Quantum computers in the next 50 years?

[Lee]

I've recently read a few articles on the subject of quantum computers, and I'm curious to when they are going to be developed to the point that they are a recognisable way to process information.

Do you think they will be in use within the next 50 years? and why?

 

[Locrian]

I'd like to toss in my vote that they will not be in use within 20 years of today, and maybe not 50 either. The problems that will be encountered as the qubits increase to a useful value will be tremendous, and I believe physicists involved have underestimated the problems in getting useful results from a device that depends on quantum computing.

 

[Lee]

A scaled down version of 7 qubits has already been developed in the lab by IBM to factor 15, though the application is trivial, but the point is that it is demonstrated in principle to be possible, so I will take my bet to be within 20 years for them to see the light of the day, also for your kind information, quantum cryptography, a branch of quantum computation as produced devices which are already there in the market and there is a full quantum network working underground in US in defense applications. 3 Firms are working to make quantum products like idquantique.

Indeed they have but it is predicted that by using that process (a liquid NMR QC) you can only really have a 10 qubit version.

 

[Lee]

Ah, ok. So are there any methods of creating quantum computers on the 100-1k qubits quantum computers on the table currently?

I found http://physicsweb.org/articles/news/10/11/10/1" article from the iop website encouraging for QCs.

 

[cartuz]

Perhaps, Quantum Mechanical Description it is possible to complete by hidden variables. Than Quantum Mechanics will be the part of Classical Physics. In this case Quantum Computer is a Classical Probability Computer. I'm not sure.

 

[cartuz]

The proposition of Bell's Theorem by John Bell and its validation by Alain Aspect in 1982 or so fundamentally rules out any Local Hidden Variable Model of QM and it is a celebrated result, there are Non-locality game in Quantum Information which has in principal no counterpart in classical physics, hence QC is fundamentally different from a classical probability computer, see Toffoli's and others work on Quantum TUring machines which are different from classical Turing machines we know of.

May be that’s right but now I have our program on PASCAL-DELPHY which can simulate the entanglement state on my notebook. But Bell's Inequality don't violet. If somebody say to you that entanglement is a Quantum Effect only you can don’t trust him.
I almost sure that in future somebody can possible the classical simulation the entanglement which will be violet Bell's Inequality! It is will be serious fundamentals results.

 

[florianb]

May be that’s right but now I have our program on PASCAL-DELPHY which can simulate the entanglement state on my notebook. But Bell's Inequality don't violet. If somebody say to you that entanglement is a Quantum Effect only you can don’t trust him.

Sure, a computer can simulate quantum effects : You can solve Schrödinger's equation on a computer, this doesn't mean that this equation is classical.

By putting a little more effort in your simulation, you will be able to violate Bell's Inequality too. There is no doubt about this.

And by the way, entanglement is clearly a quantum effect, as there is no classical equivalent to it. I'll give you a point though : There exist non-local hidden variable theories that also violate bell inequalities and basically account for all experimental facts up to now, but I wouldn't call them "classical".

 

[cartuz]

Sure, a computer can simulate quantum effects : You can solve Schrödinger's equation on a computer, this doesn't mean that this equation is classical.

By putting a little more effort in your simulation, you will be able to violate Bell's Inequality too. There is no doubt about this.

And by the way, entanglement is clearly a quantum effect, as there is no classical equivalent to it. I'll give you a point though : There exist non-local hidden variable theories that also violate bell inequalities and basically account for all experimental facts up to now, but I wouldn't call them "classical".

It is very interesting opinion.
But there is possible to think that we can to complete Classical Physics by non-local hidden variables and understand Quantum Mechanics.
What is it?
Newtonian Physics is the physics with inertial reference. But it is ideal and non-real case! If we suppose that a real system reference is non-inertial than it is will be non-local hidden variables http://xxx.lanl.gov/abs/quant-ph/0611053
From this follow that Uncertainty Relation is a Systematic Measurement Errors in non-Newtonian Physics with non-inertial systems reference.
What is non-Newtonian Physics? It is physics of non-inertial systems reference where Lagrangian is not depend of coordinate and velocity only. But in non-inertil systems reference Lagrangian is depend of acceleration and derivatives of acceleration too. I hope you are agree with this. When we simulate qubits, we don't use Schrödinger's equation but we use random non-inertial systems reference.

 

[florianb]

It is very interesting opinion.
But there is possible to think that we can to complete Classical Physics by non-local hidden variables and understand Quantum Mechanics.
What is it?
http://xxx.lanl.gov/abs/quant-ph/0611053

You're right in saying that non-local hidden variables theories can reproduce the results of Quantum Mechanics. See for example Bohmian mechanics :
http://plato.stanford.edu/entries/qm-bohm/

However, I find that the article your refer to makes assumptions that are doubtful in the context of classical mechanics :

- The form of the uncertainty relation (5) is borrowed from quantum mechanics. Why should the errors obey such an inequality ? Unless my ignorance is at fault, there is no such thing as an equation linking the "errors" on p and x together in classical mechanics.

- Assuming (5) and then expecting that the constant is Planck's constant is nothing else than assuming Heisenberg's inequality, which is quantum mechanical in essence.

The model described in this paper is therefore semi-classical.

So, this paper proves that using a semi-classical model, some results from quantum mechanics can be expected. I'm not really suprised.

All the best !

 

[cartuz]

You're right in saying that non-local hidden variables theories can reproduce the results of Quantum Mechanics. See for example Bohmian mechanics :
http://plato.stanford.edu/entries/qm-bohm/

However, I find that the article your refer to makes assumptions that are doubtful in the context of classical mechanics :

- The form of the uncertainty relation (5) is borrowed from quantum mechanics. Why should the errors obey such an inequality ? Unless my ignorance is at fault, there is no such thing as an equation linking the "errors" on p and x together in classical mechanics.

- Assuming (5) and then expecting that the constant is Planck's constant is nothing else than assuming Heisenberg's inequality, which is quantum mechanical in essence.

The model described in this paper is therefore semi-classical.

So, this paper proves that using a semi-classical model, some results from quantum mechanics can be expected. I'm not really suprised.

All the best !

It is classical description of microobjects.
Nothing Quantum is there.
This classical description is compare with QM.

 

[florianb]

It is classical description of microobjects.
Nothing Quantum is there.
This classical description is compare with QM.

You're right. I read the paper a little too fast.

Although I still maintain my point that the form of equation (5) is odd for a classical derivation (I'm waiting for pointer to a proof to the contrary), the last part of my post was not correct, and I apologize for it :

However, there are additionnal points that I still find dodgy in the paper, particularly when the comparison is made :

- Why is \alpha_3 = \cfrac{i\hbar m}{2} chosen ?
- The wavefunction e^{i\frac{S}{\hbar}} does remind me of the path integral formulation of quantum mechanics. The "path-integral" part being left out, I don't understand why this particular wavefunction is chosen.

Many thanks for clarifying those three points.

 

[florianb]

I've recently read a few articles on the subject of quantum computers, and I'm curious to when they are going to be developed to the point that they are a recognisable way to process information.

Do you think they will be in use within the next 50 years? and why?

There WILL be a quantum computer running in 49 years and 9 months. Why ? I have a bet running on this exact issue ;-)

 

[cartuz]

You're right. I read the paper a little too fast.

Although I still maintain my point that the form of equation (5) is odd for a classical derivation (I'm waiting for pointer to a proof to the contrary), the last part of my post was not correct, and I apologize for it :

However, there are additionnal points that I still find dodgy in the paper, particularly when the comparison is made :

- Why is \alpha_3 = \cfrac{i\hbar m}{2} chosen ?
- The wavefunction e^{i\frac{S}{\hbar}} does remind me of the path integral formulation of quantum mechanics. The "path-integral" part being left out, I don't understand why this particular wavefunction is chosen.

Many thanks for clarifying those three points.

In Aristotle’s Mechanic the velocity is determine the dynamic.
In Newton's Mechanic the acceleration is proportional to force but the system of reference is inertial.
Ostrogradskii's Mechanic is the general case of mechanics when consider any system of reference include the non-inertial and moving with changing acceleration. This paper Professor of Sankt-Petersburg University was written in 1850. And now we are see that this mechanic is suitable for the description of quantum objects! But this is classical description. From this mechanic is following the General Euler-Lagrange function and General Jacobi-Hamilton function. General in the sense of non-inertial systems reference and include the non-constant Coriolis forces. The General Jacobi-Hamilton equation for the action function is non-linear but if we write this equation for the exponential function we obtain the linear equation which name the Shredinger's equation. The General Jacobi-Hamilton equation for action function is equal to the Shredinger's equation for the exponential action function when \alpha_3 = \cfrac{i\hbar m}{2}. I glad to write the answer to you.

 

[florianb]

In Aristotle’s Mechanic the velocity is determine the dynamic.
In Newton's Mechanic the acceleration is proportional to force but the system of reference is inertial.
Ostrogradskii's Mechanic is the general case of mechanics when consider any system of reference include the non-inertial and moving with changing acceleration. This paper Professor of Sankt-Petersburg University was written in 1850. And now we are see that this mechanic is suitable for the description of quantum objects! But this is classical description. From this mechanic is following the General Euler-Lagrange function and General Jacobi-Hamilton function. General in the sense of non-inertial systems reference and include the non-constant Coriolis forces. The General Jacobi-Hamilton equation for the action function is non-linear but if we write this equation for the exponential function we obtain the linear equation which name the Shredinger's equation. The General Jacobi-Hamilton equation for action function is equal to the Shredinger's equation for the exponential action function when \alpha_3 = \cfrac{i\hbar m}{2}. I glad to write the answer to you.

Thanks for clarifying one of my questions.

So this \alpha_3 constant is chosen so that Schrödinger's equation is reproduced for a particular wavefunction... It doesn't look to me as a proof that QM was derived from CM. It looks a lot more like what can be found in some textbooks, where Schrödinger's equation is "derived" from classical mechanics (although most books state clearly that it is not a derivation).

In fact classical and quantum mechanics have a lot of formalism in common. That's why it is easy to quantize a classical system... and relatively easy to go from CM to QM by making some assumptions.

By the way, the paper talks about the coriolis force, which is a purely classical effect, and which can be modeled by CM. I bet that people throwing satelites into orbit have to take the coriolis force into account, and that they don't use Q-words to calculate the trajectories.

 

[Lee]

There WILL be a quantum computer running in 49 years and 9 months. Why ? I have a bet running on this exact issue ;-)

What kind of method, or even qubit do you think it shall be created from? Why do you think there willbe one working in the next 50 years? Or could anyone point me in the right direction towards some papers that would help me come to my own conclusions about this subject.

 

[florianb]

What kind of method, or even qubit do you think it shall be created from? Why do you think there willbe one working in the next 50 years? Or could anyone point me in the right direction towards some papers that would help me come to my own conclusions about this subject.

Dear Lee,

You will find a good, easy to read, review of how qubit can be implemented in Chuang and Nielsen's book : "Quantum Computation and Quantum Information", along with a lot of other interesting stuff.

To be fair, answering such a question is pure speculation. There is definitely a need for a breakthrough, because all current implementations have problems. So I replied "yes" (and was silly enough to bet on a "yes") because I think that 50 years is enough for a breakthrough to happen. This breakthrough might be a new qubit system with incredible properties, a technological advance allowing the current methods to scale or a new insight into the way we design quantum algorithms.

Hope this helped !

 

[setAI]

Cluster computation looks quite promising-


http://xxx.lanl.gov/abs/quant-ph/0508218

 

[Lee]

That book looks mighty interesting and I'll see if the university library have a copy, as I don't fancy shelling out the £40ish to buy a copy for myself, thanks for your insight into the situation.

Thanks for the link to the paper, I will try and give it a read and maybe include it in the paper I have to write.

 

[florianb]

Cluster computation looks quite promising-

http://xxx.lanl.gov/abs/quant-ph/0508218

Agreed, cluster states are quite interesting. Among their advantages :

• Each computation starts from the same state, so there is only one state we need to be able to prepare
• The only thing that need to be done afterward is to measure qubits

There are also drawbacks :

• The number of qubits needed to perform a certain calculation is higher than with a universal reversible q-computer.
• Although this is not a requirement set by the original paper, all implementation that I know have suffered from a "probabilistic" nature : The computer is run many times, until the right conditions are met. *

If you are interested in the cluster state model, I would recommend that you read the Raussendorf/Briegel papers as well : PRL 86 910 and PRL 86 5188 (2001)

If you're even more interested, there is also a recent paper on the arxiv describing the implementation of the Deutsch-Josza algorithm using a cluster state : quant-ph/0611186

happy reading !

* It seems that the "Retry until success" scheme (which I didn't know until now), also fit into this category, or, am I wrong ?