[helene_t]

Now there's this Canadian company who made an "adiabatic" quantum computer which exploits the simoultaneous quantum state of a bunch of electrons in a super-conductor. Some critics say this particular implementation is impossible but most (all?) seem to agree that in principle a quantum computer could be build. A few yeards ago a single quantum gate was claimed to have been build, based on a different implementantion.

Let's see if I can reconstruct in my onw words what the idea is:

You have a Schroedinger-cat that can be in either of 2^1000 possible states and the only way to find the single state that makes your cat purr is to scan all the states. Let's say you have 1000 electrons and each represents "1" if it has spin=up and "0" if it has spin=down. Let's say that Schoedinger's cat is put into a black box so that we cannot observe it, and that we have some mechanism that induces a particular state in the cat for each state of the electron array. Now let all the electrons be in a superimposed up+down state. Then the whole system of electorns is in a superposition of 2^1000 states, and so is the cat. So far so good.

As soon as we observe the cat the wave function collapses and we might as well have started with simple, classical electrons. Besides, even if we could preserve information about all 2^1000 states we would still have to scan all of them, or build a network with 2^1000 inputs that could select preferable states in a knock-out tourney or some such.

Maybe we could observe the interference paterns of the 2^1000 wave functions without observing the cat itself. But we aren't interested in interference between states, we're interested in individual states.

Does anybody here understand how quantum computers are supposed to work?

[P_Marlowe]

Hi Helen,

I have no idea, how quantum computers work, but the scenario
with the 2^1000 different states sounds a bit like fuzzy logic.
If it comes to fuzzy logic, I can follow the basic, but not how it
works in reality.
But in the end you add up all the probabilities and you get a
number, and with the number you receive, you can answer a
specific question.

I would assume that you observe the result of those probability
fields, because you cannot observe the state of one electron
(Heisenberg?).

With kind regards
Marlowe

[paulg]

helene_t, on Jun 26 2007, 06:53 AM, said:

Does anybody here understand how quantum computers are supposed to work?

Surely you just need to ask Quantumcat?

[Al_U_Card]

Instead of just 0 or 1, the quantum superposition states can be any number in between so pretty much unlimited computing power in almost no space....

[hotShot]

Well you will need a real Expert for quantum computing for a full explanation, but for a start:

If you know how NMR-spectra are measured, this is close to the way how quantum computing works.

Imagine that it is possible to design a system with n spin's that can have the values up and down. Now you send broadband EM-radiation to create an "exited state" in which the spins are arranged in the positions you want.
After you stopped your radiation used for excitation, the system will relax to a ground state emitting exactly the radiation that would have been needed to create it.

Mathematically speaking your broadband EM signal is the set of possible solutions, the emitted radiation is the subset that solves the problem conditions defined by the spin states.

While most of this is fiction, in reality experiments with 1-8 spins have been made up to now, so it won't help you with your cats any time soon. (But if it were available, the actual computing would only take a second.)

[bid_em_up]

helene_t, on Jun 26 2007, 12:53 AM, said:

You have a Schroedinger-cat that can be in either of 2^1000 possible states and the only way to find the single state that makes your cat purr is to scan all the states.

Is this what the cat in Heinlein's "The Cat Who Walked Thru Walls" is supposed to be?

[DrTodd13]

Quantum Computers on Wikipedia

[gwnn]

I don't know, but ask me in 2 or 3 years.

[helene_t]

Tx HotShot, let me see if I understand this:

First I have to set up my electrons so that my problem becomes one of minimizing the energy of their simoultaneous state. Then I let the electrons solve the problem and finally I observe the solution by an NMR scan.

This is actually an old idea. In the 19th century, a steel mill engineer solved the problem of locating steel mills so that the sum of the transport costs for coal, iron and end product was minimized. He drilled holes in a table corresponding to the mapped locations of coal mine, iron mine and steel market. Then he knotted three strings in a three-way knot on the table and pulled the loose string ends through the holes. Finally, he attached weights to the loose ends, coresponding to the per kilometer transport costs for the mill's need (or output for end product) for coal, iron and end product, respectively. This was, without computers, a nasty numerical problem to solve, but the physcial system solved it quickly. Today, experiemental DNA computers work by the same principles.

I suppose the reason why one might prefer quantum states instead of classical states is that the tunneling effect helps us escape a local optimum that is not a global optimum. And that the quantum computer simply converges faster (I'll would have to understand quantum dynamics in order to understand that, but the Wikipedia article seems to say that).

[mike777]

Another theory of quantum computing is the multiuniverse idea.

The brunt computing is spread out to multiuniverses that capture more possibilities quickly. The best theory now says this type of computer is limited in what it can solve but very fast within those limits. Check back in 5 or ten years for new theory.

On a similiar front assume humans with chips. Or chips with dna etc.....