Is a Reverse Quantum Computer Possible?

[ChrisWilson68]

In a standard quantum computer, the computing machinery is kept in a superposition of states, with each state doing the computation on different data. It's like having the superposition of a bunch of cats in a box with each cat doing a calculation on different data. The more complex the cat, the harder it is to keep information from leaking out of the box, and once that information leaks, the superposition collapses.

But what if you viewed the cat in the box as the observer? You keep the box around the cat and to the cat the rest of the universe is in a superposition of states. As long as you keep the cat very simple, the box around it is just as small as for a similar cat acting as a computer, and it's just as easy or hard to keep information from leaking into the box as it is to keep it from leaking out. But now instead of your computation being limited by the complexity of the cat, the outside universe is the computer, and it's much more powerful. From the cat's point of view, the outside universe would need to do some computation based on some initial random data taken from quantum measurements and it would have to send the results into the box in a superposition with the results from all the other versions of the outside universe. The cat would just have to collect the superposition of the results and distill it to the useful result, just as the outside universe has to do with the results from a quantum computer before the superposition collapses.

If such a thing could be done, it would be a lot more powerful than a standard quantum computer. Quantum computers are limited by the size of the box because as the box gets bigger, it becomes harder to keep information from leaking out. With a reverse quantum computer, the part inside the box doesn't have to do the computation, it just has to handle the superposition of the results to form a final result. The size of the box would limit the number of bits in the answer, not the amount of computation that could be done.

And superpositions of the rest of the universe can do other things besides computation. They can do searches. For example, suppose there's some lost treasure known to be buried somewhere in Texas. With a reverse quantum computer, you'd lock the cat in the box, then use quantum measurements to randomly choose one place in Texas to dig a hole. Then you send the cat the suitably encoded information about where you dug and whether you found the treasure or not. One of the many versions of the outside-the-cat universe that are in superposition will have found the treasure, and the cat will be able to communicate this information to all the universes.

Could this possibly work, even in theory? Is there some theoretical problem with this that I'm missing?

Thanks in advance for any feedback!

 

[Demystifier]

Could this possibly work, even in theory?

In principle, yes. In practice, no.

The difficulty to achieve a reverse quantum computer can be compared with a difficulty to reverse the second law of thermodynamics.

 

[JesseM]

I wonder if maybe this question is inspired by somewhat misleading popular accounts of quantum computing which suggest that a quantum computer can in some sense split up the computation into a bunch of sub-computations done in "parallel universes" (which may or may not be a good way to think about it depending on your interpretation of quantum mechanics), and then if anyone of those sub-computations comes up with the desired answer, the computer somehow merges all its versions back and alerts us to the one version that had the right answer (this part is definitely wrong). If quantum computers worked like this they would be immensely more powerful than they actually are, see Scott Aaronson's discussion of "The Limits of Quantum Computers" here:

http://www.scribd.com/doc/44305346/The-Limits-of-Quantum-Computers
(a somewhat different pdf draft of that article can be found here)

And also his comment in this blog post:

Look: if you think about quantum computing in terms of “parallel universes” (and whether you do or don’t is up to you), there’s no feasible way to detect a single universe that’s different from all the rest. Such a lone voice in the wilderness would be drowned out by the vast number of suburb-dwelling, Dockers-wearing conformist universes. What one can hope to detect, however, is a joint property of all the parallel universes together — a property that can only be revealed by a computation to which all the universes contribute.

(Note: For safety reasons, please don’t explain the above to popular writers of the “quantum computing = exponential parallelism” school. They might shrivel up like vampires exposed to sunlight.)

 

[ChrisWilson68]

Thanks, JesseM. Based on your comment, I see that part of what I said was definitely wrong. Even if a reverse (or maybe "inside out" is a better term) quantum computer were possible, the search mechanism wouldn't work as I described. You couldn't just dig one hole somewhere in Texas and know where the treasure was. But the Grover black-box algorithm would still seem to apply, in which case the number of holes you'd need to dig to find the treasure would be the square root of the number of holes that will fit in Texas. That's still potentially useful.

I also understand that an inside-out quantum computer wouldn't be compatible with interpretations of quantum theory that posit that decoherence arises spontaneously or that consciousness causes decoherence. But it seems like it would be compatible with other interpretations including multi-world and hidden-variables.

So, if I understand correctly, the article and blog post you link to point out that an inside-out quantum computer would have limits on its power. They do not, however, have anything to say about whether the basic idea of an inside-out quantum computer is possible.

Is that a fair understanding?

 

[ChrisWilson68]

In principle, yes. In practice, no.

The difficulty to achieve a reverse quantum computer can be compared with a difficulty to reverse the second law of thermodynamics.

Thanks, Demystifier.

Can you provide a little more information? Would a reverse quantum computer violate the second law itself? If so, in what way? Or is there some other physical principal that implies it would be impractical? If so, what is that principal? Or is there something I'm missing about how such a thing would be be massively complex or that it would be harder to keep information leaking into a reverse quantum computer than the problem of keeping information from leaking out of a standard quantum computer?

Thanks.

 

[ChrisWilson68]

Nobody else out there has anything to say about this? Nobody at all? No opinions?

 

[JesseM]

Thanks, JesseM. Based on your comment, I see that part of what I said was definitely wrong. Even if a reverse (or maybe "inside out" is a better term) quantum computer were possible, the search mechanism wouldn't work as I described. You couldn't just dig one hole somewhere in Texas and know where the treasure was. But the Grover black-box algorithm would still seem to apply, in which case the number of holes you'd need to dig to find the treasure would be the square root of the number of holes that will fit in Texas. That's still potentially useful.

I don't have much technical understanding of quantum computing, but unless you do and can prove that such a thing is workable, I have serious doubts that it would be possible to implement the Grover black-box algorithm a system as messy and complicated as "the outside universe". Don't quantum computer algorithms require carefully preparing all the qubits in some type of special initial state which will cause the subsequent evolution of their collective wavefunction to implement the algorithm? If you added a bunch of qubits that started in some random state and interacted in a random "thermal" manner with the qubits that were intended to implement the algorithm, I would guess that the resulting decoherence would prevent the system from computing anything useful. And of course in the "outside universe" there are a huge number of such random thermal interactions happening all the time.