It didn't learn how to play Go at all. The rules were programmed into it. If that hadn't happened, the computer wouldn't have played a single move.
Roughly speaking, what it was programmed to do, was "learn".
Very roughly. That's how we explain it to people who are more comfortable with anthropomorphization than with the actual workings of neural networks. It is more appropriately described as a curve fitting problem, with a heuristic in place because the domain is too large to actually fit in computers.
but the extent of human intervention in "learn" is decreasing considerably, and fairly soon "learn" will be completely general.
This is the sort of comment that makes me think you've never written a neural network, and actually don't understand them at all. There is soooooo much human intervention involved. More importantly there is no clear path from neural networks like we have now to general AI. Whoever gave you that impression was dead wrong.
but being the first nation in 50 years to use them in warfare will bring a new kind of political hell that you really don't want to defend against.
So that's an interesting question.......what do you think would happen if a country used nukes? It would probably depend on which country used them........
That is irrelevant. You entirely missed the point. It did exactly what it was programmed to do, and it is easy to distinguish between AlphaGo and a human. Even you can distinguish.
This is an old topic, and it's been hashed out philosophically. It's only confusing to people who don't know the definition of strong AI. When you have strong AI, it'll be able to play Go, then in the afternoon converse on the finer points of Slashdot moderation without needing to be reprogrammed. Or even programmed to do those things at all: it will be able to teach itself to play parcheesi or anything else not programmed into it in advance.
In other words, you are looking at it wrong: Strong AI is not a skill level, it's human cognitive equivalence. No one will be able to tell if it's a human or a Strong AI in the Chinese room.
The future of the U.S. Air Force may well depend on advanced platforms like F-35s commanding fleets of unmanned drones which can act as additional ears, eyes, and shooters in the sky during battles."
That works great until there is a jammer. In other words, it works fine against small, overpowered nations against whom there are already a myriad of options.
If a player plays pure, cold, calculation, then the way to counter that is to put place a big bet on the flop. If the computer has a mediocre hand, it will not calculate that the hand is worth playing and will fold. Then you keep doing that, the computer keeps folding, and you win.
Now consider that the computer realizes your strategy, and starts calling your bluffs. That's when you have a game.
"People think that bluffing is very human," Mr Brown told Bloomberg, "It turns out that's not true."
"A computer can learn from experience that if it has a weak hand and it bluffs, it can make more money."
Nah. They had a good market position thanks to IBM (and they also had 'sharp' business practices, and other times completely unethical business practices), but Apple and Commodore made strategic mistakes, too.
That hosts file routes everything to 127.0.0.1, which can be kind of inconvenient if you have a web server running on port 80 on your local computer. Better to route it to 0.0.0.0
I think there's a little non-sequitur there, inasmuch as you need to acquire a monopoly on a market before you can abuse it. Certainly Microsoft did without a doubt abuse their monopoly once they had it, but it's not how they grew their monopoly in the first place.
The headline is really wrong, they didn't detect vulnerability, they detected volatility. Some months income was above average, some months it was below average. Salesmen are really familiar with this, but that's just how some jobs are.
In some American dialects (specifically, mine, whatever it is), college and university are the same thing.
It's really not worth arguing about though, unless we can quantify exactly how much lower your taxes would be. We would also have to quantify how much college would cost, because that can vary dramatically.
My parents didn't pay for my college though, and I imagine that's true of a lot of kids, otherwise they wouldn't be complaining about student loans.
Wow, learn to read, I'll bold the relevant part of the original quote for you:
TBH that's probably not worth it. Hard to be sure because obviously the 40% + 25% goes to other things as well. You've probably spent a lot extra in taxes than you would have if you'd just paid for college.
Slashdot Mirror
Astrologers are kicking ass this month!
Aries, the Ram, is quite good at that.
You're probably right, but I would have responded by linking it to a Big Mac
Roughly speaking, what it was programmed to do, was "learn".
Very roughly. That's how we explain it to people who are more comfortable with anthropomorphization than with the actual workings of neural networks. It is more appropriately described as a curve fitting problem, with a heuristic in place because the domain is too large to actually fit in computers.
but the extent of human intervention in "learn" is decreasing considerably, and fairly soon "learn" will be completely general.
This is the sort of comment that makes me think you've never written a neural network, and actually don't understand them at all. There is soooooo much human intervention involved. More importantly there is no clear path from neural networks like we have now to general AI. Whoever gave you that impression was dead wrong.
but being the first nation in 50 years to use them in warfare will bring a new kind of political hell that you really don't want to defend against.
So that's an interesting question.......what do you think would happen if a country used nukes? It would probably depend on which country used them........
That is irrelevant. You entirely missed the point. It did exactly what it was programmed to do, and it is easy to distinguish between AlphaGo and a human. Even you can distinguish.
This is an old topic, and it's been hashed out philosophically. It's only confusing to people who don't know the definition of strong AI. When you have strong AI, it'll be able to play Go, then in the afternoon converse on the finer points of Slashdot moderation without needing to be reprogrammed. Or even programmed to do those things at all: it will be able to teach itself to play parcheesi or anything else not programmed into it in advance.
In other words, you are looking at it wrong: Strong AI is not a skill level, it's human cognitive equivalence. No one will be able to tell if it's a human or a Strong AI in the Chinese room.
The future of the U.S. Air Force may well depend on advanced platforms like F-35s commanding fleets of unmanned drones which can act as additional ears, eyes, and shooters in the sky during battles."
That works great until there is a jammer. In other words, it works fine against small, overpowered nations against whom there are already a myriad of options.
If you have trouble because your emotions are clouding your thoughts, the solution is to do some zen meditation or similar.
Computers definitely respond to bluffing. When someone puts down a large bet, they have to decide whether to raise or fold, just like the rest of us.
If a player plays pure, cold, calculation, then the way to counter that is to put place a big bet on the flop. If the computer has a mediocre hand, it will not calculate that the hand is worth playing and will fold. Then you keep doing that, the computer keeps folding, and you win.
Now consider that the computer realizes your strategy, and starts calling your bluffs. That's when you have a game.
"People think that bluffing is very human," Mr Brown told Bloomberg, "It turns out that's not true." "A computer can learn from experience that if it has a weak hand and it bluffs, it can make more money."
I don't know that this is the official moment when AI becomes smarter than us,
It's not, this is just weak AI. Unless you also think a calculator is smarter than us, which, from a certain perspective, it is.
Nah. They had a good market position thanks to IBM (and they also had 'sharp' business practices, and other times completely unethical business practices), but Apple and Commodore made strategic mistakes, too.
Uh, did you read the part in bold?
They're working on that, too. Firefox is fast becoming the best browser.
That hosts file routes everything to 127.0.0.1, which can be kind of inconvenient if you have a web server running on port 80 on your local computer. Better to route it to 0.0.0.0
I think there's a little non-sequitur there, inasmuch as you need to acquire a monopoly on a market before you can abuse it. Certainly Microsoft did without a doubt abuse their monopoly once they had it, but it's not how they grew their monopoly in the first place.
The headline is really wrong, they didn't detect vulnerability, they detected volatility. Some months income was above average, some months it was below average. Salesmen are really familiar with this, but that's just how some jobs are.
Average gasoline prices is currently $0.68 / gallon in the USA.
That would be great but it's actually closer to $3 a gallon. The rest of your comment is on point, though.
Which route only has United?
Uber knows they are screwed in this situation, so almost anything they can change gives them a better chance.
Google's LIDAR designs are all patented, so they are public anyway. They don't need to hide it in a trial.
Not sure if you meant college or university.
In some American dialects (specifically, mine, whatever it is), college and university are the same thing.
It's really not worth arguing about though, unless we can quantify exactly how much lower your taxes would be. We would also have to quantify how much college would cost, because that can vary dramatically.
My parents didn't pay for my college though, and I imagine that's true of a lot of kids, otherwise they wouldn't be complaining about student loans.
It would be a stretch to say that self-driving cars exists today,
That's kind of true, too. We're definitely not at SAE level 4 autonomous cars.
That's because the program is rewriting itself while it runs.
No, that's a common misconception. Once it's trained, it doesn't. It wasn't re-training itself while playing against Lee Sedol.
TBH that's probably not worth it. Hard to be sure because obviously the 40% + 25% goes to other things as well. You've probably spent a lot extra in taxes than you would have if you'd just paid for college.