j********e 发帖数: 124 | 1 *How does a piece of code have self-awareness?
*A robot should not allow anything to turn off its power.
*A robot should have consistant "will" to be qualified as human-like.
"Guns don't kill, only people do." This saying is based on the fact that a gun
is a machine that don't have a "will".
What qualifies a machine to be having "will"?
Self-awareness is one necessary condition for having a will.
*Does a virus have self-awareness / will?
............ cell ........................................ | f*****r 发帖数: 229 | 2
From the self-diagnose perspective, we can consider a kind of self-awareness:
the trace/log of interactions between components. In an autonomic computing
field, each component is an autonomic part, and exposes well-defined interface
to the outside of the world. In this world, only interaction(input/output or
stimulus/response) makes a component knows itself. For example, the component
can learn what is the normal behavior of itself (note that a component's
behavior depends on outside inputs
【在 j********e 的大作中提到】 : *How does a piece of code have self-awareness? : *A robot should not allow anything to turn off its power. : *A robot should have consistant "will" to be qualified as human-like. : "Guns don't kill, only people do." This saying is based on the fact that a gun : is a machine that don't have a "will". : What qualifies a machine to be having "will"? : Self-awareness is one necessary condition for having a will. : *Does a virus have self-awareness / will? : ............ cell ........................................
| j********e 发帖数: 124 | 3 You inputs are very interesting to me!
For "A robot should not allow anything to turn off its power.", I mean more
the following:
Just like a human can be "brain-washed", which I don't find solid cases, a
robot can be injected with new control software when it is powered-off. The
harm is that its central control software is abruptly changed, not consistant
self-adaption. In this case, the robot's consistant identity is violated.
Thus, I think if I am a robot, I would do all I can to stop anythin
【在 f*****r 的大作中提到】 : : From the self-diagnose perspective, we can consider a kind of self-awareness: : the trace/log of interactions between components. In an autonomic computing : field, each component is an autonomic part, and exposes well-defined interface : to the outside of the world. In this world, only interaction(input/output or : stimulus/response) makes a component knows itself. For example, the component : can learn what is the normal behavior of itself (note that a component's : behavior depends on outside inputs
| a******n 发帖数: 100 | 4 I have the same random thought over many years.
The most fundamental question I think is "self-awareness". Even
mathematically, many fundamental theoretical results are based on
self-(but)-contradiction: e.g. Turing's proof of HALTing problem. Can one
really reason about himself?
It seems very hard to find 'self'. Do you really think human has
free-will-of-its-own ? Are you just a machine that God build? (just as
nowadays human build computers?) Even before the film "Matrix" is shot, I
have
【在 j********e 的大作中提到】 : *How does a piece of code have self-awareness? : *A robot should not allow anything to turn off its power. : *A robot should have consistant "will" to be qualified as human-like. : "Guns don't kill, only people do." This saying is based on the fact that a gun : is a machine that don't have a "will". : What qualifies a machine to be having "will"? : Self-awareness is one necessary condition for having a will. : *Does a virus have self-awareness / will? : ............ cell ........................................
| a******n 发帖数: 100 | 5
I have thought about a slightly different toy many many years ago, which can
be trivially built with today's technology:
Basically it's a toy car, with rechargable battery, and a monitor to indicate
the battery level, plus a sensor to find the charger in the room.
Precondition: there is always a charger can be found within distance
(battery.level > 0.10)
while(true) {
if (battery.level > 0.50) then { randomly travel in the room; }
else { find the neareast charger; and charge itself; }
}
Do
【在 j********e 的大作中提到】 : You inputs are very interesting to me! : For "A robot should not allow anything to turn off its power.", I mean more : the following: : Just like a human can be "brain-washed", which I don't find solid cases, a : robot can be injected with new control software when it is powered-off. The : harm is that its central control software is abruptly changed, not consistant : self-adaption. In this case, the robot's consistant identity is violated. : Thus, I think if I am a robot, I would do all I can to stop anythin
| c******n 发帖数: 4965 | 6 I think human intelligence has a large part to do with
random exploration,
that's probably the difference between a human and a TM
【在 a******n 的大作中提到】 : I have the same random thought over many years. : The most fundamental question I think is "self-awareness". Even : mathematically, many fundamental theoretical results are based on : self-(but)-contradiction: e.g. Turing's proof of HALTing problem. Can one : really reason about himself? : It seems very hard to find 'self'. Do you really think human has : free-will-of-its-own ? Are you just a machine that God build? (just as : nowadays human build computers?) Even before the film "Matrix" is shot, I : have
| a******n 发帖数: 100 | 7
That's also an interesting topic I discussed with people near me.
It fits nicely with how human (or creature in general) evolve, according to
Darwin's theory.
But intelligence ?= genetic algorithm + a *true* random number generator +
natural (or God) selection
Something we still not feel very comfortable with. Even in physics, Einstein
said "I don't believe God play dice".
Well, the good thing about this theory is that, different things (physics,
biology, computer science) start to converge to
【在 c******n 的大作中提到】 : I think human intelligence has a large part to do with : random exploration, : that's probably the difference between a human and a TM
| h*******y 发帖数: 864 | 8 Actually, you said an interesting piont that intelligence was due to
random exploration. In fact, there are many different views.
Some think that intelligence is inherited from ancestors. Some think that
intelligence is shaped by culture and society you live in. Some think that
intelligence evolved as human evolved.Some think that intelligence is
developed because children are actively explorate the world around them.
Moreover, their genes set them up to explorate the world.
You know, many exper
【在 c******n 的大作中提到】 : I think human intelligence has a large part to do with : random exploration, : that's probably the difference between a human and a TM
| h*******y 发帖数: 864 | 9
my
I think the answer is that monkeys eventually think about the world, him or
his family and so forth. You toy do not think. In addition, your toy can not
even remember the place for recharging, and it do not have a goal for finding
the place recharging while traveling randomly.
【在 a******n 的大作中提到】 : : That's also an interesting topic I discussed with people near me. : It fits nicely with how human (or creature in general) evolve, according to : Darwin's theory. : But intelligence ?= genetic algorithm + a *true* random number generator + : natural (or God) selection : Something we still not feel very comfortable with. Even in physics, Einstein : said "I don't believe God play dice". : Well, the good thing about this theory is that, different things (physics, : biology, computer science) start to converge to
| h*******y 发帖数: 864 | 10 You think well!!!
It is pleasure to read such post.:)
Actaully, the most fundamental question is "consciousness". Whether a robot
can percive the world, serve his best interest and so forth. I actually bought
a book called "Robot Companions". It is an introductory book.The book talks
about some questions like those you have.
in
No one can answer that question. It is the classical debate whether humans
have free will or humans are determined. However, human do have free will,but
not always.
to
T
【在 a******n 的大作中提到】 : I have the same random thought over many years. : The most fundamental question I think is "self-awareness". Even : mathematically, many fundamental theoretical results are based on : self-(but)-contradiction: e.g. Turing's proof of HALTing problem. Can one : really reason about himself? : It seems very hard to find 'self'. Do you really think human has : free-will-of-its-own ? Are you just a machine that God build? (just as : nowadays human build computers?) Even before the film "Matrix" is shot, I : have
| | | h*******y 发帖数: 864 | 11 偶对AI也不太懂,只是从偶学的COGNITION来回答的。
希望不要嘲笑。偶挺爱讨论这中类似问题的。很有意思。 | j********e 发帖数: 124 | 12 3x 3x a lot for all the inputs!!!
And here is a material, somewhat related to our topic. It is about a case of
current brain surgery, where the patient should be remaining concious during
the process. The surgery is for removing a brain tumor, which is close to the
patient's language core. In order to make sure the language is not harmed, the
patient should be asked to talk from time to time, during the surgery. Sigh!
Hanible in real life!
The detail is in the latest issue of National Geogrphic:
【在 j********e 的大作中提到】 : *How does a piece of code have self-awareness? : *A robot should not allow anything to turn off its power. : *A robot should have consistant "will" to be qualified as human-like. : "Guns don't kill, only people do." This saying is based on the fact that a gun : is a machine that don't have a "will". : What qualifies a machine to be having "will"? : Self-awareness is one necessary condition for having a will. : *Does a virus have self-awareness / will? : ............ cell ........................................
| a******n 发帖数: 100 | 13 fengeer, you missed our point, or you haven't thought deep enough:
What you said about "learning from books/people.. reduce the search space" is
already achieved by a Turing Machine. e.g. a computer can behave correctly (or
"intelligently") like play chess once the apporate rules(program) is loaded
("learned").
What we (creation and me) think important is how to generate the *rules*
itself. e.g. let's talking about human, we don't know how to calculate the
force between two objects until 1700 w
【在 f*****r 的大作中提到】 : : From the self-diagnose perspective, we can consider a kind of self-awareness: : the trace/log of interactions between components. In an autonomic computing : field, each component is an autonomic part, and exposes well-defined interface : to the outside of the world. In this world, only interaction(input/output or : stimulus/response) makes a component knows itself. For example, the component : can learn what is the normal behavior of itself (note that a component's : behavior depends on outside inputs
| m******t 发帖数: 2416 | 14 Recommend you guys read "Speaker for the Dead" and "Xenocide". Orson Scott
Card has some very insightful philoshopical thoughts on what life forms can be
considered sentient. | d*l 发帖数: 400 | 15 I don't think random process can explain intelligence. let monkeys write
random letters on a tape, how many years would it take for them to write a
master piece, like Hong Lou Meng?
is
(or
existed
"force
think
will
【在 a******n 的大作中提到】 : fengeer, you missed our point, or you haven't thought deep enough: : What you said about "learning from books/people.. reduce the search space" is : already achieved by a Turing Machine. e.g. a computer can behave correctly (or : "intelligently") like play chess once the apporate rules(program) is loaded : ("learned"). : What we (creation and me) think important is how to generate the *rules* : itself. e.g. let's talking about human, we don't know how to calculate the : force between two objects until 1700 w
| m******t 发帖数: 2416 | 16
Well, I guess these guys' point is that the universe has come to be for
billions of years, so the monkey has actually typed for long enough to
come up with Hong Lou Meng. 8-)
【在 d*l 的大作中提到】 : I don't think random process can explain intelligence. let monkeys write : random letters on a tape, how many years would it take for them to write a : master piece, like Hong Lou Meng? : : is : (or : existed : "force : think : will
|
|