Are you a follower of Ray Kurzweil, by any chance? I only ask because I don't often see Searle dismissed outright by anyone competent unless they also happen to be a singularity nut.
I ask because I actually know people who are starting to demonstrate the rudiments of intelligence using simulations of ~100,000 neurons.
Per upthread, that's a long way from a brain, and in fact we don't even know how all of the brain is wired, let alone how it works. But you might want to consider this and this and this.
If they're attempting the impossible, you should let them know not to waste their money.
What's actually new in the neural net business? That's a real question - not a sarcastic or rhetorical one.
What's reportedly new is the ability to train feed-forward networks with many layers. They have never trained well with backpropagation because the backpropagated error estimate becomes "diluted" the further back it goes, and as a result most of the training happens to the weights closest to the output end.
The notion that the first hidden layer is a low-level feature detector and each successive layer is a higher-level feature layer is ancient lore in the ANN research community. The claims of the Deep Learning people is that they can actually make it work on deep networks.
IMO their techniques sound very plausible, but I say "reportedly" above, because I don't know whether the methods are actually delivering on expectations.
Artificial neural nets were suggested and tried for AI at least 50 years ago. They were bashed by the old Minsky/McCarthy AI crowd, who didn't like the competition's idea (always better to write another million lines of Lisp).
There has been a lot of bad blood between camps in the machine intelligence research community because the ANN guys have never gotten over the suspicion that Minsky & Papiert's book was a hit job on ANNs. However, it said a lot of nice things about them, in addition to pointing out their limitations. And exposing those limitations shouldn't have had the effect they had, because we already knew that networks of perceptrons could do things that individual perceptrons cannot.
Interestingly, Hinton is one of the people who rehabilitated ANNs in the mid-1980s, as co-author of the very influential Parallel Distributed Processing (PDP) book. (It made the backpropagation algorithm well known, though IIRC it had been invented independently a couple of times before then.)
More recently, SVNn have brought on another winter for ANNs, and here's the same Hinton breathing new life into the field. Good luck to him, if he pulls it off twice in one lifetime.
So what's really improved w/ neural nets these days?
So far as this story is concerned, the news is that people claim to be able to train deeply layered networks with autoassociative methods to produce a hierarchy of feature detectors that have Amazing Powers(tm) for pattern recognition. But per what I said above, I think it's too soon to say whether those claims should be interpreted as facts, expectations, or hype.
First of all, while they have been inspired by the brain, they don't "mimic" it.
That is true.
Neural networks are based on some neurons having negative weights, reversing the polarity of the signal, which doesn't happen in the brain.
There are in fact inhibitory connections in the brain.
They are also linear
That is false. The only way an ANN could be linear is if each "neuron" used a squashing function f(x) = x. Then they'd just be doing linear algebra, namely change-of-basis computations. But no one uses that. Even the super-simple heaviside squash used in the 1950s perceptrons made them do nonlinear computations.
Second neural networks are only good at things when they have to immediately react to an input. Originally, neural networks didn't have memory, and while it's possible to add it, it doesn't fit right into the system and is hard to work with. While neural networks make good reflex machines, even simple stateful tasks like a linear or cyclic multi-step motion are nontrivial to implement in them.
That is also false. I suspect there are limits to what kind of stateful computations you can do with an ANN, but you can certainly do some of them. For example, the POMDP version of the pole balancing problem got so easy to solve with neuroevolution that no one even uses it for a benchmark anymore.
Historical linguists basically laughed Renfrew out of town for his 1987 "out of Anatolia" hypothesis about Indo-European origins.
Also, he is an archaeologist, not a linguist. IMO archeologists know exactly diddly about historical linguistics, and reveal it almost every time they say anything on the topic.
There is a strong suspicion that the m-words for mother and p-words for father arise cross-culturally because they are both labial articulations, and an infant can easily see how you are doing the articulation.
Some anthropologists think our ancestors already "had language" when our species began to spread around the world. If so, it may be that every language in the world is related. (The alternative being that language was invented independently more than once, and that more than one lineage has survived to the present.)
The problem is how you demonstrate it rigorously. Every historical linguist accepts the relatedness of languages in 5000-year-old families. But for proposed older relations (e.g., Nostratic, 10,000-15,000 ybp), the number of linguists that accept them is pretty much inversely proportional to the time depth.
As one of the linked summary articles points out, the further back you go the less evidence you have (lexical replacement), and the more noise (spurious similarities arising from chance). Beyond a certain point you just can't demonstrate relatedness reliably, though exactly what that point is is up for debate.
You would expect a few out of sheer randomness. Especially when you're using a vague notion of similarity.
That's why most historical linguists utterly reject Greenberg's mass-comparison method. (And why cranks latch on to it: they can use it to "prove" any language relationship they care to peddle.)
The New Scientist article -- whoops, guest editorial -- is titled "Psychiatry divided as mental health 'bible' denounced", but 'denounced' is a ridiculous overstatement. NIH/NIMH are simply announcing a new cross-category funding program that will step back and question the field's traditional assumptions.
Either the guest editorialist didn't RTFA, or else is just using the occasion to inject their personal views into public sight.
Slashdot Mirror
This crap is for week-minded fools who lack the will power
to abstain from time-wasting activities.
...he posted to Slashdot.
We can tell you've got teh True Geek... ...a normal guy would have made it make the girls' knickers fall down.
Bah, details!
We've always been at war with Eurasia.
Are you a follower of Ray Kurzweil, by any chance? I only ask because I don't often see Searle dismissed outright by anyone competent unless they also happen to be a singularity nut.
No, I think Kurzweil is a crank.
They should have used wireless...
Are you claiming that symbol grounding is a non-solvable problem?
the north pole's ice cover disappears every year, so for those people the ice age comes and goes with the school year.
Actually it doesn't completely disappear - yet.
But that's exactly what I was talking about.
It proved useful a few times.
For example?
What precisely are those long-standing problems?
I ask because I actually know people who are starting to demonstrate the rudiments of intelligence using simulations of ~100,000 neurons.
Per upthread, that's a long way from a brain, and in fact we don't even know how all of the brain is wired, let alone how it works. But you might want to consider this and this and this.
If they're attempting the impossible, you should let them know not to waste their money.
What's actually new in the neural net business? That's a real question - not a sarcastic or rhetorical one.
What's reportedly new is the ability to train feed-forward networks with many layers. They have never trained well with backpropagation because the backpropagated error estimate becomes "diluted" the further back it goes, and as a result most of the training happens to the weights closest to the output end.
The notion that the first hidden layer is a low-level feature detector and each successive layer is a higher-level feature layer is ancient lore in the ANN research community. The claims of the Deep Learning people is that they can actually make it work on deep networks.
IMO their techniques sound very plausible, but I say "reportedly" above, because I don't know whether the methods are actually delivering on expectations.
Artificial neural nets were suggested and tried for AI at least 50 years ago. They were bashed by the old Minsky/McCarthy AI crowd, who didn't like the competition's idea (always better to write another million lines of Lisp).
There has been a lot of bad blood between camps in the machine intelligence research community because the ANN guys have never gotten over the suspicion that Minsky & Papiert's book was a hit job on ANNs. However, it said a lot of nice things about them, in addition to pointing out their limitations. And exposing those limitations shouldn't have had the effect they had, because we already knew that networks of perceptrons could do things that individual perceptrons cannot.
Interestingly, Hinton is one of the people who rehabilitated ANNs in the mid-1980s, as co-author of the very influential Parallel Distributed Processing (PDP) book. (It made the backpropagation algorithm well known, though IIRC it had been invented independently a couple of times before then.)
More recently, SVNn have brought on another winter for ANNs, and here's the same Hinton breathing new life into the field. Good luck to him, if he pulls it off twice in one lifetime.
So what's really improved w/ neural nets these days?
So far as this story is concerned, the news is that people claim to be able to train deeply layered networks with autoassociative methods to produce a hierarchy of feature detectors that have Amazing Powers(tm) for pattern recognition. But per what I said above, I think it's too soon to say whether those claims should be interpreted as facts, expectations, or hype.
First of all, while they have been inspired by the brain, they don't "mimic" it.
That is true.
Neural networks are based on some neurons having negative weights, reversing the polarity of the signal, which doesn't happen in the brain.
There are in fact inhibitory connections in the brain.
They are also linear
That is false. The only way an ANN could be linear is if each "neuron" used a squashing function f(x) = x. Then they'd just be doing linear algebra, namely change-of-basis computations. But no one uses that. Even the super-simple heaviside squash used in the 1950s perceptrons made them do nonlinear computations.
Second neural networks are only good at things when they have to immediately react to an input. Originally, neural networks didn't have memory, and while it's possible to add it, it doesn't fit right into the system and is hard to work with. While neural networks make good reflex machines, even simple stateful tasks like a linear or cyclic multi-step motion are nontrivial to implement in them.
That is also false. I suspect there are limits to what kind of stateful computations you can do with an ANN, but you can certainly do some of them. For example, the POMDP version of the pole balancing problem got so easy to solve with neuroevolution that no one even uses it for a benchmark anymore.
I post as an anonymous coward so as not to harm my career (any further) by stating *truth* which is not politically nor academically correct.
Also, you wisely don't want your name associated with a positive view of Searle's nonsense.
Neural networks? Is it news?
No, it's misrepresentation. This isn't any more akin to neuroscience than any of the other techniques used with artificial neural networks.
It will be a great thing, though, if it lives up to expectations.
As long as there are still polar ice sheets, the ice age hasn't ended.
If you insist on the plural, the ice age will be ending pretty soon.
Unga bunga
That has evolved to cowabunga. We conclude that 'ung' is the ancient word for cow.
Historical linguists basically laughed Renfrew out of town for his 1987 "out of Anatolia" hypothesis about Indo-European origins.
Also, he is an archaeologist, not a linguist. IMO archeologists know exactly diddly about historical linguistics, and reveal it almost every time they say anything on the topic.
There is a strong suspicion that the m-words for mother and p-words for father arise cross-culturally because they are both labial articulations, and an infant can easily see how you are doing the articulation.
Some anthropologists think our ancestors already "had language" when our species began to spread around the world. If so, it may be that every language in the world is related. (The alternative being that language was invented independently more than once, and that more than one lineage has survived to the present.)
The problem is how you demonstrate it rigorously. Every historical linguist accepts the relatedness of languages in 5000-year-old families. But for proposed older relations (e.g., Nostratic, 10,000-15,000 ybp), the number of linguists that accept them is pretty much inversely proportional to the time depth.
As one of the linked summary articles points out, the further back you go the less evidence you have (lexical replacement), and the more noise (spurious similarities arising from chance). Beyond a certain point you just can't demonstrate relatedness reliably, though exactly what that point is is up for debate.
This is a pretty lame summary. If there are words preserved from the Ice Age, list like five of them!
Or give us the Iceageish translation for "Jeez, it's cold out there."
Take this one with a grain of salt...
Really. They're showing that two words are related, without reference to what the words actually are? Oh, please...
Too bad I can't see the article. I suspect that they're capturing some interesting properties of language (in the abstract, not "languages").
OTOH, maybe they're just showing that lots of languages have a word for "I". The descriptions in the summaries are pretty vague about their methods.
You would expect a few out of sheer randomness. Especially when you're using a vague notion of similarity.
That's why most historical linguists utterly reject Greenberg's mass-comparison method. (And why cranks latch on to it: they can use it to "prove" any language relationship they care to peddle.)
The New Scientist article -- whoops, guest editorial -- is titled "Psychiatry divided as mental health 'bible' denounced", but 'denounced' is a ridiculous overstatement. NIH/NIMH are simply announcing a new cross-category funding program that will step back and question the field's traditional assumptions.
Either the guest editorialist didn't RTFA, or else is just using the occasion to inject their personal views into public sight.
Or else just trolling.
Were already well known. Considering we don't know too much about the organic causes of most mental disorders I'm curious about what they mean.
Good point. If they're launching this as a research initiative, we should expect a decade or more before it starts traditional methods.
Careful there, paranoia is already established as a mental illness.