannot defend the position as I am not involved enough into the
> thoughts behind it.
> André
And if you google a bit, you should be able to find competent rebuttle of
essentially every claim the ID folk make. The statistical arguments that I
have seen from ID are naive, to be charitable. They certainly haven't
"explained that it is not possible that a brain can evolve".
Simon.
Gorbag
"Joe Marshall" <j...@ccs.neu.edu> wrote in message
news:oept5n...@ccs.neu.edu...
No, no, it's 0. You must have rounded somewhere. Did you use floating point,
or exact bignums?
Andreas Scholta
> Well, as I understood it, ID says that information cannot appear out of
> nothing. You are describing some order/regularity (like structures that
> you can find in nature) and not about information. For example it would
> be hard to find some trees in nature who grew in a way (when observed
> from a helicopter) that they form the letters which itself are the source
> code of a scheme compiler.
good day..
I have been lurking here in the shadows for quite some time now, finally
feeling the urge to step into the light (knowing that after posting i am
probably going to regret this).
The difference between "Chaos" and "Information" lies in our
understanding and interpretation. We create information by relating
observations to past experiences and aquired knowledge.
more clearly:
If something is chaotic to us it is only because we are unable to see a
pattern.
If something is information to us it is because we are able to see a
pattern and can connect the observation made with our already aquired
knowledge.
it is not really important that a letter has a sender who crammed a lot
of stuff into it for us to read and understand. it is that we as the
recipients can relate to the letter's content. while knowing that there
was a sender who supposedly HAD to say SOMEthing in that letter makes us
search more keenly for bits of information while reading it.. the
crucial part is not the assembling, but the disassembling. we can
disassemble almost anything the way we see fit.
if you and me were to hop into a helicopter, fly around until we found a
wood, I would show you how out of the many trees you saw there, I could
pick some out, shaping the character #\A.
have fun,
Andreas Scholta
Joe Marshall
It is customary to suppress the trailing zeroes on a fraction, so the
rightmost digit cannot be zero. Since we are in base two, that leaves
us few options. (The leftmost digit of anything in base two is 1
also.)