Thani Shamsi
A Quantum (Computing) Leap - From Vision to Reality (2014)
This essay was written as part of a working group on quantum computing for a technology conference.
Imagine a controversial topic that is the subject of heated discussions in both the business and academic worlds and that is also confusing. Quantum computing is such a topic. It promises to be the next great revolution after the rise of the microprocessor. In theory, it has the potential to exponentially accelerate processing speeds when compared to classical computers, for example, rendering current cryptography irrelevant.
As revolutionary as the theory is, in practice the developments are more evolutionary and nowhere near being able to meet the ambitious visions. Yet, investors like Amazon's Jeff Bezos and customers like Google and Lockheed Martin are investing millions in quantum computers. Are we at the tipping point of the quantum computing era?
The world has been looking with excitement and curiosity at the developments in quantum computing since at least 1994. Twenty years ago, Peter Shor formulated a quantum algorithm that – run on a large quantum computer – would be able to break public-key cryptography. That includes, for example, the widely used RSA scheme, which is based on the inability of classical computers to factor large numbers into their prime factors in a reasonable amount of time. While factoring a number like 77 into its prime factors 7 and 11, the time classical computers need to factor larger numbers grows exponentially. Shor's algorithm proves that prime factorization is (quantum) computationally feasible, since it requires significantly less time.
Another aspect that makes quantum computing even more interesting is the fact that there is a hard limit to Moore's Law, considering transistor size reaching atomic level and the necessary amount of electrical power needed. The exponential growth of the number of transistors on a chip is the fundamental backbone of the ICT industry, fueling its exceptional growth and technological progress. Quantum computing, with its ability to exponentially accelerate certain computations, could be the next logical step to this revolution. But what differentiates a quantum computer from a classical transistor computer and where does the potential exponential speedup come from?
Much of the confusion surrounding quantum computing might be based on the theory behind it: Quantum mechanics. The nanoscopic world on the quantum level is fundamentally different from the macroscopic world we know. Very strange and counterintuitive phenomena occur on that level – which quantum computers exploit to perform computations. While we know a world of yes and no, true or false, on and off, 0 and 1, a world where information is stored in bits, quantum computers work with quantum bits, so-called "qubits." Surprisingly, a qubit can be in the state 0, 1, or 0 and 1 both at the same time, a state called superposition. Superposition and quantum entanglement, a phenomenon describing a strange non-local correlation between particles, are both required to enable the desired quantum speedup.
A functional quantum computer would need a large number of entangled qubits in a stable and coherent system to actually perform reliable computations and outperform today's supercomputers.
Quantum computing set out to change the game, in theory. So where do we stand today? The most straightforward benchmark would be to look at the ability of quantum computers today in prime factorization (where a quantum speedup would break public-key cryptography). The largest number researchers were able to factor – using Shor's algorithm – is 21 (prime factors 3 and 7). This result conveys two important messages. First, quantum computers are real and the theory works in practice. Second, quantum computers designed to run Shor's algorithm are merely proof-of-concepts and nowhere near breaking public-key cryptography.
The reality is that, on one hand, quantum computers are extremely hard to build and, on the other, hard to keep coherent when operated: The states of qubits are really fragile; qubits lose their "quantumness" through minimal external disturbances. They need to be isolated from the environment and any influences such as temperature changes and magnetic fields, which cause computations to fail completely. Compared to classical hardware, quantum hardware is extremely unreliable, making a lot of error correction necessary. Improving the reliability of quantum computer architectures requires massive engineering efforts and is directly linked to their scalability – the major issue of building a useful quantum computer. Right now, the most promising developments are superconducting solid-state technologies where new levels of reliability have been reached.
But as we know, reality is weird, surprising, and different in the quantum world. Not only did Chinese researchers claim to have factored the number 143 (prime factors 11 and 13) with a quantum algorithm, but the first quantum computers have already been shipped. D-Wave Systems, a company that was recently featured in Time magazine with a cover story, claims to have built the world's first commercial quantum computer. And their customers are well-known: Lockheed Martin, Google, and NASA bought and installed D-Wave's machines for an estimated $10 million each.
Unlike the aforementioned approach of building a universal gate-model quantum computer able to run Shor's algorithm and more similar to classical transistor computers running sequential operations, the Chinese researchers and D-Wave are using something called adiabatic quantum computing (AQC), which is based on quantum annealing, only exploiting a subset of quantum mechanics. AQCs are not able to run Shor's algorithm, but might be useful in the future for solving extremely complex optimization problems. Their functionality is best described by imagining a landscape with mountains and valleys. The lowest point in this landscape stands for the best solution, the highest for the worst. Rather than a classical algorithm, which would traverse every point on the landscape one-by-one, an AQC considers all possible solutions simultaneously and returns a set of good solutions.
The advantage of D-Wave's AQC architecture is its higher tolerance of error prone qubits, compared to the digital gate-model quantum computing architecture. D-Wave System's latest machine, the D-Wave Two (a quantum annealing co-processor), operates with 512 qubits and a model with 1,024 qubits is under development. And even if the D-Wave machines are not able to run Shor's algorithm, their application could be revolutionary for industries dealing with extremely complex optimization problems. Although researchers found that at least eight out of 512 qubits were actually entangled, a quantum speedup has yet to be proven.
Considering all of these facets, which scenarios could we possibly face regarding the future of quantum computing?
One thing is for sure: the future of quantum computing remains in a superposition. It can be both success and failure at the same time, we just don't know yet. The theory of quantum computing holds great potential; we just don't know yet if current approaches, gate-model and adiabatic quantum computers, will lead to the desired quantum speedup. The developments don't seem to be disruptive enough to indicate broad commercialization in the near future – if anything in the next two to three decades. No matter what approach succeeds, a model of "quantum-computing-as-a-service" for dealing with complex optimization problems, artificial intelligence or even big data would be extremely intriguing. Basically, the question on the future of quantum computing is: Revolution or evolution?