This is a nice paper by respected researchers in AI+Vision, however pretty much the entire content of the journal this was published in (IEEE Pattern Analysis and Machine Intelligence) is up to that level. Why single out that particular paper ?
Interested readers can browse the content of PAMI current and back issues and either go to their local scientific library (PAMI is recognisable from afar by its bright yellow cover) or search on the web for interesting articles. Often researchers put their own paper on their home page. For example, here is the publication page of one of the authors (I'm not him).
For the record, I think justifying various ad-hoc vision/image analysis techniques using approximations of biological underpining is of limited interest. When asked if computer would think one day, Edsgerd Dijkstra famously answered by "can submarine swim?". In the same manner, it has been observed that (for example) most neural network architectures make worse classifiers than standard logistic regression, not to mention Support Vector Machines, which what this article uses BTW.
The summary by our friend Roland P. is not very good:
This versatile model could one day be used for automobile driver's assistance, visual search engines, biomedical imaging analysis, or robots with realistic vision
There already exist working automated driving software. The december 2006 issue of IEEE Computers magazing was on them last month. Read about the car that drove a thousand miles on Italy's road thanks to Linux, no less.
Visual search engine exist, at the research level. The whole field is called "Content Based Retrieval", and the main issue is not so much to search, but to formulate the question.
Biomedical image analysis has been going strong for decades and is used every day in your local hospital. Ask your doctor !
Robotic vision is pretty much as old as computers themselves. There are even fun robot competitions like robocup.
I could go on with lists and links but the future is already here, generally inconspicuously. Read about it.
This is not a general-purpose quantum computer, it is supposedly designed to solve quadratic integer programming (efficiently), which is NP-complete, but I don't believe it can do it, even in theory, let alone in practice. Indeed AFAIK there is no published way to efficiently solve any known NP-complete problems using quantum computers.
There is a well-known quantum algorithm to efficently solve factorisation, which is definitely NP but probably not NP-complete. This QC is not designed to solve factorisation. I don't know what it will do, we'll have to wait.
Maybe this QC can solve small quadratic integer programming problems. This would be interesting and groundbreaking but not enough to really change computing as we know it.
No there isn't. There is an artificial restriction of people who could go on to become doctors, but it doesn't follow that these people would necessarily be any good.
Yeah, I remember in the late 70s Microsoft used to do a pretty good BASIC. Ever since they came up with this DOS thing, I reckon it's been going downward.
First, it is perfectly possible -- and, in fact, dead easy -- to write software today on a beta of 10.5 that will run on every OS back to 10.0 on every machine capable of running 10.0. That's coming up on 8 years of hardware and 6 years of software that said app will run on. Applications on the Mac do not "require new frameworks" or any such thing. There is a robust groundwork of commonality in all versions of OS X that make it possible to build even the most advanced applications in a backwards compatible way.
Yes it is possible, if you limit yourself to pretty much the old NeXT toolkit of 1994 or so, or alternatively if you use Carbon, or perhaps if you give up the fancy aqua GUI and move to X11, but it is not dead easy, far less so since the move to Intel, as Carbon is not compatible with Universal binary, AFAIK. Remember it was a huge issue for Adobe, who was still using Carbon for its Creative Suite, and the main reason for not having a CS3 yet.
In fact it is so not dead easy that very few OS/X developers maintain software that will run on 10.0.x ; I don't know of any except some F/OSS folk. Apple itself doesn't bother, not to mention any of the big development houses.
Hello, for a start, ask around people who are not young adults with few brushes with serious illness if they think healh care is a luxury.
Second, this makes no sense:
Rights should be limited to those things which can not be taken away from me, not those things which must be given to me.
Everything can be taken away from you, including life and liberty. Are these not rights ?
This also does not compute:
But please do explain to me how something others provide to you can be your right?
Do you consider voting a right ? Pretty much everything you have is provided to you by society. Precious democracy is still a novel thing not available everywhere on the planet, yet that you may enjoy (I don't know your particulars). Do you consider that you provide yourself unaided the right to vote ? I think not. What about the rest of what you consider your rights ?
If anything, your health is something that is an essential part of you. It should be your right, and in many countries it is, for it to be maintained to an acceptable degree by society. Sure a doctor must care for you, but in what way is it different than a politician or a journalist fighting for your right to freedom of speech ?
Personal health insurance is not enough. I personnally know people who contacted serious enough, however curable a disease that their insurance refused to pay for under some weird excuse. They had to fight this throught the courts. How can that be an effective way to run a society?
You don't trust your current government to run health care, fine. Why don't you use your society-maintained voting rights to put someone in place you think you will trust for this task, assuming you find it worthwhile ?
You mean, in the USA ? Why not ? In many countries around the world, it is. Unsurprisingly, perhaps, people live measurably longer in these countries.
Among countries that have a longer average lifespan than the ol' US of A, you'll find some Asian countries where the custom is to pay your doctor only if you *don't* get sick. If you do, they look after you for free. Interesting concept, no ?
It is not known whether Quantum computers can solve NP-complete problems any better than normal computers. Quantum computers can solve some NP problems like the factorisation problem (using Shor's algorithm) efficiently, however these are not believed to be NP-complete problems.
By itself a OLPC instance is useless. What you describe assumes a whole infrastructure including a network connected to the Internet. That is assuming an enormous amount : very qualified people will have to maintain it and work on it daily. This is not going to happen.
In the poorest countries in Africa, something as complicated as a well quickly becomes unusable because of lack of maintenance. In past decades people in such countries knew how to dig and maintain wells, but due to AIDS, famines, warfare, rural exodus and general poverty so much of the knowledge held by trained adult just vanished. Also wells have to be dug deeper these days due to desertification.
Wells are often now dug by first world volunteers, which is all good, but these people require first world technology meaning pumps and engines that require petrol. Eventually the volunteers leave and the local people are left to fend for themselves with technology they know very little about. The volunteers often train some local people to look after the pumps, but these people with such skills can find employment elsewhere and leave. Experience shows that the wells quickly becomes unusable.
OLPC is nice but reflects the huge gap of understanding required to lift most of Africa out of its quagmire. In fact its a toy for 1st world technologists.
States have expenses too. Maybe not armies, but roads, schools, employees and so forth. Some of these expenses hopefully benefit the public. They have to be paid by taxes, and if Google doesn't pay these 3 millions a year, rest assured that someone else will, most probably taxpayers in one form or another.
Because most people cannot hear the artifacts caused by AAC 128kbit compression, at least not without direct comparison with the original uncompressed file. However, combined mp3 and AAC compression artifacts might be more readily audible.
Note that used CD can feature gaps, cracks and even unreadable tracks. No need to be an audiophile to notice these.
Finally even CD data is compressed (PCM, 16-bit, 22.5kHz sampling). It's all a matter of how much compression artifacts you can live with.
The point has been made several times that the resulting mp3 are of lesser quality than the original DRM-encumbered AAC files, and that this action requires time.
Check the Apple license agreement to Darwin, the APSL. I'm pretty sure that for any code to be checked back in to Apple's Darwin repository, copyright must be assigned to Apple.
Note that the FSF does the same thing with its various official GNU projects. If you want to contribute code to, say, GCC, you must give up your copyright to the FSF, together with a signed sheet of paper that says you do in fact own the copyright of the stuff you are contributing. This is non-negociable and meant to avoid SCO-like lawsuits associated to GNU projects. So far it has worked rather well from that point of view.
No real advantage besides using the same measurement units than all the other countries in the world, including Britain. Regarding Phoenix, I would have thought that either north-south avenues would visibly diverge or that these 1sq-mile square aren't really perfect. I mean if avenues run perfectly north-south then they must converge as they go towards the north pole. Therefore a square north of the city must be slightly shorter on its side than a corresponding square further south, right ?
Given the time, it was not so bad a result, although it could have been better because the lawmakers used provisional estimations from the measuring expedition because they were in a hurry for a result. Some things never change.
eMusic has heaps of jazz and the classical section is quite good. It's not a replacement for iTunes, but I now spend more money at eMusic than iTunes.
It's interesting to see that in spite of its somewhat meagre catalogue, unwieldy web site and dearth of poular artists, eMusic manages to cinch the #2 spot.
Slashdot Mirror
There's no telling what this damn bird flu could do. Big deal if it chokes the net for a few weeks even and annoys a few slashdotters.
Rightly or wrongly, this idea would mean the end of mankind.
Interested readers can browse the content of PAMI current and back issues and either go to their local scientific library (PAMI is recognisable from afar by its bright yellow cover) or search on the web for interesting articles. Often researchers put their own paper on their home page. For example, here is the publication page of one of the authors (I'm not him).
For the record, I think justifying various ad-hoc vision/image analysis techniques using approximations of biological underpining is of limited interest. When asked if computer would think one day, Edsgerd Dijkstra famously answered by "can submarine swim?". In the same manner, it has been observed that (for example) most neural network architectures make worse classifiers than standard logistic regression, not to mention Support Vector Machines, which what this article uses BTW.
The summary by our friend Roland P. is not very good
I could go on with lists and links but the future is already here, generally inconspicuously. Read about it.
Sorry, no, I wasn't clear.
This is not a general-purpose quantum computer, it is supposedly designed to solve quadratic integer programming (efficiently), which is NP-complete, but I don't believe it can do it, even in theory, let alone in practice. Indeed AFAIK there is no published way to efficiently solve any known NP-complete problems using quantum computers.
There is a well-known quantum algorithm to efficently solve factorisation, which is definitely NP but probably not NP-complete. This QC is not designed to solve factorisation. I don't know what it will do, we'll have to wait.
Maybe this QC can solve small quadratic integer programming problems. This would be interesting and groundbreaking but not enough to really change computing as we know it.
Since we don't know whether P = NP, both are correct.
So what you are writing is that capitalism is the absolute worst economic system, except for all the others ?
No there isn't. There is an artificial restriction of people who could go on to become doctors, but it doesn't follow that these people would necessarily be any good.
Yeah, I remember in the late 70s Microsoft used to do a pretty good BASIC. Ever since they came up with this DOS thing, I reckon it's been going downward.
Yes it is possible, if you limit yourself to pretty much the old NeXT toolkit of 1994 or so, or alternatively if you use Carbon, or perhaps if you give up the fancy aqua GUI and move to X11, but it is not dead easy, far less so since the move to Intel, as Carbon is not compatible with Universal binary, AFAIK. Remember it was a huge issue for Adobe, who was still using Carbon for its Creative Suite, and the main reason for not having a CS3 yet.
In fact it is so not dead easy that very few OS/X developers maintain software that will run on 10.0.x ; I don't know of any except some F/OSS folk. Apple itself doesn't bother, not to mention any of the big development houses.
That would work if there were an oversupply of (good) doctors, which is not the case. Good medical talent is hard to come by, like any true talent.
Second, this makes no sense
Everything can be taken away from you, including life and liberty. Are these not rights ?
This also does not compute
Do you consider voting a right ? Pretty much everything you have is provided to you by society. Precious democracy is still a novel thing not available everywhere on the planet, yet that you may enjoy (I don't know your particulars). Do you consider that you provide yourself unaided the right to vote ? I think not. What about the rest of what you consider your rights ?
If anything, your health is something that is an essential part of you. It should be your right, and in many countries it is, for it to be maintained to an acceptable degree by society. Sure a doctor must care for you, but in what way is it different than a politician or a journalist fighting for your right to freedom of speech ?
Personal health insurance is not enough. I personnally know people who contacted serious enough, however curable a disease that their insurance refused to pay for under some weird excuse. They had to fight this throught the courts. How can that be an effective way to run a society?
You don't trust your current government to run health care, fine. Why don't you use your society-maintained voting rights to put someone in place you think you will trust for this task, assuming you find it worthwhile ?
You mean, in the USA ? Why not ? In many countries around the world, it is. Unsurprisingly, perhaps, people live measurably longer in these countries.
Among countries that have a longer average lifespan than the ol' US of A, you'll find some Asian countries where the custom is to pay your doctor only if you *don't* get sick. If you do, they look after you for free. Interesting concept, no ?
Correction, factoring is not known to be NP-complete.
It is not known whether Quantum computers can solve NP-complete problems any better than normal computers. Quantum computers can solve some NP problems like the factorisation problem (using Shor's algorithm) efficiently, however these are not believed to be NP-complete problems.
Thanks for the reply, but you assume too much. There is no need to interpret what I wrote.
By itself a OLPC instance is useless. What you describe assumes a whole infrastructure including a network connected to the Internet. That is assuming an enormous amount : very qualified people will have to maintain it and work on it daily. This is not going to happen.
In the poorest countries in Africa, something as complicated as a well quickly becomes unusable because of lack of maintenance. In past decades people in such countries knew how to dig and maintain wells, but due to AIDS, famines, warfare, rural exodus and general poverty so much of the knowledge held by trained adult just vanished. Also wells have to be dug deeper these days due to desertification.
Wells are often now dug by first world volunteers, which is all good, but these people require first world technology meaning pumps and engines that require petrol. Eventually the volunteers leave and the local people are left to fend for themselves with technology they know very little about. The volunteers often train some local people to look after the pumps, but these people with such skills can find employment elsewhere and leave. Experience shows that the wells quickly becomes unusable.
OLPC is nice but reflects the huge gap of understanding required to lift most of Africa out of its quagmire. In fact its a toy for 1st world technologists.
States have expenses too. Maybe not armies, but roads, schools, employees and so forth. Some of these expenses hopefully benefit the public. They have to be paid by taxes, and if Google doesn't pay these 3 millions a year, rest assured that someone else will, most probably taxpayers in one form or another.
...these things are. Keep getting bullied by the BSA ? ditch them, you don't need them. Did I mention non-BSA software is sometime free, even Libre ?
Because most people cannot hear the artifacts caused by AAC 128kbit compression, at least not without direct comparison with the original uncompressed file. However, combined mp3 and AAC compression artifacts might be more readily audible.
Note that used CD can feature gaps, cracks and even unreadable tracks. No need to be an audiophile to notice these.
Finally even CD data is compressed (PCM, 16-bit, 22.5kHz sampling). It's all a matter of how much compression artifacts you can live with.
They can also fine them.
The point has been made several times that the resulting mp3 are of lesser quality than the original DRM-encumbered AAC files, and that this action requires time.
Check the Apple license agreement to Darwin, the APSL. I'm pretty sure that for any code to be checked back in to Apple's Darwin repository, copyright must be assigned to Apple.
Note that the FSF does the same thing with its various official GNU projects. If you want to contribute code to, say, GCC, you must give up your copyright to the FSF, together with a signed sheet of paper that says you do in fact own the copyright of the stuff you are contributing. This is non-negociable and meant to avoid SCO-like lawsuits associated to GNU projects. So far it has worked rather well from that point of view.
No real advantage besides using the same measurement units than all the other countries in the world, including Britain. Regarding Phoenix, I would have thought that either north-south avenues would visibly diverge or that these 1sq-mile square aren't really perfect. I mean if avenues run perfectly north-south then they must converge as they go towards the north pole. Therefore a square north of the city must be slightly shorter on its side than a corresponding square further south, right ?
Actually they made a 0.02% error (0.2mm in 1m). See also the obligatory wikipedia link.
Given the time, it was not so bad a result, although it could have been better because the lawmakers used provisional estimations from the measuring expedition because they were in a hurry for a result. Some things never change.
eMusic has heaps of jazz and the classical section is quite good. It's not a replacement for iTunes, but I now spend more money at eMusic than iTunes.
It's interesting to see that in spite of its somewhat meagre catalogue, unwieldy web site and dearth of poular artists, eMusic manages to cinch the #2 spot.