Monday, September 10, 2012

First Across the Rubicon: Vesuvius, the Black Box, and the Quantum Mind

In May 2011 a deal was struck between the Vancouver-based research firm D-wave, and the US' Lockheed Martin for the purchase of what has been ostensibly claimed as the first "quantum computer," for $10 million US.  Between that time and up until recently, D-wave had not endeavored to prove that their computer was performing quantum operations whatsoever.  However, a publication in nature by our friends at D-wave has addressed the widespread criticism of its product.

To my limited understanding, the paper demonstrates that D-wave is performing quantum annealing operations, through several levels of inference and games of theoretical physics. The details can be read here. However:

To clarify, D-wave's "quantum computer" is not a UNIVERSAL quantum computer. What that means is that it doesn't take advantage of the quantum phenomenon called entanglement. Instead, the processor takes advantage of another quantum process called annealing. Annealing is essentially the tendency for energy to move to its lowest state. A similar analogy would be to shake a jar of marbles consisting of different sizes, finding the smallest marbles always sink to the bottom and the largest always float to the top.

 As a result, the D-Wave system is good for creative optimization processing. More importantly, the D-wave system is setting a precedent for integrated quantum bits (qubits). The architecture of the chip, in fact, is strikingly similar to other proposed memristor and nanometer wafer designs, excluding the obvious fact of it being quantum versus classical.

Above, you can see that in Vesuvius (the d-wave QC), the qubits are highly integrated and scaled, the long-sought after goal of a quantum computing.  At this moment, independent research is being conducted to test whether D-wave's claims are indeed true.  The paper published in nature seems to address some of the issues that have come up, but not all of them. 

 In previous posts, I had discussed the elements of a quantum mind, and what might be possible with quantum computing's convergence with metamaterials. The Rubicon is an apt metaphor for the journey to quantum computing, with the important difference that when it is crossed, the bridge burns itself. No going back!  You can read more about the "BlackBox" programming system for the D-wave here. In short, the annealing process "chooses" the best binary strings for an output of a Real number. To use the examples outlined on D-wave's site:

G(x1, x2,...xN)

This is a binary string output - binary consisting of a string of zeroes and ones that represent a real number (once the annealing process is complete). Additionally, the BlackBox system allows the user to program for highly complex parallel problems like the "travelling salesman" problem...

"The input bit string represents a 'potential solution', and the number returned by the function given that input gives a measure of the goodness of the potential solution - the lower G(x1,x2,...,xN) is, the better the potential solution (x1,x2,...,xN) is."

 So clearly what we're seeing is not a quantum mind. Aside from the fact that the D-wave processor is not performing at the level of a universal computer, the amount of qubits on the chip is insufficient for that purpose anyhow.  But twice that many qubits on a universal QC would be inconceivably powerful (literally, inconceivable because we'd have to build a universe-sized classical computer to simulate it). Such a computer would be quite good at thinking on its feet (on its chips?).

There is a distinction between human thought and "quantum thought" that is important to outline.  Neurons act like elecrochemical logic gates - action potentials are either inhibitory or excitory. As such, binary is still the programming language of choice, even though our brain uses neurotransmitters, cells, and genes instead of electrons, silicon, and copper! While it is enough to consider that there is an eery universality of thought (researchers have built computers using water droplets, gears and levers, beakers of DNA, bacterial colonies, etc.); what is more ominous, in fact, is the idea that quantum computing isn't really bound by time in the way that we are.

Our minds are causal generators. We build hierarchical models of the world in a series of catergorizations; of expected results in cause and effect.  Causality, however, is itself a binary process. 1-1-0-0-1-1-0-1-1 - each number or event occurs in succession, whereas a quantum computer performs those binary choices simultaneously. As such, the unachievable idea of infinity inches closer. The true irony of the quantum journey is that if we are ever successful in building a QC, we will have no possible way to confirm it! Decoherence occurs, and observation destroys quantum phenomenon, as exhibited in the two-slit experiment.  So when we raise the hood of the QC, the qubits all become hushed and suspicious ones and zeroes, until we close it again and the party can resume.  In this way, a quantum computer would have to become self-aware in order to prove that it is indeed quantum. Although I think it will be obvious when such a computer begins performing quantum algorithms successfully, I'm much more interested in the first instance where mathematics can go beyond itself, in the Godel sense of things. When this new form of processing inverts, as many classical systems tend to, we will likely see the emergence of an entirely new class of mind, a speciation that branches from technology and humanity's long-standing marriage. We may all be in the thrall of a QC before long, not because of its power, but because of the potential wisdom (in human terms, those are never the same thing).  So here's to the cat in the box - dead or alive - I'll pay attention when it proclaims "hear me, for I am quantum."

No comments:

Post a Comment