...is always released to the press before being published in a peer reviewed journal.:P
"Intelligence tests are hocus-pocus used by psychologists to prove that they are smart and that their clients are stupid. The general acceptance of these tests suggests that this claim may not be without foundation." --Thomas Szasz
An "infinite number" of mirrors can't exist, so not being able to determine what they would do is not a failing of physics; it's a failing of your thought experiment.
In a more abstract statement of the problem, given an infinite series:
1 + 1/2 + 1/3 + 1/4 + 1/5 +...
is the denominator of the last term even or odd?
Well, if it's an infinite series, there is no "last term" and thus the question is ill posed. The whole point of an infinite series is that it doesn't end (that's where the "infinite" part comes from).
Backpropagation is a supervised learning method. That means you have to tell it "right" or "wrong, it was this..." in order to get a backprop net to learn.
This technique is unsupervised, which means it figures out different categories (the number depends on learning parameters and the data presented) and clusters the inputs into one or more of these clusters. Example clusters would be "animal-like" or "furry".
The interesting thing about this technique seems to be the combination of these clusters into larger clusters. In other words, it doesn't just learn the categories, it learns the components that make up the categories too.
Your photoreceptors in your retina absorb and respond better to particular wavelengths of photons. These are the same in everybody because of the chemical nature of the photopigments. Because of this, different people's eyes respond similarly to every stimulus.
I'm sure you could really mess up children though, teaching them the wrong color names for different objects. They'd be real confused when they came into contact with children taught the traditional color names. I can almost hear the arguments now:
The problems people have had with first person shooters is that it desensitizes people from violence. In a flight sim, you shoot down and blow up objects. You are not getting desensitized from killing living creatures but just destroying objects.
Sure, those planes are just equipment.And their pilots aren't people, they're the ENEMY!
Slavery is legal in the U.S. It says so in the Constitution! Of course, you have to be convicted of a felony and imprisoned through due process first. Luckily, we're making that easier and easier, and the prison population is skyrocketing. Of course, we can't let all that potential labor go to waste! That would be bad for business, and therefore bad for America.
Prisoner labor is big business here in the U.S. We don't even have to go to neighboring countries to find where we are screwing people over.
One of the posts above put it quite nicely:
"I guess people all over the world are the same inside, no matter what their governments say."
This may sound like just a bunch of preachy BS, but it's very disturbing...
No, it's just preachy BS.
...and the creator of the first one that says "I'm Sorry Dave" will think he is god.
It's obvious you are just tripping. The creator of the first one that says "I'm Sorry Dave" will think he is Dave, or that he has been mistaken for someone who IS Dave.
...people were glorifying themselfs, when they were really just mimmicking what they already saw with the human brain.
As far as "glorifying themselfs" for mimicking what they already saw with the human brain, WTF are you talking about? Of course they're proud of themselves! They did something with a computer that nobody had ever been able to do before. That right there is fucking cool! I'm proud of them too, and I go to UCLA! (the researchers in question were from USC, in case anybody missed that)
It doesn't say anything at all about language understanding. It says the network "recognizes words" in noisy conditions better than humans. That is a far cry from understanding language!
The most interesting thing in this is that it used so few neurons yet it can recognize speech patterns and categorize them! This technique has applications ranging far outside of just speech recognition!
[...] either have someone with the device planted on them track the suspect or have an array of said devices placed in public areas where the suspect is known to hang out (bus terminals, bars, etc.). Sooner or later, the suspect might slip up and the authorities have enough evidence needed for an arrest.
My god, you're right! They could never do this with a tape recorder or a digital sampler because it would never recognize the person's voice! It would just record everything!
Oh wait. That would work just fine.:P
Someone would just have to listen to the tape. Maybe one of these things could be used to scan the tape though...
>> When it comes down to it, a NN is just a >> non-linear function.
> Specific nonlinear and linear model structures > can be used to do the exact same things neural nets can do IF YOU KNOW THE CORRECT STRUCTURE. > This is where the power of neural networks lies. > A neural network can (I think) match arbitrary > nonlinearity, given enough hidden nodes.
Yes, it can "match" it if you are lucky with the training. The problem is that there is no guarantee that this will occur.
Also, it IS possible to extract a nonlinear function from a trained neural network. It just isn't pretty (or easy).
You're right, the journalists missed the point. The part that's "wrong" is the over-application of the gaussian distribusion as a model of everything .
You just don't get it! We don't want any damn assistants! Ever! We'll do it ourselves and to hell with the planet or the cat or whoever the fuck M$ representative comes bundled with our bloatware!
The key to the illusion of making a computer look like it has more cycles is to pre-render as much as possible. This could work with mode probabilities. If a player is exceptionally good at taking out a behemoth tank with mines and finishing it off with an air strike, for example, the ai probability could opt to go into "defense from behemoth tank with mines and air strike" mode. I have not seen this variant on modes in games yet.
The reason you have not seen this variant on modes is that it is insanely hard to enumerate all the possible strategies and defences. The real key will be when someone develops a method whereby the AI can learn new strategies and plug them into this architecture.
The only thing that really bugged me about this book was the Gatesian fear of new techniques to "factor large prime numbers" (p.55)
I'd like to take this opportunity to make the claim that I can instantly factor ANY prime number NO MATTER HOW BIG IT IS! The problem with deciphering these coded messages is not factoring large prime numbers, but rather in factoring composite numbers which are multiples of large prime numbers.
Overlooking that small point, the rest of the book is entertaining but not addicting. That could be a good thing or not, depending on your pov.
This is right down the alley with those "facts" about how the average human uses 5-10% of the capability of his brain. Where does that measurement come from?
I always thought it came from the fact that out of every 10 cells in the brain, about 9 are glial cells and one is a neuron. Since glial cells don't seem to do much communication, we're only using 10% of the cells in our brain for processing.
> what is wrong with a company that says "you can work with our code and if you make money we can have some"?
Here's one problem: how much $$ do they get? Will it depend on how much of their code is copy/pasted into your final product? (how do they know?) What if you went through and changed all the variable names? What if you did minor modifications to the program structure? What about major modifications? What if you just got some good IDEAS from their code? In many of these cases, it's tough to see how they're going to prove that you used their code if you say you didn't.
Now suppose you're going to be generous (and honest) and let everyone know that your product is derived in part from their code. Now they want money from you. I can see how this could work to the benefit of both you and the code-providers, but I can also see how it could provide ammunition to legal departments! Read the licensing agreements *very* *very* *carefully*.
Slashdot Mirror
Maybe beautiful women just make the stupid psychological tests seem less important by comparison.
...is always released to the press before being published in a peer reviewed journal. :P
"Intelligence tests are hocus-pocus used by psychologists to prove that they are smart and that their clients are stupid. The general acceptance of these tests suggests that this claim may not be without foundation." --Thomas Szasz
In a more abstract statement of the problem, given an infinite series:
is the denominator of the last term even or odd?
Well, if it's an infinite series, there is no "last term" and thus the question is ill posed. The whole point of an infinite series is that it doesn't end (that's where the "infinite" part comes from).
I believe these kids are guilty of hate crimes against stupid people.
For any prime p, its factors are p and 1.
If we allow negatives, then -p and -1 work too.
How fast is that?
This technique is unsupervised, which means it figures out different categories (the number depends on learning parameters and the data presented) and clusters the inputs into one or more of these clusters. Example clusters would be "animal-like" or "furry".
The interesting thing about this technique seems to be the combination of these clusters into larger clusters. In other words, it doesn't just learn the categories, it learns the components that make up the categories too.
I'm sure you could really mess up children though, teaching them the wrong color names for different objects. They'd be real confused when they came into contact with children taught the traditional color names. I can almost hear the arguments now:
"Nun uuuuh! It's not red, it's green!"
I have two of those. I call them 'eyes'
Give me a break.
Slavery is legal in the U.S. It says so in the Constitution! Of course, you have to be convicted of a felony and imprisoned through due process first. Luckily, we're making that easier and easier, and the prison population is skyrocketing. Of course, we can't let all that potential labor go to waste! That would be bad for business, and therefore bad for America.
Prisoner labor is big business here in the U.S. We don't even have to go to neighboring countries to find where we are screwing people over.
One of the posts above put it quite nicely:
No, it's just preachy BS.
It's obvious you are just tripping. The creator of the first one that says "I'm Sorry Dave" will think he is Dave, or that he has been mistaken for someone who IS Dave.
As far as "glorifying themselfs" for mimicking what they already saw with the human brain, WTF are you talking about? Of course they're proud of themselves! They did something with a computer that nobody had ever been able to do before. That right there is fucking cool! I'm proud of them too, and I go to UCLA! (the researchers in question were from USC, in case anybody missed that)
The most interesting thing in this is that it used so few neurons yet it can recognize speech patterns and categorize them! This technique has applications ranging far outside of just speech recognition!
My god, you're right! They could never do this with a tape recorder or a digital sampler because it would never recognize the person's voice! It would just record everything!
Oh wait. That would work just fine. :P
Someone would just have to listen to the tape. Maybe one of these things could be used to scan the tape though...
>> non-linear function.
> Specific nonlinear and linear model structures
> can be used to do the exact same things neural nets can do IF YOU KNOW THE CORRECT STRUCTURE.
> This is where the power of neural networks lies.
> A neural network can (I think) match arbitrary
> nonlinearity, given enough hidden nodes.
Yes, it can "match" it if you are lucky with the training. The problem is that there is no guarantee that this will occur.
Also, it IS possible to extract a nonlinear function from a trained neural network. It just isn't pretty (or easy).
You're right, the journalists missed the point. The part that's "wrong" is the over-application of the gaussian distribusion as a model of everything .
You just don't get it! We don't want any damn assistants! Ever! We'll do it ourselves and to hell with the planet or the cat or whoever the fuck M$ representative comes bundled with our bloatware!
I'd like to take this opportunity to make the claim that I can instantly factor ANY prime number NO MATTER HOW BIG IT IS! The problem with deciphering these coded messages is not factoring large prime numbers, but rather in factoring composite numbers which are multiples of large prime numbers.
Overlooking that small point, the rest of the book is entertaining but not addicting. That could be a good thing or not, depending on your pov.
I always thought it came from the fact that out of every 10 cells in the brain, about 9 are glial cells and one is a neuron. Since glial cells don't seem to do much communication, we're only using 10% of the cells in our brain for processing.
Here's one problem: how much $$ do they get? Will it depend on how much of their code is copy/pasted into your final product? (how do they know?) What if you went through and changed all the variable names? What if you did minor modifications to the program structure? What about major modifications? What if you just got some good IDEAS from their code? In many of these cases, it's tough to see how they're going to prove that you used their code if you say you didn't.
Now suppose you're going to be generous (and honest) and let everyone know that your product is derived in part from their code. Now they want money from you. I can see how this could work to the benefit of both you and the code-providers, but I can also see how it could provide ammunition to legal departments! Read the licensing agreements *very* *very* *carefully*.