It's not that hard. A Hubble-sized space telescope could see a kilowatt laser over 4 light years. The hard part is sending something big enough to generate that kind of power 4 light years.
A typical nuclear reactor puts out around 1 GW. So you need 65 of them (or one with 65 times the capacity). Not trivial, but not exactly requiring engineering breakthroughs either.
If you have 20 models and each one measures an effect with a confidence of, say, 0.1 sigma, the ensemble is MORE than 0.1 sigma. Only if you've got some showing an effect one way and other showing an effect the other way does the confidence go down.
In fact, this very article is about such a situation. IIRC the results from the LHC and SLAC individually are about 3 sigma, but both in the same direction, so when you combine them you get 4.
Climate models all generally agree, at least in the direction of the changes. That means the total confidence is greater than the confidence from any single model.
Given no prior knowledge, 2 sigmas is roughly 95% confidence that the observation is not a false positive. Given the risk of spending trillions to avert a possible false positive or not doing anything and having billions displaced and the world's food production severely damaged in the case of a false negative, doing something seems like a reasonable course no?
It's academic anyway. There's no way the world is politically going to do much. Fortunately it looks like the economic choice will be non-fossil in the next few years anyway. Perhaps this is the solution to the Fermi paradox: you only avoid cooking yourselves back to the stone age (with no easily exploitable oil for next time) if you mange to invent economically superior solar panels in time.
No, Physics assumes that particles behave in consistent ways, at least statistically. Simplified, the observation here seems to be that tau particles decay a little faster than we predict they should. There's some uncertainty, because there are errors in our measurements, and the decay time itself is stochastic. As you collect more data you can put narrower limits on the average decay time and become more confident that it is really different from prediction.
Two sigma is the usual standard. The magic p 0.05.
Some fields use higher standards for very high profile findings, either directly or indirectly. The FDA for example, usually requires at least two independent trials with p 0.05 to approve a new drug. In high energy particle physics it is possible to do very tightly controlled experiments that are expensive to set up but then relatively cheap to run. With this situation, getting more data is fairly inexpensive, so you can achieve very high confidence on your results just by waiting a little longer.
You should never accept anything as "true." In science we accept things as "most likely" and "meeting our prespecified level of evidence."
For observational sciences like parts of astronomy and climate, you might just have to live with what you get. For most of medicine and some of climate, we could certainly achieve five sigma confidence. It would cost orders of magnitude more than the two or three sigma we generally use now.
Both of those. Do you not follow the news? There are self-driven cars driving around all over the place. Any number of robots could cross a road as well, such as the combat robots Google makes.
I think your example reinforces his point. Q-learning, or more generally reinforcement learning, is a learning algorithm. You don't program the system, you set up some basic infrastructure and then train it by example. We've learned that such systems can learn to do things we don't understand, and cannot program.
They were. Go is quite resistant to the brute force and play dictionary techniques used in the past on checkers and chess, which is why people wax poetic about the complexity of Go.
AlphaGo is trained using reinforcement learning, which, frankly, is such a twitchy thing that it's still surprising how well it can work.
Kasparov was beaten by a big computer programmed to play chess. AlphaGo is a very different thing.
Air isn't much of a problem. As Slashdot pundits are fond of pointing out, the entire planet is covered with perchlorates, which are an excellent source of oxygen. Curiosity has found nitrogen too.
I don't know... an editor that screws around with your text without telling you, and doesn't have an option to turn it off? That sounds like a good problem to fix with high priority to me.
I suppose it might have been a good idea fifteen years ago. Possibly even ten, but they would have had to have been fast. But by five years ago there were already many smartphone apps that offered to do the same thing... for free.
Every IDE contains an editor. I find such posts as yours simply idiotic.
Hey old friend! Still don't read posts hey? If english is not your first language it's understandable, but you really shouldn't criticize then. Let me rewrite my post for you, slowly.
OP said "Maybe I'm using the wrong editor...."
I said (edited for the angels): "You're using the wrong editor. [It's probably part of] an IDE? IDEs are evil."
By the way, I find most of your posts pointless, and the others demonstrably wrong. I *think* you managed to score both on this one:
Whitespace IS visible, and the interpreter will happily point out where you (or your editor) screwed up. In Python? No. Either the code can be interpreted then it runs, but gives the wrong result... or it is a recognized error, and no one is complaining about the later. All hate the former.
Python gives an IndentationError exception when you have whitespace the interpreter can't parse. It points out the exact location. So if you slip in a space by accident (because you should be using tabs) then the interpreter helpfully points out exactly where it is so you can delete it.
You're using the wrong editor. Probably an IDE? IDEs are evil.
Personally, I find braces the bad option. Whitespace IS visible, and the interpreter will happily point out where you (or your editor) screwed up. Braces are hard to keep track of, particularly when heavily nested. Of course, if you use whitespace appropriately the braces are much easier to track....
Re:No, because meaningful whitespace (Score:?) by Anonymous Coward on Monday June 12, 2017 @09:16AM I consider that a feature actually. Use tabs for code. When you copy and paste something from the web you have to go through line by line and convert the spaces to tabs. That forces you to go through that code you just downloaded off the web line by line....
Actually, you can almost always do a search and replace for multiple spaces, but don't tell that to my summer students.
Edit: it appears more recent evidence makes the Sahara older, although there do seem to have been wet periods in the meantime. Even so, pre-human hominids clearly made it between Ethiopia and Morocco.
The Sahara didn't exist 300,000 years ago. Instead, it seems it was a fairly lush grassland. You probably wouldn't have one dude walking between the two fathering children as he went, but it's not an unreasonable geography for genes to cross back and forth, over a thousand generations.
Clearly pre-technology hominids COULD cross the barrier because their bones have been found in both places (as have apes, other primates, and lots of other animals). The question is whether their genes could slosh around enough to keep the whole area from going off on individual evolutionary tangents. It's not unreasonable that happened.
He may well have meant there is no hope of ever observing the phenomenon directly. In 1920 there was considerable argument about whether certain "nebula" were independent galaxies or just parts of our own, and a comprehensive star survey was being prepared... containing a quarter of a million stars.
Without modern telescope technology, CCD cameras, computers, and a knowledge of just how many stars there are, it would be nearly impossible to detect that kind of effect in reasonable timeframes. Also, in 1920 only three white dwarfs were known, and nobody knew that they were compact objects.
Future knowledge of all those things would have been hard to anticipate in 1920.
Slashdot Mirror
It's not that hard. A Hubble-sized space telescope could see a kilowatt laser over 4 light years. The hard part is sending something big enough to generate that kind of power 4 light years.
You be missin' some zeros. The world generates about 24,000 TWh/y currently:
https://en.wikipedia.org/wiki/...
A typical nuclear reactor puts out around 1 GW. So you need 65 of them (or one with 65 times the capacity). Not trivial, but not exactly requiring engineering breakthroughs either.
You keep mentioning VR and 3d TVs. Are you familiar with the red herring fallacy?
If you have 20 models and each one measures an effect with a confidence of, say, 0.1 sigma, the ensemble is MORE than 0.1 sigma. Only if you've got some showing an effect one way and other showing an effect the other way does the confidence go down.
In fact, this very article is about such a situation. IIRC the results from the LHC and SLAC individually are about 3 sigma, but both in the same direction, so when you combine them you get 4.
Climate models all generally agree, at least in the direction of the changes. That means the total confidence is greater than the confidence from any single model.
Well, that's an opinion all right. Now it's in your posting history. Come back in five to ten years and review.
Given no prior knowledge, 2 sigmas is roughly 95% confidence that the observation is not a false positive. Given the risk of spending trillions to avert a possible false positive or not doing anything and having billions displaced and the world's food production severely damaged in the case of a false negative, doing something seems like a reasonable course no?
It's academic anyway. There's no way the world is politically going to do much. Fortunately it looks like the economic choice will be non-fossil in the next few years anyway. Perhaps this is the solution to the Fermi paradox: you only avoid cooking yourselves back to the stone age (with no easily exploitable oil for next time) if you mange to invent economically superior solar panels in time.
No, Physics assumes that particles behave in consistent ways, at least statistically. Simplified, the observation here seems to be that tau particles decay a little faster than we predict they should. There's some uncertainty, because there are errors in our measurements, and the decay time itself is stochastic. As you collect more data you can put narrower limits on the average decay time and become more confident that it is really different from prediction.
Two sigma is the usual standard. The magic p 0.05.
Some fields use higher standards for very high profile findings, either directly or indirectly. The FDA for example, usually requires at least two independent trials with p 0.05 to approve a new drug. In high energy particle physics it is possible to do very tightly controlled experiments that are expensive to set up but then relatively cheap to run. With this situation, getting more data is fairly inexpensive, so you can achieve very high confidence on your results just by waiting a little longer.
You should never accept anything as "true." In science we accept things as "most likely" and "meeting our prespecified level of evidence."
For observational sciences like parts of astronomy and climate, you might just have to live with what you get. For most of medicine and some of climate, we could certainly achieve five sigma confidence. It would cost orders of magnitude more than the two or three sigma we generally use now.
Both of those. Do you not follow the news? There are self-driven cars driving around all over the place. Any number of robots could cross a road as well, such as the combat robots Google makes.
I think your example reinforces his point. Q-learning, or more generally reinforcement learning, is a learning algorithm. You don't program the system, you set up some basic infrastructure and then train it by example. We've learned that such systems can learn to do things we don't understand, and cannot program.
Perception is harder than human-level chess, but not Go.
Now we've got systems that perform perception tasks AND play GO better than humans.
They were. Go is quite resistant to the brute force and play dictionary techniques used in the past on checkers and chess, which is why people wax poetic about the complexity of Go.
AlphaGo is trained using reinforcement learning, which, frankly, is such a twitchy thing that it's still surprising how well it can work.
Kasparov was beaten by a big computer programmed to play chess. AlphaGo is a very different thing.
Mars One planned to fund the project by producing a reality show.
Too bad when it gets cancelled.
Air isn't much of a problem. As Slashdot pundits are fond of pointing out, the entire planet is covered with perchlorates, which are an excellent source of oxygen. Curiosity has found nitrogen too.
Sounds like an opportunity to release a replacement. Complete with web scraper to copy over all your old... Slacks?
Sure. Because it's an app, and everything old is new again, as an app.
Slack is okay as a chat platform. Being able to divide different groups into different chats is nice. Having to create a new account each time is not.
Of course, except for some trendy colourful graphics, it doesn't really do anything a private IRC server wouldn't.
I don't know... an editor that screws around with your text without telling you, and doesn't have an option to turn it off? That sounds like a good problem to fix with high priority to me.
I suppose it might have been a good idea fifteen years ago. Possibly even ten, but they would have had to have been fast. But by five years ago there were already many smartphone apps that offered to do the same thing... for free.
Hey old friend! Still don't read posts hey? If english is not your first language it's understandable, but you really shouldn't criticize then. Let me rewrite my post for you, slowly.
OP said "Maybe I'm using the wrong editor...."
I said (edited for the angels): "You're using the wrong editor. [It's probably part of] an IDE? IDEs are evil."
By the way, I find most of your posts pointless, and the others demonstrably wrong. I *think* you managed to score both on this one:
Python gives an IndentationError exception when you have whitespace the interpreter can't parse. It points out the exact location. So if you slip in a space by accident (because you should be using tabs) then the interpreter helpfully points out exactly where it is so you can delete it.
You've never actually used Python, have you?
You're using the wrong editor. Probably an IDE? IDEs are evil.
Personally, I find braces the bad option. Whitespace IS visible, and the interpreter will happily point out where you (or your editor) screwed up. Braces are hard to keep track of, particularly when heavily nested. Of course, if you use whitespace appropriately the braces are much easier to track....
Re:No, because meaningful whitespace (Score:?)
by Anonymous Coward on Monday June 12, 2017 @09:16AM
I consider that a feature actually. Use tabs for code. When you copy and paste something from the web you have to go through line by line and convert the spaces to tabs. That forces you to go through that code you just downloaded off the web line by line....
Actually, you can almost always do a search and replace for multiple spaces, but don't tell that to my summer students.
Edit: it appears more recent evidence makes the Sahara older, although there do seem to have been wet periods in the meantime. Even so, pre-human hominids clearly made it between Ethiopia and Morocco.
The Sahara didn't exist 300,000 years ago. Instead, it seems it was a fairly lush grassland. You probably wouldn't have one dude walking between the two fathering children as he went, but it's not an unreasonable geography for genes to cross back and forth, over a thousand generations.
Clearly pre-technology hominids COULD cross the barrier because their bones have been found in both places (as have apes, other primates, and lots of other animals). The question is whether their genes could slosh around enough to keep the whole area from going off on individual evolutionary tangents. It's not unreasonable that happened.
He may well have meant there is no hope of ever observing the phenomenon directly. In 1920 there was considerable argument about whether certain "nebula" were independent galaxies or just parts of our own, and a comprehensive star survey was being prepared... containing a quarter of a million stars.
Without modern telescope technology, CCD cameras, computers, and a knowledge of just how many stars there are, it would be nearly impossible to detect that kind of effect in reasonable timeframes. Also, in 1920 only three white dwarfs were known, and nobody knew that they were compact objects.
Future knowledge of all those things would have been hard to anticipate in 1920.