Slashdot Mirror
First Quantum Computing Gate on a Chip
An anonymous reader writes "After recent success in using quantum computing for superconducting qubits, researchers from Delft have formed the first Controlled-NOT quantum gate. 'A team has demonstrated a key ingredient of such a computer by using one superconducting loop to control the information stored on a second. Combined with other recent advances, the result may pave the way for devices of double the size in the next year or two--closer to what other quantum computing candidates have achieved, says physicist Hans Mooij of the Delft University of Technology in the Netherlands. Unlike today's computers, which process information in the form of 0s and 1s, a quantum computer would achieve new levels of power by turning bits into fuzzy quantum things called qubits (pronounced cue-bits) that are 0 and 1 simultaneously. In theory, quantum computers would allow hackers to crack today's toughest coded messages and researchers to better simulate molecules for designing new drugs and materials.'"
I wouldn't know for sure, but I don't think it's valid to compare any form of computing based on binary logic with quantum computing. I searched for "quantum transistor" and found this, which makes use of the term to refer to transistors that rely on principles of quantum mechanics to function properly. This would be relevant to conventional computing, but not quantum computing. If I understand correctly, quantum computing is not a replacement for binary logic computing, but an alternative or supplement.
Most of the hard logistics problems are forms of nonlinear optimization and usually reduce to integer programming, which is solved by a pretty blind search (excluding tricks like branch and bound, pruning; it's still a tree search). Improving these solutions by a few percent translates into a lot of money very quickly.
Esoteric, maybe (it'll be a computer in a lab, not a PC), but usable nonetheless.
Wrong, get Gordan Freeman to throw the switch, no amount of Marines are gonna be enough.
Disclaimer: I am a quantum information scientist.
If you re-read the article, you'll see that the gate is a controlled-NOT (aka CNOT) gate, rather than a simple NOT gate. CNOT is a two-bit (or, in this case, two-qubit) gate. Simply being able to make and maintain single qubits is challenging (at least, for superconducting systems), manipulating single qubits is more challenging, and performing two-qubit operations is extremely hard. It's worth noting that this result is not the first example of a two-qubit gate in a superconducting system; rather, it appears to be the most complete so far.
It's also worth noting that quantum CNOT gates were achieved years ago in other physical systems (NMR, ion traps, etc.). Part of the reason people are so excited about this is that, by virtue of it being on a chip, we may be able to apply the enormous amounts of chip-fab technology we already posses to scaling up the system.
Disclaimer: I am a quantum information scientist
Qubits represent a probability of being a 0 or 1. Observing a qubit destroys that probability, and you "read" only a zero or a one.
This is at best an incomplete description of what happens. Qubits are quantum states, not probabilities. Quantum states are sometimes called "probability amplitudes", in that taking the square of the magnitude of the coefficient for a particular basis state gives you the probability of getting that state if you measure in that basis. There are a few very important points: (1) we're dealing with complex numbers, and things work in such a way as to give us the possibility of "interference" of probability amplitudes; (2) quantum states are real states, not just probabilities representing our ignorance of which classical state you'll find when you measure.
A brief intro to the math:
Let's take some qubit in some arbitrary state, which we'll call | psi > (I'm using Dirac notation). We can completely describe the state as follows:
| psi > = a | 0 > + b | 1 >,
where a and b are complex numbers, and have the property that |a|^2+ |b|^2 = 1. We see that we have an uncountably infinite number of possible states for just a single qubit. If psi were a classical bit instead of a quantum bit, we could use essentially the same description, except that the requirement on a and b would then be that exactly one of them is 1, while the other is 0 (only two possible states). If psi were a "classical" analog "bit" or a probabilistic bit, the requirement would be that a, b in [0,1], and a+b=1.
What happens if we measure psi? It depends on the basis we choose to measure in, but if we go to measure psi in the {| 0 >, | 1 >} basis, we'll get | 0 > with probability |a|^2, and | 1 > with probability |b|^2. Figuring out probabilities for other bases requires only a basis transformation (simple linear algebra).
Now, this qubit business seems horribly messy--we have an infinite number of states for a single qubit--how can we possibly describe the action of a two-qubit gate like controlled-NOT (CNOT)? Fortunately, quantum mechanics is linear, which means that if we describe how a gate operates on each of the possible input basis states, we've completely specified the gate. For two qubits, we can use the following basis: {| 00 >, | 01 >, | 10 >, | 11 >}. Labeling the rows and columns in that order, we get the following truth table for the CNOT gate:
1 0 0 0
0 1 0 0
0 0 0 1
0 0 1 0
In other words, if the first bit is 0, do nothing to the second bit. If the first bit is 1, flip the second bit.
It turns out that CNOT plus a bunch of different single qubit gates is universal, meaning you can use that set of gates to implement any "quantum circuit".