Andy Hall, a creator physics games, has tweeted about a rather amazing article in Ars Technica,
Download FoldIt now
The people behind FoldIt had previously created Rosetta@home which (just like the Mersenne prime project and others) used the idle time of computers of many volunteers to solve "protein folding" tasks for the molecular biochemists.
With Rosetta@home, users were frustrated that they saw solutions that their computer didn't but they had no way to help their pet computers in the desperate situation. So the creators of FoldIt created something that uses one of the most clever peripherals attached to the users' computers most of the time: the users' brains. ;-)
Since 2010, tens of thousands of players of FoldIt were producing results that solved tasks that the professional biochemists couldn't, and those results were published in Nature and elsewhere. However, there is a brand new development described in Nature today:
Instead of FoldIt, EteRNA, and EyeWire which were remarkable but only used the gamers' biological brains' idle time to solve basically classical tasks, the authors of the article above described the results of Quantum Moves, the first volunteer distributed game that actually solves serious optimization tasks in quantum physics!
The combination of the high-brow and clever aspects employed by this game is rather amazing.
Here you have a 3-weeks-old demo of Quantum Moves, the game:
The game is available for Windows (just 12 MB), Mac, iOS, as well as Android. You're basically doing some moves with something representing wave functions that are meant to be "equivalent" to operations done by a quantum computer with 300 atoms that the research group is trying to build.
I've played the game Рand have beaten its 23 levels in an hours (with some score I don't claim to be amazing). You are basically reshaping a smooth potential \(V(\hat x)\), basically by moving and/or deepening a well or several wells at the same time (while the potential has other wells or terms in it) Рwhich will be done to the atom by lasers Рwhile you are trying to keep the atom's wave function (which is shown at all times) "mostly near the center of the dip", and perhaps move the atom elsewhere at the same moment. The wave function obeys the time-dependent Schr̦dinger's equation given by the potential you influence. And the wave function behaves like a "liquid" of a special sort Рthis language is obviously misleading because the real interpretation of the wave function is probabilistic while you can "observe" its model on the screen. With minutes or hours you invest into playing, you are learning how the "liquid" responds to you moves.
In the paper, the authors remarkably demonstrated that using their intuition and heuristic approaches, the human players were able to find solutions to tasks in which the well-known classical optimization algorithms don't work well – but the quantum computers would. The well-known classical optimization algorithms fail especially near the "quantum speed limit", when the shortest process duration is combined with perfect fidelity. It means that many human players can move the potential and the wave function to its nearly precise target shape (perfect fidelity) at minimum time (yes, it's better if you solve each level quickly), while computers with traditional algorithms couldn't do such things.
Obviously, I don't think that it proves that the human brains' physiology is an example of a quantum computer – and I am confident that they don't suggest such a thing, either. The temperature in the brain is too high and the decoherence is too fast, I think. But the heuristic strategies used by the humans seem new. The researchers were able to figure out what these strategies were and improve their "professional" methods to deal with such problems, too.
At the end, I think that the lesson is that all the basically repeated human work may be replaced by computers. But the heuristic brains of (even rather ordinary) humans seem to be able to do many clever things that the "experts in optimization" just don't immediately exploit. Maybe, one could interpret this situation by suggesting that the experts (and even many other people) think that the right or best computing and thinking (and even the laws of physics) are much more "deterministic and mechanical" than the actual one.
A classical computer says "do this exact step, that exact step". Sometimes, an algorithm may use a (pseudo)random generator to achieve certain outcomes. But this is still working within the framework of classical physics. The quantum mechanical laws are "more general" and produce outcomes even if the intermediate states and steps aren't objectively described or seem "fuzzy". Similarly, the human thinking often ends up with answers even though it's not exactly clear how it was done, what was happening at each moment.
Just like in the funny cartoon, "here a miracle occurs" sometime in the middle, and the human brain manages to "tunnel" to an answer that is sometimes very clever. Quantum mechanics is sort of similar in the sense that it predicts nonzero probabilities for final outcomes (of measurements) "directly", without decomposing the calculation to totally well-defined steps. But again, I emphasize that it is just a vague analogy of the two spirits; I believe that the human brain is not a quantum computer in any sense.
Because the authors of Quantum Moves have become so amazed by the power of the human mind, they are already releasing a new game, Quantum Mind, so far for Macs but soon for Windows, too.
Gamers help satisfy the need for speed in quantum manipulations,which talks about the successes of a high-brow game named FoldIt and developed initially in 2008 and especially about its young sibling, Quantum Moves.
Download FoldIt now
The people behind FoldIt had previously created Rosetta@home which (just like the Mersenne prime project and others) used the idle time of computers of many volunteers to solve "protein folding" tasks for the molecular biochemists.
With Rosetta@home, users were frustrated that they saw solutions that their computer didn't but they had no way to help their pet computers in the desperate situation. So the creators of FoldIt created something that uses one of the most clever peripherals attached to the users' computers most of the time: the users' brains. ;-)
Since 2010, tens of thousands of players of FoldIt were producing results that solved tasks that the professional biochemists couldn't, and those results were published in Nature and elsewhere. However, there is a brand new development described in Nature today:
Exploring the quantum speed limit with computer gamesWhen you click this hyperlink at TRF, you only get the abstract, but if you click the very same "10.1038" link at ArsTechnica, the full PDF paper opens. This is what the SJWs call "discrimination". ;-)
Instead of FoldIt, EteRNA, and EyeWire which were remarkable but only used the gamers' biological brains' idle time to solve basically classical tasks, the authors of the article above described the results of Quantum Moves, the first volunteer distributed game that actually solves serious optimization tasks in quantum physics!
The combination of the high-brow and clever aspects employed by this game is rather amazing.
Here you have a 3-weeks-old demo of Quantum Moves, the game:
The game is available for Windows (just 12 MB), Mac, iOS, as well as Android. You're basically doing some moves with something representing wave functions that are meant to be "equivalent" to operations done by a quantum computer with 300 atoms that the research group is trying to build.
I've played the game Рand have beaten its 23 levels in an hours (with some score I don't claim to be amazing). You are basically reshaping a smooth potential \(V(\hat x)\), basically by moving and/or deepening a well or several wells at the same time (while the potential has other wells or terms in it) Рwhich will be done to the atom by lasers Рwhile you are trying to keep the atom's wave function (which is shown at all times) "mostly near the center of the dip", and perhaps move the atom elsewhere at the same moment. The wave function obeys the time-dependent Schr̦dinger's equation given by the potential you influence. And the wave function behaves like a "liquid" of a special sort Рthis language is obviously misleading because the real interpretation of the wave function is probabilistic while you can "observe" its model on the screen. With minutes or hours you invest into playing, you are learning how the "liquid" responds to you moves.
In the paper, the authors remarkably demonstrated that using their intuition and heuristic approaches, the human players were able to find solutions to tasks in which the well-known classical optimization algorithms don't work well – but the quantum computers would. The well-known classical optimization algorithms fail especially near the "quantum speed limit", when the shortest process duration is combined with perfect fidelity. It means that many human players can move the potential and the wave function to its nearly precise target shape (perfect fidelity) at minimum time (yes, it's better if you solve each level quickly), while computers with traditional algorithms couldn't do such things.
Obviously, I don't think that it proves that the human brains' physiology is an example of a quantum computer – and I am confident that they don't suggest such a thing, either. The temperature in the brain is too high and the decoherence is too fast, I think. But the heuristic strategies used by the humans seem new. The researchers were able to figure out what these strategies were and improve their "professional" methods to deal with such problems, too.
At the end, I think that the lesson is that all the basically repeated human work may be replaced by computers. But the heuristic brains of (even rather ordinary) humans seem to be able to do many clever things that the "experts in optimization" just don't immediately exploit. Maybe, one could interpret this situation by suggesting that the experts (and even many other people) think that the right or best computing and thinking (and even the laws of physics) are much more "deterministic and mechanical" than the actual one.
A classical computer says "do this exact step, that exact step". Sometimes, an algorithm may use a (pseudo)random generator to achieve certain outcomes. But this is still working within the framework of classical physics. The quantum mechanical laws are "more general" and produce outcomes even if the intermediate states and steps aren't objectively described or seem "fuzzy". Similarly, the human thinking often ends up with answers even though it's not exactly clear how it was done, what was happening at each moment.
 |
Just like in the funny cartoon, "here a miracle occurs" sometime in the middle, and the human brain manages to "tunnel" to an answer that is sometimes very clever. Quantum mechanics is sort of similar in the sense that it predicts nonzero probabilities for final outcomes (of measurements) "directly", without decomposing the calculation to totally well-defined steps. But again, I emphasize that it is just a vague analogy of the two spirits; I believe that the human brain is not a quantum computer in any sense.
Because the authors of Quantum Moves have become so amazed by the power of the human mind, they are already releasing a new game, Quantum Mind, so far for Macs but soon for Windows, too.
Citizen scientists (gamers) beat experts, PCs in protein folding, quantum tasks
Reviewed by MCH
on
April 14, 2016
Rating:
Reviewed by MCH
on
April 14, 2016
Rating:
No comments: