The way I see the human mind, and the way I've often observed my own mind to be, is like that of a very large, very complex computer program.
You are in good company, some AI folks have believed so for decades. IMHO the computer analogy is a much too simplified model for our minds. It leads to all sorts of weird conclusions, such as the idea that there might be a sequence of words (a program in your analogy) that would cause the mind to hang. Also the mind performs many things in parallel, for instance all sensory input (vision, audio etc) is preprocessed in massive parallelism before it reaches conscious experience.
On a side note, if it is true, that we do continually form new neurons, then why do older people seem to learn much slower and often forget things? Is it even possible to really ever forget something?
New information can replace old by means of synaptic changes--happens all the time in ANNs, where no new nodes are added once training has started. I doubt that every time you learn something, you get a new neuron. And the case that old people seem to learn much slower: It takes less time to learn a completely new concept for a fresh mind, than to try to incorporate a new concept into an existing body of knowledge.
So reluctance to learning might be a cultural trait? Probably, but I think this is a secondary effect that merely reinforces the trend. The main cause I believe is that (as you mentioned) the need to learn decreases, and the need to learn decreases because we know a lot of useful stuff already...
However I think you could take it a bit further. If the mind already has a model that explains a phenomenon sufficiently well, there is a resistance to learn a new model that might be better, instead you just make small modifications on your old one. Just like the MS Word app BTW;)
Back in the 50s everyone thought nurture was the case, then in the late 70s things switched to nature, and now we are back to the correct answer: nurture.
The world is not binary, at least not the one outside my window;) The debate is not about if it is nature OR nurture. It's about to which degree they both matter.
Age has nothing to do with new tricks. Old dogs _can_be taught new tricks, and old people _can_ learn new things. It's a stupid cliche really. The real deal is "You can't teach an old dog STUPID tricks."
I agree completely. But as you grow older you'll tend to find more things stupid, because you know better--or at least think you do;)
But... there are a lot of things that are universal- across all cultures. No society would think it's a good thing to kill your folks. That society would disappear after just a few generations.
Ever heard of mercy killing?
IMHO universal right and wrong is just an illusion we get from the abstract concepts the use of language impose on us. Ofcourse "Murder" is wrong, in the abstract sense, but as soon as you deal with an actual experienced situation, the choice of whether it is right or wrong will depend on the actual situation, and previous experiences of the subject deciding between right and wrong.
The nuns don't get Alzheimer's because they use their brains constantly, keeping diaries and staying active well into their later years.
My grandmother regularly read newspapers, and led an active social life. She even practised memorising long sequences of digits to keep her mind crisp. She now has Alzheimer's...
There could be a lot of other reasons for the phenomenon you describe. For instance, nuns have a lot of habits that differ from those most other people have. Maybe some special kind of diet, and plenty of physical exercise did the trick?!
On the other hand, maybe mental exercise does prevent Alzheimer's, but only in the statistical sense...
A short excerpt: Everyone who believes, as I do, that all the corporeal and mental organs (excepting those which are neither advantegous or disadvantegous to the posessor) of all beings have been developed through natural selection, or the survival of the fittest, together with use or habit[4], will admit that these organs have formed so that their possessors may compete succesfully with other beings, and thus increase in number.
Sure Darwin was religious, and sure he did believe in a "Creator God", but only before he set out with the Beagle.
See, those creationists may not be so kooky as you think!
It's always sound to be sceptical about all theories, that much I can agree with you. But, while most scientists that study evolution are sceptical, almost no creationists are. Disregarding all the details about the theories I find it far more open minded to study a phenomenon, and then formulate a theory, as Darwin did, than to adopt a dogma and then try to convince others...
The only thing I didn't like was when the little brat says "This is a UNIX system... I know this"
When I wathced JP on the first night it showed, the cinema was filled with computer science students. This line made over half the crowd go "Wooow!", and then laugh. The rest of the audience was left looking rather puzzled:-)
I concluded however that the movie misses the point. The really frightening thing about DNA analysis is not that it makes prejudice based on DNA possible, prejudice is already deeply rooted in our society, but that that (unlike frenology which was just bullshit) prejudice based on genes would be right.
I'd rather conclude that you missed the point:-) Predjudice based on genes would be only partly right, but could easily be seen as completely justified (as you seem to think). My perception of the movie was that it was this danger it was out to pinpoint.
The fitness of a human is a combination of inherited and learned traits. This is why Ethan Hawke's character (sorry, forgot the name) was able to beat his brother despite his inferior physique: He had developed a stronger will-power.
I loved GATTACA, and found it a wonderful, and thought-provoking movie.
The other possibility would be to reduce the population; simply remove the industrial working class by randomly sterilizing a large number of people.
This problem has already been dealt with by another insightful SF author: Douglas Adams. Simply tell the redundant part of the population that the world is doomed, and send them to colonize another world:-)
Seriously, I think this problem will solve itself, since robotics breakthroughs will not happen overnight. Most likely we will gradually have more people employed in service proffessions, like shop keepers, telephone sanitisers (couldn't resist;) ), accountants etc.
The problem is, that computers don't see. They look, but they don't really see.
No, exactly. And to make a computer really see, is what's known as the vision problem. My question (as you will see if you read it again) was: Why isn't this a suitable problem for AI?
Hmm... I've read some of the conversations there now. Judging from the ratio of gender words in the replies, I'd say that the system is either very human or operated by a guy/some guys in their twenties:-)
Any other ideas for good AI benchmarks? The benchmarks have to be there to encourage funding and some research, so some test needs to be decided as a standard.
Computer vision is a decent test, but it has to be under such tight constraints. Other senses aren't worth the time, so we are left we only a few options.
This remark leaves me clueless. What's wrong with the vision problem? If you have a computer system that can see, wouldn't that be useful? What are those constraints? If you are referring to the great need of processing power, I'd say thats more of a challenge (if the algorithms require too much processing power, then maybe they need rethinking)!
The problem with the Turing Test is that it tries to make a computer human...
Actually it does not even encourage us to make a computer human. It encourages us to make a computer program that produces sentences that sound human. Compare the size of the cortical areas devoted to speech processing, to the total area of the brain. This is my estimate of the Turing test's relevance:-)
The general problem with the Turing test, as with most of the rest of the classical AI genre, is that it assumes that all relevant information processing should be symbolic. More likely only a small fraction of the information processing ought to be symbolic, the rest sub symbolic (ANN-s, fuzzy logic etc).
Look at ants, rabbits, dogs etc -- They cannot do symbolic information processing (cannot speak) but the feat they accomplish is still pretty impressive!
Seriously, the way to go in neural networks is most likely NOT to have 40,000-60,000 separate classes on the response side. Such a system cannot generalize at all, and space is wasted. ie. many links will be the same for similar words. There should be an intermediary representation (not a hidden layer, but a sparse/semi local representation of the output concepts).
White noise is certainly random - but background noise in real world situations is hardly going be that random. Rather, it's going to be a chaotic blend of non-random signals - each of which may (or may not) be a valid speech signal in it's own right.
Actually it does not matter whether the background signal is completely white, or not. As long as the speech signal is the most correlated one, you can find it. The coctail party problem (to isolate one speech signal in a crowd of speakers) is of course more difficult. The technique can be extended to separate more sources, if one adds more microphones/ears (see independent component analysis), one extra microphone per source you want to isolate, but that would be to cheat, wouldn't it;)
The net has to know what it is listening for inside of the noise before it can actually pick it out.
NO, IT DOES NOT!!
It all comes down to statistics. Speech is a non-white signal. Noise is white. If you have two microphones/ears, you simply search for the linear combination of the two signals that is the most un-correlated temporally, and voala! You have found the speech signal. This is known as blind separation.
Yes, this passage about just 11 neurons connected by a mere 30 links makes me wonder what this net actually does. "Speech Recognition" could of course also mean the ability to recognize that an audio signal contains speech:-)
Of course the task of net could also be to separate the noise signal from the speech, aka blind separation, a problem that has been solved before (for instance by independent component analysis)
If this is merely ICA with a time coded neural net, it is IMHO still pretty cool, and much more impressive than all those commercial systems that rely on dumb correlation and processing power.
Anyway, instead of just having me guessing, could someone please point to their paper:-)
Slashdot Mirror
Pardon me, but what does the icon for this topic mean?
Looks like a flowerpot to me... ?!
Didn't it use to be a cup of steaming hot coffe?
The way I see the human mind, and the way I've often observed my own mind to be, is like that of a very large, very complex computer program.
You are in good company, some AI folks have believed so for decades. IMHO the computer analogy is a much too simplified model for our minds. It leads to all sorts of weird conclusions, such as the idea that there might be a sequence of words (a program in your analogy) that would cause the mind to hang. Also the mind performs many things in parallel, for instance all sensory input (vision, audio etc) is preprocessed in massive parallelism before it reaches conscious experience.
Sorry, a typo. Here is what I meant to say:
right and wrong are not subjective, rather whatever society agrees upon. these are known as laws.
Oh, it's that simple. Silly me. But, I happen to dislike some of the laws in my country, In my point of view they are wrong, yet I obey them.
Apparently, what you call right and wrong, seems to be what I call legal and illegal.
right and wrong are not subjective, rather whatever society agrees upon. these are known as laws.
Oh, it's that simple. Silly me. But, I happen to dislike some of the laws in my country, I my point of view they wrong, yet I obey them.
Apparently, what you call right and wrong, seems to be what I call legal and illegal.
On a side note, if it is true, that we do continually form new neurons, then why do older people seem to learn much slower and often forget things? Is it even possible to really ever forget something?
New information can replace old by means of synaptic changes--happens all the time in ANNs, where no new nodes are added once training has started. I doubt that every time you learn something, you get a new neuron.
And the case that old people seem to learn much slower: It takes less time to learn a completely new concept for a fresh mind, than to try to incorporate a new concept into an existing body of knowledge.
Nice comment, should have higher score IMHO...
So reluctance to learning might be a cultural trait? Probably, but I think this is a secondary effect that merely reinforces the trend. The main cause I believe is that (as you mentioned) the need to learn decreases, and the need to learn decreases because we know a lot of useful stuff already...
I just love the MS analogy ;)
;)
However I think you could take it a bit further. If the mind already has a model that explains a phenomenon sufficiently well, there is a resistance to learn a new model that might be better, instead you just make small modifications on your old one.
Just like the MS Word app BTW
Back in the 50s everyone thought nurture was the case, then in the late 70s things switched to nature, and now we are back to the correct answer: nurture.
;)
The world is not binary, at least not the one outside my window
The debate is not about if it is nature OR nurture. It's about to which degree they both matter.
Age has nothing to do with new tricks. Old dogs _can_be taught new tricks, and old people _can_ learn new things.
;)
It's a stupid cliche really. The real deal is "You can't teach an old dog STUPID tricks."
I agree completely. But as you grow older you'll tend to find more things stupid, because you know better--or at least think you do
Thanks!
IMHO this one is much more well written than the NYT one. I'll check out the mentioned article in Science (Oct 15 issue) tomorrow...
But... there are a lot of things that are universal- across all cultures. No society would think it's a good thing to kill your folks. That society would disappear after just a few generations.
Ever heard of mercy killing?
IMHO universal right and wrong is just an illusion we get from the abstract concepts the use of language impose on us. Ofcourse "Murder" is wrong, in the abstract sense, but as soon as you deal with an actual experienced situation, the choice of whether it is right or wrong will depend on the actual situation, and previous experiences of the subject deciding between right and wrong.
The nuns don't get Alzheimer's because they use their brains constantly, keeping diaries and staying active well into their later years.
My grandmother regularly read newspapers, and led an active social life. She even practised memorising long sequences of digits to keep her mind crisp. She now has Alzheimer's...
There could be a lot of other reasons for the phenomenon you describe. For instance, nuns have a lot of habits that differ from those most other people have. Maybe some special kind of diet, and plenty of physical exercise did the trick?!
On the other hand, maybe mental exercise does prevent Alzheimer's, but only in the statistical sense...
But even Darwin believed in a Creator God - and he said so himself.
This is what Darwin actually said about religion: http://www.update.uu.se/~fbend z/library/cd_relig.htm
A short excerpt:
Everyone who believes, as I do, that all the corporeal and mental organs (excepting those which are neither advantegous or disadvantegous to the posessor) of all beings have been developed through natural selection, or the survival of the fittest, together with use or habit[4], will admit that these organs have formed so that their possessors may compete succesfully with other beings, and thus increase in number.
Sure Darwin was religious, and sure he did believe in a "Creator God", but only before he set out with the Beagle.
See, those creationists may not be so kooky as you think!
It's always sound to be sceptical about all theories, that much I can agree with you. But, while most scientists that study evolution are sceptical, almost no creationists are. Disregarding all the details about the theories I find it far more open minded to study a phenomenon, and then formulate a theory, as Darwin did, than to adopt a dogma and then try to convince others...
The only thing I didn't like was when the little brat says "This is a UNIX system... I know this"
:-)
When I wathced JP on the first night it showed, the cinema was filled with computer science students. This line made over half the crowd go "Wooow!", and then laugh. The rest of the audience was left looking rather puzzled
I concluded however that the movie misses the point. The really frightening thing about DNA analysis is not that it makes prejudice based on DNA possible, prejudice is already deeply rooted in our society, but that that (unlike frenology which was just bullshit) prejudice based on genes would be right.
:-)
I'd rather conclude that you missed the point
Predjudice based on genes would be only partly right, but could easily be seen as completely justified (as you seem to think). My perception of the movie was that it was this danger it was out to pinpoint.
The fitness of a human is a combination of inherited and learned traits. This is why Ethan Hawke's character (sorry, forgot the name) was able to beat his brother despite his inferior physique: He had developed a stronger will-power.
I loved GATTACA, and found it a wonderful, and thought-provoking movie.
The other possibility would be to reduce the population; simply remove the industrial working class by randomly sterilizing a large number of people.
:-)
;) ), accountants etc.
This problem has already been dealt with by another insightful SF author: Douglas Adams.
Simply tell the redundant part of the population that the world is doomed, and send them to colonize another world
Seriously, I think this problem will solve itself, since robotics breakthroughs will not happen overnight. Most likely we will gradually have more people employed in service proffessions, like shop keepers, telephone sanitisers (couldn't resist
The problem is, that computers don't see. They look, but they don't really see.
No, exactly. And to make a computer really see, is what's known as the vision problem. My question (as you will see if you read it again) was: Why isn't this a suitable problem for AI?
Hmm... I've read some of the conversations there now. Judging from the ratio of gender words in the replies, I'd say that the system is either very human or operated by a guy/some guys in their twenties :-)
But it's a funny site. I recommend it!
Any other ideas for good AI benchmarks?
The benchmarks have to be there to encourage funding and some research, so some test needs to be decided as a standard.
How about RoboCup?
Computer vision is a decent test, but it has to be under such tight constraints. Other senses aren't worth the time, so we are left we only a few options.
This remark leaves me clueless. What's wrong with the vision problem? If you have a computer system that can see, wouldn't that be useful?
What are those constraints? If you are referring to the great need of processing power, I'd say thats more of a challenge (if the algorithms require too much processing power, then maybe they need rethinking)!
The problem with the Turing Test is that it tries to make a computer human...
:-)
Actually it does not even encourage us to make a computer human. It encourages us to make a computer program that produces sentences that sound human. Compare the size of the cortical areas devoted to speech processing, to the total area of the brain. This is my estimate of the Turing test's relevance
The general problem with the Turing test, as with most of the rest of the classical AI genre, is that it assumes that all relevant information processing should be symbolic. More likely only a small fraction of the information processing ought to be symbolic, the rest sub symbolic (ANN-s, fuzzy logic etc).
Look at ants, rabbits, dogs etc -- They cannot do symbolic information processing (cannot speak) but the feat they accomplish is still pretty impressive!
To train such a network would take years :-)
Seriously, the way to go in neural networks is most likely NOT to have 40,000-60,000 separate classes on the response side. Such a system cannot generalize at all, and space is wasted. ie. many links will be the same for similar words. There should be an intermediary representation (not a hidden layer, but a sparse/semi local representation of the output concepts).
White noise is certainly random - but background noise in real world situations is hardly going be that random. Rather, it's going to be a chaotic blend of non-random signals - each of which may (or may not) be a valid speech signal in it's own right.
;)
Actually it does not matter whether the background signal is completely white, or not. As long as the speech signal is the most correlated one, you can find it. The coctail party problem (to isolate one speech signal in a crowd of speakers) is of course more difficult. The technique can be extended to separate more sources, if one adds more microphones/ears (see independent component analysis), one extra microphone per source you want to isolate, but that would be to cheat, wouldn't it
The net has to know what it is listening for inside of the noise before it can actually pick it out.
NO, IT DOES NOT!!
It all comes down to statistics. Speech is a non-white signal. Noise is white. If you have two microphones/ears, you simply search for the linear combination of the two signals that is the most un-correlated temporally, and voala! You have found the speech signal. This is known as blind separation.
Yes, this passage about just 11 neurons connected by a mere 30 links makes me wonder what this net actually does. "Speech Recognition" could of course also mean the ability to recognize that an audio signal contains speech :-)
:-)
Of course the task of net could also be to separate the noise signal from the speech, aka blind separation, a problem that has been solved before (for instance by independent component analysis)
If this is merely ICA with a time coded neural net, it is IMHO still pretty cool, and much more impressive than all those commercial systems that rely on dumb correlation and processing power.
Anyway, instead of just having me guessing, could someone please point to their paper