You gave a good summary of the article's logic, but TFA's author is wrong to exclude the 7th younger boy possibility. The author assumes that any case where two boys are born on Tuesday is identical, and so duplicates should be removed; this is not true.
It's simply not the case that there are duplicates. There is just one event there are two boys who are born on a Tuesday.
<eldest boy born on Tuesday, youngest boy born on Tuesday>
There are two events where there are two boys and one is born on a Monday and the other is born on a Tuesday.
<eldest boy born on Monday, youngest boy born on Tuesday>
<eldest boy born on Tuesday, youngest boy born on Monday>
Your confusion arises from introducing a new random variable (whether you've met X or not) and enumerating over this, but only for the both-born-on-a-Tuesday event. You would need to do this for all the other events as well; this would have the effect of needlessly multiplying events but wouldn't change the relative frequencies of the original events.
You are introducing probabilities too early. We start by just counting events. This will lead in turn to frequencies and thence to probabilities.
So: there is only one event where two boys are born on Tuesday, and it is this:
<Eldest boy is born on Tuesday, Youngest boy is born on Tuesday>
There are no other events where both boys are born on Tuesday. There are two events where there are two boys who are born on a Monday and Tuesday, and they are these:
<Eldest boy is born on Monday, Youngest boy is born on Tuesday> <Eldest boy is born on Tuesday, Youngest boy is born on Monday>
Note that there are only 27 outcomes as both boys being born on a Tuesday is only one outcome, not two. There are 13 outcomes labelled BB, so the probability of having two boys is 13/27. Hope that's clear now.
First, The question doesn't say the other (this does not mean older or younger...) child was not born on a Tuesday, maybe the questioner meant to include this info but they failed to.
It's not part of the problem that only one of the children can be born on a Tuesday, you might be dealing with two boys of differing ages, both of whom were born on a Tuesday. In Keith Devlin's analysis (which is correct, given certain assumptions about the problem), there is only one outcome where there are two boys born on a Tuesday. This is why, once you've accounted for this by enumerating the outcomes where the elder child is a boy born on Tuesday, you can't count it again when enumerating the outcomes for the younger child.
What's with all the pessimism? Strong AI is a matter of inevitability.
Sadly, we're not even in a position where we can say whether it's inevitable or not. We don't know what strong AI is or how we could tell if we had it (the Turing Test isn't a good way of evaluating this). I agree that it's short-sighted to claim that we'll never develop strong AI, but you're wrong to claim that it's inevitable.
Having spent over ten years working professionally in the field, I've found this kind of thing to be all too common, especially among ambitious-but-talentless academics where grant applications are concerned, particularly where said grant applications contain the words "semantic" and "web" in close proximity - now that really is pure, undiluted snake oil. The semantic web community has received hundreds of millions of dollars/euros/pounds in funding and they've delivered nothing of any use. Zippo, zilch, zero. Compare this with enormous amounts of useful functionality delivered by the machine learning community. The difference is that machine learning is rigorous and can be really quite difficult, whereas the semantic web is based on the belief that 3-tuples are really neat.
It only seems weird if you fail to recognise that decisions that politicians have an effect on the future.
Take the current banking crisis in the UK. As far as I can tell, many people who know far more about this kind of thing than I do, say that this is directly attributable to the stripping away of regulatory control of the financial sector that happened in the 80s.
Now, whether or not this is true, it is clearly plausible that the current economic shitstorm is directly attributable to the economic policies of Thatcher.
You don't need to store massive amounts of electricity, just massive amounts of energy, in whatever form. That's what pumped-storage hydroelectric power stations do - they pump water uphill overnight using the cheap electricity produced by the coal and nuclear power stations that can't be turned off, and then let it fall during the day, driving the turbines.
+1 for the book by Bishop (don't know about the others). In addition, have a look at Information Theory by David Mackay which I found stunningly good. There is a free on-line version available, but you should buy it:
http://www.inference.phy.cam.ac.uk/itprnn/book.html
The best argument against this kind of ridiculous assertion that somehow random information will somehow give rise to intelligence is provided in the old movie Short Circuit.
I agree. I don't know why Searle bothered with all that Chinese Room nonsense. The answers to all of the great philosophical questions of our age are to be found in the movies of Steve Guttenberg.
Hmmm, either I meant (a) "the academic world isn't a reality" (which would clearly be a ludicrous claim), or (b) something else. Guess what? It's (b). The semantic web is not grounded in the real world in that it does not take into account that real people (as opposed to the kind of people that exist in research funding proposals) generally make decisions about how much time or money they invest in something based on the benefit they expect to obtain from that effort. It just seems hugely unlikely that people are going to decide to use "semantic" mark-up (by which I mean mark-up that is grounded in some formally-defined ontology which has been committed to by some community) when there seems very little benefit in them doing so. Will it improve their page-rank? Not at the moment. Some people are trying to bootstrap the semantic web by automatically annotating existing bodies of knowledge (e.g. Wikipedia). The obvious problem with this is that there is no way it's going to of sufficient quality (current NLP just isn't good enough) and this lack of precision completely wipes out the claimed advantages of the semantic web, which is that semantic annotations will bring increased quality of search results for more specific kinds of search queries.
it's precisely in academic circles that the semantic web is already a reality But this is precisely the problem with the semantic web - it's not grounded in reality (apologies for the poor logic pun). I'm not going to indulge in academic-bashing, having spent a fair amount of time wandering up and down ivory towers in my time, but the semantic web does not measure up to even a cursory cost-benefit analysis. It provides very little qualitative benefit over current search - especially for content providers, who unfortunately are the people it requires a huge commitment from.
Absolutely right, and it's not as if these academic chancers don't have form in this area - the importance of this "breakthrough" can indeed be overestimated, just like previous "breakthroughs"). The semantic web is nothing more than snake oil.
Agreed with the "most of it is garbage". See the Language Log for informed debunking of 45 and 57 (btw, the Language Log is highly recommended. In the best Reithian traditions, it manages to "inform, educate and entertain" - something the BBC website would do well to take note of). The quality of science and technology reporting on the BBC (and this goes for all news organisations that I'm aware of) is just pitiful.
Well, I could recapitulate Searle's counter to this (imagine the man memorising the book) but I'm sure we're not going to resolve this ancient debate here.
I think I must have this wrong, but the only interpretation I can think of is that you're saying that the "whole system" (the man and the rulebook) understands Chinese?
Couldn't agree more with the parent. I used to work in the AI department of BT's research labs, and this guy was a constant embarassment to us with his ill-informed drivel. We'd try hard to build some kind of reputation in the field, and this moron would undo it all with his "robots will destroy humanity by the middle of next week" toss. He's like a less-scientific Captain Cyborg (if such a thing is possible).
That's not the point about the Chinese Room. It's an analogy, where the man in the room corresponds to a symbol-manipulating computer. So the question is not whether "the system as a whole" understands Chinese but whether the man in the room does. He doesn't.
Slashdot Mirror
You gave a good summary of the article's logic, but TFA's author is wrong to exclude the 7th younger boy possibility. The author assumes that any case where two boys are born on Tuesday is identical, and so duplicates should be removed; this is not true.
It's simply not the case that there are duplicates. There is just one event there are two boys who are born on a Tuesday.
<eldest boy born on Tuesday, youngest boy born on Tuesday>
There are two events where there are two boys and one is born on a Monday and the other is born on a Tuesday.
<eldest boy born on Monday, youngest boy born on Tuesday>
<eldest boy born on Tuesday, youngest boy born on Monday>
Your confusion arises from introducing a new random variable (whether you've met X or not) and enumerating over this, but only for the both-born-on-a-Tuesday event. You would need to do this for all the other events as well; this would have the effect of needlessly multiplying events but wouldn't change the relative frequencies of the original events.
You are introducing probabilities too early. We start by just counting events. This will lead in turn to frequencies and thence to probabilities.
So: there is only one event where two boys are born on Tuesday, and it is this:
<Eldest boy is born on Tuesday,
Youngest boy is born on Tuesday>
There are no other events where both boys are born on Tuesday. There are two events where there are two boys who are born on a Monday and Tuesday, and they are these:
<Eldest boy is born on Monday,
Youngest boy is born on Tuesday>
<Eldest boy is born on Tuesday,
Youngest boy is born on Monday>
That's all there is to it.
Please point out where this analysis goes astray.
Gladly. The available outcomes are actually:
Note that there are only 27 outcomes as both boys being born on a Tuesday is only one outcome, not two. There are 13 outcomes labelled BB, so the probability of having two boys is 13/27. Hope that's clear now.
First, The question doesn't say the other (this does not mean older or younger...) child was not born on a Tuesday, maybe the questioner meant to include this info but they failed to.
It's not part of the problem that only one of the children can be born on a Tuesday, you might be dealing with two boys of differing ages, both of whom were born on a Tuesday. In Keith Devlin's analysis (which is correct, given certain assumptions about the problem), there is only one outcome where there are two boys born on a Tuesday. This is why, once you've accounted for this by enumerating the outcomes where the elder child is a boy born on Tuesday, you can't count it again when enumerating the outcomes for the younger child.
What's with all the pessimism? Strong AI is a matter of inevitability.
Sadly, we're not even in a position where we can say whether it's inevitable or not. We don't know what strong AI is or how we could tell if we had it (the Turing Test isn't a good way of evaluating this). I agree that it's short-sighted to claim that we'll never develop strong AI, but you're wrong to claim that it's inevitable.
Having taken several courses on AI, I never found a contributor to the field that promised it to be the silver bullet
<cough>Ontologies: A Silver Bullet for Knowledge Management and Electronic Commerce</cough>
Having spent over ten years working professionally in the field, I've found this kind of thing to be all too common, especially among ambitious-but-talentless academics where grant applications are concerned, particularly where said grant applications contain the words "semantic" and "web" in close proximity - now that really is pure, undiluted snake oil. The semantic web community has received hundreds of millions of dollars/euros/pounds in funding and they've delivered nothing of any use. Zippo, zilch, zero. Compare this with enormous amounts of useful functionality delivered by the machine learning community. The difference is that machine learning is rigorous and can be really quite difficult, whereas the semantic web is based on the belief that 3-tuples are really neat.
It seems that you forgot about the use of 'sie' for the third person plural, so the GP's quote was orthographically valid.
It only seems weird if you fail to recognise that decisions that politicians have an effect on the future. Take the current banking crisis in the UK. As far as I can tell, many people who know far more about this kind of thing than I do, say that this is directly attributable to the stripping away of regulatory control of the financial sector that happened in the 80s. Now, whether or not this is true, it is clearly plausible that the current economic shitstorm is directly attributable to the economic policies of Thatcher.
You don't need to store massive amounts of electricity, just massive amounts of energy, in whatever form. That's what pumped-storage hydroelectric power stations do - they pump water uphill overnight using the cheap electricity produced by the coal and nuclear power stations that can't be turned off, and then let it fall during the day, driving the turbines.
How about the state Bush left the fucking *world* in when he left office? Asshole.
+1 for the book by Bishop (don't know about the others). In addition, have a look at Information Theory by David Mackay which I found stunningly good. There is a free on-line version available, but you should buy it: http://www.inference.phy.cam.ac.uk/itprnn/book.html
For a really good dissection of that critique of map-reduce, have a look at the following: http://scienceblogs.com/goodmath/2008/01/databases_are_hammers_mapreduc.php
The best argument against this kind of ridiculous assertion that somehow random information will somehow give rise to intelligence is provided in the old movie Short Circuit.
I agree. I don't know why Searle bothered with all that Chinese Room nonsense. The answers to all of the great philosophical questions of our age are to be found in the movies of Steve Guttenberg.
Hmmm, either I meant (a) "the academic world isn't a reality" (which would clearly be a ludicrous claim), or (b) something else. Guess what? It's (b). The semantic web is not grounded in the real world in that it does not take into account that real people (as opposed to the kind of people that exist in research funding proposals) generally make decisions about how much time or money they invest in something based on the benefit they expect to obtain from that effort. It just seems hugely unlikely that people are going to decide to use "semantic" mark-up (by which I mean mark-up that is grounded in some formally-defined ontology which has been committed to by some community) when there seems very little benefit in them doing so. Will it improve their page-rank? Not at the moment. Some people are trying to bootstrap the semantic web by automatically annotating existing bodies of knowledge (e.g. Wikipedia). The obvious problem with this is that there is no way it's going to of sufficient quality (current NLP just isn't good enough) and this lack of precision completely wipes out the claimed advantages of the semantic web, which is that semantic annotations will bring increased quality of search results for more specific kinds of search queries.
That's no cliché. It's a snowclone.
Worst article ever.
Well, your IQ may be 140, but you don't understand IQ tests or probability distributions if you think 2*IQ == twice as smart.
Absolutely right, and it's not as if these academic chancers don't have form in this area - the importance of this "breakthrough" can indeed be overestimated, just like previous "breakthroughs"). The semantic web is nothing more than snake oil.
Agreed with the "most of it is garbage". See the Language Log for informed debunking of 45 and 57 (btw, the Language Log is highly recommended. In the best Reithian traditions, it manages to "inform, educate and entertain" - something the BBC website would do well to take note of). The quality of science and technology reporting on the BBC (and this goes for all news organisations that I'm aware of) is just pitiful.
It's all very well these hucksters peddling the semantic web to funding bodies who don't know any better, as long as they don't start pretending it's anything other than the new FIPA - a collection of committees generating specifications that the world will continue to ignore.
Well, I could recapitulate Searle's counter to this (imagine the man memorising the book) but I'm sure we're not going to resolve this ancient debate here.
I think I must have this wrong, but the only interpretation I can think of is that you're saying that the "whole system" (the man and the rulebook) understands Chinese?
Couldn't agree more with the parent. I used to work in the AI department of BT's research labs, and this guy was a constant embarassment to us with his ill-informed drivel. We'd try hard to build some kind of reputation in the field, and this moron would undo it all with his "robots will destroy humanity by the middle of next week" toss. He's like a less-scientific Captain Cyborg (if such a thing is possible).
That's not the point about the Chinese Room. It's an analogy, where the man in the room corresponds to a symbol-manipulating computer. So the question is not whether "the system as a whole" understands Chinese but whether the man in the room does. He doesn't.