Friday, May 16, 2014

A Quantum Computing "Mind Machine" To Solve All Human Problems? Don't Believe It!

In an earlier blog post I pooh-poohed the notion of the "singularity" and "ultimate computers" manifested in some ultra-artificial intelligence embodied in a supercomputer. Though Stephen Hawking too is evidently freaking out about the potential for these things to emerge, and wreak havoc - I continue to consign them to scifi flicks like "The Forbin Project".'

At the core of most of the novel hype is none other than the quantum computer. Never mind that a properly working one -   that can perform a multiplicity of tasks-   is likely over a century from manifestation, people and techie media mavens are obsessed with them. (In physics, the main issue and problem concerns "entanglement" which up to now no one has been able to solve.).

Imagine then a computer - call it a "quantum computer" - that obeys quantum rules. The most fundamental of these is known as superposition. Thus, if I have some wave function U, it can be written as:

U = U1  + U2.  + U3.  +. ..........UN

Where the sum represents a superposition of wave states contained within U. According to the Principle of Superposition, each wave state can be added to the other wave states in the ensemble without changing the nature of U. In Paul Dirac's terminology it implied that "the state of a system is defined as a state of undisturbed motion that it restricted by as many conditions as possible without mutual interference or contradiction". ('The Principles of Quantum Mechanics, p. 11)

Now, applied to computing, one confronts the quantum bit or qubit, as opposed to the ordinary bit. The latter may be 1 or 0 but the former can be 1 AND 0 at the same time, The linkage of 1 and 0 is then a superposition of states, i.e. U = U(1) + U(0).  Thus, in its superposed state  a quantum bit exists as two equally probable possibilities. According to one hypothesis from David Deutsch, the quantum bit is operating in two slightly different universes at the same time - one for which it's 1 and the other U(0) for which it's 0.  In Deutsch's parlance "it's the first technology to allow useful tasks to be performed in collaboration between parallel universes."

Most noteworthy here, if a qubit can be in two states at one time, it can perform two computations at the same time.  Then two qubits could perform four simultaneous computations, three could perform 2^3 = 8 and so on. Now, if the supercooled niobium chip at the heart of one specific quantum computer - called D Wave 2, has 512 qubits in play, that means in theory it can perform 2^512 operations simultaneously.

Quick! Trick question: Identify the coldest place or entity in the universe. Most astronomers, astrophysicists would doubtless name the Boomerang Nebula 5,000 light years away where the temperature is estimated to be minus 458 C of nearly at absolute zero Kelvin.   But even that isn't the coldest place. Nope, it's actually in a burg called Burnaby - directly east of Vancouver, B.C.  Burnaby is the HQ for D-Wave Two, a putative quantum computer of which there are now 5 in existence. Externally one sees a 10' high black box and if you could peer inside you'd see a cylindrical cooling apparatus containing a niobium computer chip chilled to 20 millikelvins or minus 459.6 K, nearly 2C colder than the Boomerang Nebula.

As fancy and formidable as D-Wave 2 might be, it's nowhere near ready to act as a quantum "mind machine" or indeed, any basis for artificial intelligence.  In many ways quantum computers are a solution still looking for the right problem. Example: Feed it the finest resolution solar images in H-alpha, monochromatic light and UV then all the ancillary magnetic data - say from vector magnetographs - and have it build a self-consistent model that accounts for both CMEs (Coronal mass ejections) and solar flares. It can't do it! The problem is beyond its capacity (as it is for all human researchers).

The bugbear? Existing so-called quantum computers have to be maximally isolated, like the D-Wave 2 in Burnaby, B.C. No information can be allowed to escape because any interaction with the outside world will cause errors to creep into any calculations, say just to find the changing magnetic gradient at any instant in a complex (quadripolar) sunspot field.  Even the most basic computations are made even more difficult by the fact that in the isolated state you still have to control them.

Not surprisingly, quantum computer mavens and techies have had to get around these problems by compromise. Enter then what's called an adiabatic quantum computer which works by means of a process called quantum annealing. Basically, the qubits are linked together by couplings. These couplings are then programmed using a special algorithm that specifies certain interactions between the qubits. (If this is a 1, then that has to be a 0, and so forth).  The qubits are then put into a state of quantum superposition in which they're free to explore all those two to the nth qubits possibilities simultaneously.   They are then allowed to settle back into a 'classical' state to become separate 1s and 0s again.   The qubits naturally seek out the lowest energy state consistent with the requirements specified in the original algorithm - and the answer can be read in the final qubit configuration.

What's the hang up? Well, the adiabatic quantum computer can only solve one class of problem, which goes by the moniker "discrete combinatorial optimization". This type  entails finding the best (optimum), the shortest or the fastest (or the cheapest or most efficient) way of doing a task.  For example, say a European traveler wishes to go to Paris, London, Berlin, Z├╝rich and Rome all in one week with the cheapest transport feasible but also the highest quality he can buy. How does he do it? Quantum annealing computers can provide an answer,

But alas, they can't provide an answer to the solar model problem described earlier. That one is simply not in the nature of the problems it's able to solve. Some might argue, 'We'll maybe it can select the best solar model - say for flares'.  But this can't work since it's not really a (routing) optimization problem. The computer would still have to compare a vast number of plasma parameters forecast for each model, as well as the topological (i.e. twisting) conditions to attain maximal magnetic stress for each - given as a quantitative estimate for a set of the same ARs, and then make a determination. There's no way currently isolated quantum computers could do it.

On the positive side, we can be thankful that existing quantum computers (which the NSA is also interested in) won't let them come up with the ultimate de-encryption scheme to break any encryption now matter how many primes or combinations thereof are use. Thus, the adiabatic quantum computer is little or no use to the NSA (say in designing an ultra-PRISM) program because it won't run the sort of code-cracking algorithm the spooks are interested in. Which is good for the rest of us who value the 4th amendment.

Oh sure, it's only a matter of time before current limitations on quantum computing are eventually surpassed, but I don't see anything approaching a "mind machine" or true AI for a long time - if ever.  The ultimate cap on that aspiration may well be Peak Oil. Because if you run out of the energy (as well as water) to keep these computers running you aren't going to be able to do much at all. Not even compute the best route for a postman to take in a dog-infested city neighborhood.

See also:

No comments: