I would still call my statement an untrue generalization in the strict sense--certainly some Americans who don't appreciate satire are not (possibly vocal) conservatives. Even classifying people in such generalized categories is problematic: what is a conservative, and what does it mean to be a vocal one? What does it mean to appreciate satire? Must you like all satire, even that which would deeply offend nearly everyone (think eating babies)? Is it in fact satire itself these people don't like, or its execution in specific instances--for instance, could they like Voltaire's Candide and hate South Park? Putting people into such large categories should be done cautiously as I believe history has shown. So, I marked my statement as a generalization to be taken with a grain of salt.
That said, I imagine my statement is "mostly true", whatever that means. Language is wonderful--we can express remarkably vague notions with it. Hopefully these vague notions have "content".
I don't have access to the paper (aside from the abstract I copied in another comment) and the article is lean on details, but perhaps machine learning took care of most of the issues people have been bringing up. Maybe "calm" callers with real emergencies show signs aside from overt stress reactions that the algorithms picked up on. For instance, I imagine I would speak particularly slowly and clearly if I were actively trying to be calm, and would speak faster otherwise. It's not clear what the 4.2% error rate refers to--incorrectly recognized urgency of 9-1-1 (or equivalent) calls? Incorrect detections of "strong emotion" in general? Without more info, I tend to give researchers the benefit of the doubt.
It looks like the copy&paste didn't quite work. The words "difficult", "benefit", and "effect" were spelled correctly, but used a single double-f/fi character instead of two characters. The other grammatical issues (sentence fragments and so on) were present on the page I linked.
The abundance of calls to emergency lines during crises is dicult to handle by the limited number of operators. Detecting if the caller is experiencing some extreme emotions can be a solution for distinguishing the more urgent calls. Apart from these and there are several other applications that can benet from awareness of the emotional state of the speaker. This paper describes the design of a system for selecting the calls that appear to be urgent and based on emotion detection. The system is trained using a database of spontaneous emotional speech from a call-centre. Four machine learning techniques are applied and based on either prosodic or spectral features and resulting in individual detectors. As a last stage and we investigate the eect of fusing these detectors into a single detection system. We observe an improvement in the Equal Error Rate (eer) from 19.0 % on average for 4 individual detectors to 4.2 % when fused using linear logistic regression. All experiments are performed in a speaker independent cross-validation framework.
In my experience (as an American) American perspectives vary wildly. Check out South Park or Family Guy; they're American shows that involve heavy satire. They're very popular with certain demographics, but they're hated, often vehemently, by other segments of the population. I think it would be more accurate to say there's a large, often vocal, conservative block in our population that doesn't appreciate satire. [Of course even this refinement is a generalization and isn't strictly true.]
As for me, I couldn't care less about the DKF multiplayer "controversy". It seems inconsistent to only get upset about hitting a woman/carrying her as a flag when the game is about killing people.
I think that's a different definition of "center of the universe" than the GP was implying (admittedly they were vague). Defining away the problem doesn't solve it.
Feynman's a very good counterexample (I didn't say there were none). Perelman though? Uhm, what? He's basically a hermit and never officially published his proof of the Poincare conjecture. How has he himself popularized his field? Note that there's a difference between generating press by solving a Millenium problem and actively popularizing one's field.
The properties of Euclidean space were basically made up from nothing. I don't see how adding time as a dimension (presumably you mean in spacetime) is any less natural than, say, completing the rationals to get the reals for use geometrically. It's all contrived in some sense. One hopes the end result is useful is all.
"Mathematically proved" would seem to imply something like reasoning that could be cast in formal logic, which I don't think you mean. It's an important point that scientific proof and mathematical proof are different. In science, you basically never prove things without any doubt--you prove likely error margins on data that supports or doesn't support your model. In math, the whole point is to prove things without any doubt whatsoever subject to the constraints of your system of logic/rules of deduction, and axioms. (Math proofs aren't usually written quite that formally since it's just incredibly tedious, but in principle it should be possible.)
I'd bet most of the top people in their field don't take the time to make their field publicly accessible. Steven Hawking comes to mind as a counterexample with a few books, but I can't think of a single mathematician counterexample. My point is Michio Kaku doesn't have to be a "top physicist", and I wouldn't even expect him to be. That he popularizes technical stuff is enough for me.
He probably has a good point, too, that at least eventually Moore's law failing will have strong economic impacts, and it's unlikely that an exponential law can continue indefinitely.
With the variation in people and how many of us there are, I'd be surprised if not one single person honestly changed their sexual preference throughout their lifetime. Maybe they get bored, or experience makes them appreciate different traits over time for whatever reason. I imagine it would be rare--significantly rarer than homosexuality/bisexuality, which themselves aren't *that* common. The continuity of one's sexuality depends a lot on the definition used, which is a subtle but probably pointless discussion. In general though, I tend to agree that true "converts" probably were just bisexual and never seriously tried the opposite sex.
That reminds me of something I've wondered before. There are people trying to turn gays straight. Does anyone try to do the reverse, i.e. turn straight people gay? If so, how successful is it?
The meta study that's the source behind your quote (abstract here) concludes that male sexual orientation and 2D:4D ratios are unrelated, but female sexual orientation and 2D:4D ratios are related. The GP's claims included both genders, so at least from the meta analysis some of them are plausible.
We know, we know, tons of people on/. love to hate MS. Enough. If you don't think they should be on this list, don't just accuse them of bribery--cite some instances of unethical behavior. Don't make them vague, make them specific and look-up-able, preferably with links. It would also be great if you would cite how MS is worse than companies not on the list.
To be clear, I'm not necessarily supporting Microsoft. I'm just tired of idiots spouting hatred that makes them cool in the eyes of their friends, rather than injecting real content into the discussion. (Personally I don't have nearly enough information to have an opinion. I didn't look at thousands of companies to compare their records on ethics--who am I to comment on such a list?)
I agree, I wouldn't call arithmetic an "unsolvable" problem, but I might call summing two arbitrary 10^12345 digit numbers "unsolvable with conventional computers" (which is how the summary was worded). But my 10^12345 point was just illustrating that "Computers in the real world can currently solve every NP-complete problem" is not true by giving an instance that real world computers cannot solve. Only an abstract idealization of real world computers can solve such a problem.
Hah! I can just imagine a future AI examining the long-running computer's code, saying "this is really inefficient", swapping the algorithm mid-computation, and finishing it off a few seconds later.
No, they can't. Solve a 10^12345 city TSP problem on current hardware for me, please. It's not possible, since it won't even fit in memory, disregarding problems of the hardware breaking down/the sun exploding before the problem finishes/etc. This is the distinction between abstract algorithms and real-world implementations I was getting at. When you say "computers in the real world can currently solve every NP-complete problem with current algorithms", what you mean is that abstract Turing machines, which real-world computers approximate, can solve every NP-complete problem with current algorithms.
"Uncomputable" has a well-defined meaning in CS. I don't recall my CS courses using "unsolvable" as a synonym. I did a little searching and a few people seem to use them as synonyms, though they're mostly older sources or if they're current they're in the vast minority. "Solvable" and "effectively solvable" seem to be in current use, but "unsolvable" doesn't seem to be.
Slashdot Mirror
I would still call my statement an untrue generalization in the strict sense--certainly some Americans who don't appreciate satire are not (possibly vocal) conservatives. Even classifying people in such generalized categories is problematic: what is a conservative, and what does it mean to be a vocal one? What does it mean to appreciate satire? Must you like all satire, even that which would deeply offend nearly everyone (think eating babies)? Is it in fact satire itself these people don't like, or its execution in specific instances--for instance, could they like Voltaire's Candide and hate South Park? Putting people into such large categories should be done cautiously as I believe history has shown. So, I marked my statement as a generalization to be taken with a grain of salt.
That said, I imagine my statement is "mostly true", whatever that means. Language is wonderful--we can express remarkably vague notions with it. Hopefully these vague notions have "content".
I don't have access to the paper (aside from the abstract I copied in another comment) and the article is lean on details, but perhaps machine learning took care of most of the issues people have been bringing up. Maybe "calm" callers with real emergencies show signs aside from overt stress reactions that the algorithms picked up on. For instance, I imagine I would speak particularly slowly and clearly if I were actively trying to be calm, and would speak faster otherwise. It's not clear what the 4.2% error rate refers to--incorrectly recognized urgency of 9-1-1 (or equivalent) calls? Incorrect detections of "strong emotion" in general? Without more info, I tend to give researchers the benefit of the doubt.
It looks like the copy&paste didn't quite work. The words "difficult", "benefit", and "effect" were spelled correctly, but used a single double-f/fi character instead of two characters. The other grammatical issues (sentence fragments and so on) were present on the page I linked.
The abundance of calls to emergency lines during crises is dicult to handle by the limited number of operators. Detecting if the caller is experiencing some extreme emotions can be a solution for distinguishing the more urgent calls. Apart from these and there are several other applications that can benet from awareness of the emotional state of the speaker. This paper describes the design of a system for selecting the calls that appear to be urgent and based on emotion detection. The system is trained using a database of spontaneous emotional speech from a call-centre. Four machine learning techniques are applied and based on either prosodic or spectral features and resulting in individual detectors. As a last stage and we investigate the eect of fusing these detectors into a single detection system. We observe an improvement in the Equal Error Rate (eer) from 19.0 % on average for 4 individual detectors to 4.2 % when fused using linear logistic regression. All experiments are performed in a speaker independent cross-validation framework.
Taken from here. The International Journal of Intelligent Defence Support Systems web site doesn't appear to be updated with this year's 2nd volume yet. [The English is a bit clunky, but the researchers appear to be Dutch so I forgive them that, at least.]
In my experience (as an American) American perspectives vary wildly. Check out South Park or Family Guy; they're American shows that involve heavy satire. They're very popular with certain demographics, but they're hated, often vehemently, by other segments of the population. I think it would be more accurate to say there's a large, often vocal, conservative block in our population that doesn't appreciate satire. [Of course even this refinement is a generalization and isn't strictly true.]
As for me, I couldn't care less about the DKF multiplayer "controversy". It seems inconsistent to only get upset about hitting a woman/carrying her as a flag when the game is about killing people.
I think that's a different definition of "center of the universe" than the GP was implying (admittedly they were vague). Defining away the problem doesn't solve it.
Feynman's a very good counterexample (I didn't say there were none). Perelman though? Uhm, what? He's basically a hermit and never officially published his proof of the Poincare conjecture. How has he himself popularized his field? Note that there's a difference between generating press by solving a Millenium problem and actively popularizing one's field.
The properties of Euclidean space were basically made up from nothing. I don't see how adding time as a dimension (presumably you mean in spacetime) is any less natural than, say, completing the rationals to get the reals for use geometrically. It's all contrived in some sense. One hopes the end result is useful is all.
"Mathematically proved" would seem to imply something like reasoning that could be cast in formal logic, which I don't think you mean. It's an important point that scientific proof and mathematical proof are different. In science, you basically never prove things without any doubt--you prove likely error margins on data that supports or doesn't support your model. In math, the whole point is to prove things without any doubt whatsoever subject to the constraints of your system of logic/rules of deduction, and axioms. (Math proofs aren't usually written quite that formally since it's just incredibly tedious, but in principle it should be possible.)
I'd bet most of the top people in their field don't take the time to make their field publicly accessible. Steven Hawking comes to mind as a counterexample with a few books, but I can't think of a single mathematician counterexample. My point is Michio Kaku doesn't have to be a "top physicist", and I wouldn't even expect him to be. That he popularizes technical stuff is enough for me.
He probably has a good point, too, that at least eventually Moore's law failing will have strong economic impacts, and it's unlikely that an exponential law can continue indefinitely.
This is OT, but did your buggy whip reference come from the movie Other People's Money?
With the variation in people and how many of us there are, I'd be surprised if not one single person honestly changed their sexual preference throughout their lifetime. Maybe they get bored, or experience makes them appreciate different traits over time for whatever reason. I imagine it would be rare--significantly rarer than homosexuality/bisexuality, which themselves aren't *that* common. The continuity of one's sexuality depends a lot on the definition used, which is a subtle but probably pointless discussion. In general though, I tend to agree that true "converts" probably were just bisexual and never seriously tried the opposite sex.
That reminds me of something I've wondered before. There are people trying to turn gays straight. Does anyone try to do the reverse, i.e. turn straight people gay? If so, how successful is it?
The meta study that's the source behind your quote (abstract here) concludes that male sexual orientation and 2D:4D ratios are unrelated, but female sexual orientation and 2D:4D ratios are related. The GP's claims included both genders, so at least from the meta analysis some of them are plausible.
We know, we know, tons of people on /. love to hate MS. Enough. If you don't think they should be on this list, don't just accuse them of bribery--cite some instances of unethical behavior. Don't make them vague, make them specific and look-up-able, preferably with links. It would also be great if you would cite how MS is worse than companies not on the list.
To be clear, I'm not necessarily supporting Microsoft. I'm just tired of idiots spouting hatred that makes them cool in the eyes of their friends, rather than injecting real content into the discussion. (Personally I don't have nearly enough information to have an opinion. I didn't look at thousands of companies to compare their records on ethics--who am I to comment on such a list?)
Ah, good point, though that was a season 1 episode. Like most of season 1, it's probably best forgotten ;).
R2D2 could understand speech but not speak.
Maybe he was just too cool for primitive speech. He always did seem to have an attitude.
I'm curious, how do you think they misused Data?
I like base 16 in light of the BBP formula. The formula gives base 16 digits of pi, without computing the previous digits.
slashslashslashdot anyone?
Three slashes?
I agree, I wouldn't call arithmetic an "unsolvable" problem, but I might call summing two arbitrary 10^12345 digit numbers "unsolvable with conventional computers" (which is how the summary was worded). But my 10^12345 point was just illustrating that "Computers in the real world can currently solve every NP-complete problem" is not true by giving an instance that real world computers cannot solve. Only an abstract idealization of real world computers can solve such a problem.
Hah! I can just imagine a future AI examining the long-running computer's code, saying "this is really inefficient", swapping the algorithm mid-computation, and finishing it off a few seconds later.
No, they can't. Solve a 10^12345 city TSP problem on current hardware for me, please. It's not possible, since it won't even fit in memory, disregarding problems of the hardware breaking down/the sun exploding before the problem finishes/etc. This is the distinction between abstract algorithms and real-world implementations I was getting at. When you say "computers in the real world can currently solve every NP-complete problem with current algorithms", what you mean is that abstract Turing machines, which real-world computers approximate, can solve every NP-complete problem with current algorithms.
"Uncomputable" has a well-defined meaning in CS. I don't recall my CS courses using "unsolvable" as a synonym. I did a little searching and a few people seem to use them as synonyms, though they're mostly older sources or if they're current they're in the vast minority. "Solvable" and "effectively solvable" seem to be in current use, but "unsolvable" doesn't seem to be.