What gives quantum computers that extra oomph over their classical digital counterparts? An intrinsic, measurable aspect of quantum mechanics called contextuality, it now emerges. See Article p.351
For decades, researchers have struggled with the question of what makes quantum computers so powerful, and the answer has been as elusive as an understanding of quantum physics itself. Is there some unique feature of quantum physics that is responsible for enabling quantum computers to perform certain computations faster than their conventional digital counterparts? Many of the more exotic properties of quantum mechanics have been put forward as possible candidates, but so far none has held up to scrutiny. On page 351 of this issue, Howard et al.1 uncover a remarkable connection between the power of quantum computers and one of the stranger properties of quantum theory known as contextuality.
Designs for quantum computers often mirror those of conventional computers, in that they are built out of basic components such as logic gates that perform elementary operations on quantum bits of information. A commonly used set of operations for a quantum processor is known as the stabilizer operations2. These operations are designed using the rules of quantum physics, but are in many ways similar to those used by a classical machine. For example, initializing quantum bits to a value of 0 or 1, reading out these binary values or flipping them are all stabilizer operations (as well as more-exotic ones). In fact, within this limited set of building blocks, it is often possible to imagine that the quantum bits are simply described by pairs of bits (0's and 1's) that are initialized, processed and measured by stabilizer operations much like the bits in a digital computer3,4. The restricted 'classical' nature of the stabilizer operations lets quantum engineers design error-correcting codes and logic gates that are tolerant when things go wrong.
Any quantum machine that computes using only these stabilizer operations is no more powerful than your desktop computer5,6. So how can we supplement this set to build a quantum computer? There are several approaches, but by far the most common is to provide the computer with a large number of additional quantum bits that are initialized in a peculiar way, using a quantum superposition of the usual stabilizer initializations7. A quantum bit described by a superposition possesses characteristics of both binary values 0 and 1 simultaneously. This way of initializing the quantum bits is called magic — a rather suitable name for some of the quantum weirdness that contradicts our everyday experience. Supply a processor that uses only stabilizer operations with quantum bits initialized as magic states and — hey presto! — that limited machine is endowed with the full power of a quantum computer.
If a quantum computer that uses only stabilizer operations is stuck in the slow lane together with today's run-of-the-mill digital computers, but can be 'boosted' to a powerful quantum computer by being supplied with magic states, then these magic states must hold the key to the quantum computer's increased performance. So what is so special about magic states? The answer provided by Howard et al. comes from studying how these states might also be described using pairs of bits for each quantum bit, as we could for stabilizer operations. The authors formalize this perspective by using a non-contextual hidden-variable theory, which is a way of describing the properties of a quantum particle or device using the values (hidden variables) of a number of bits. The non-contextuality comes from the desire to have these bits take consistent values throughout the computation, regardless of when and how we might hypothetically take a peek at their values (the context in which we measure the bits).
We have long known that not all of quantum physics can be described by a non-contextual hidden-variable theory, and there are experimental tests that can be used to prove that quantum systems are contextual and so evade any possible classical description. In their study, Howard and colleagues show that what makes magic states special is precisely their contextuality. Specifically, they find that magic states possess exactly the properties needed to prove that quantum physics is contextual using an experimental test that relies only on stabilizer operations. That is, the authors demonstrate that this particular measurable aspect of quantum weirdness — contextuality — is the source of a quantum computer's power.
A few curious details remain unresolved. First, there are some subtleties that limit what these results can say about quantum bits — the most elementary quantum systems — as opposed to larger quantum systems. The limitations could simply be a vagary of the proof technique used by the authors, or could be a hint of something deeper. There also remain some unanswered questions regarding the power of states with vanishingly small amounts of magic. And finally, does contextuality power other quantum-computing architectures that supplement stabilizer operations in other ways than supplying magic states, such as those designed around quantum measurements8? Further refinements of the possible tests of contextuality to the most general situations could clarify these outstanding issues.
Knowing that contextuality supplies the magic for quantum computers is much more than a satisfying connection. This finding also promises to help researchers design better architectures for quantum machines. In many of the most sophisticated models for a potential quantum computer, just manipulating magic states into a usable form consumes most of the processor time. New architectures that are thriftier in their use of this contextuality resource may be much easier to build.