The Lure of Quantum Computing – Part 2

Spike Narayan-

Spike Narayan

Spike Narayan is a seasoned hi-tech executive managing exploratory research in science and technology at IBM

In Part 1 of this two-part series, we discussed why quantum computing (QC) is such a big deal and how there is a significant effort and investment across the globe to move the field forward. Meaningful advances are already visible, and the impact of quantum computing is closer than what many (not just skeptics) predicted even 2 years ago.

In this article let us step into the wonderful world of quantum computing and see what makes this technology so unique and potentially so powerful and what are the challenges lie ahead.

In part 1 we saw that a mere 160 qubits (equivalent to 1048 conventional bits) were needed to solve some potentially intractable problems and we also saw that IBM already has a 127 qubit quantum computer. So, are we nearly done? Not quite. The reasons are both fascinating and daunting. There are at least two dimensions to this. The first is, we need perfect hardware which means some number of “perfect” qubits. The challenges I will discuss here have to do with getting to that level of perfection, or at least close to it, with less than perfect quantum hardware that we can build today. A second dimension of the challenge has to do with ease of use for engineers and scientists who are not trained in quantum physics. Pervasive use and massive adoption is what will eventually drive innovation and greatly accelerate this computational field. We will discuss these two challenges.

To discuss the first dimension, what do we mean by perfect hardware? It has to do with how quantum computers work today. As we saw in Part 1 of this two-part article there are a few different types of QC technologies and the definition of a perfect bit will vary some depending on the technology of choice. In this article, I will talk about the superconducting qubit technology that IBM has chosen to pursue.

These qubits have three magical powers that conventional compute bits do not have. They are Superposition, Interference, and Entanglement. There is a great video from IBM’s Research Director that introduces these three fundamental and powerful ideas. You can also use your favorite search engine to learn more about these three properties of qubits. It is still worthwhile to note a few things here.

In today’s conventional computing architecture like the one used in all computers today, the information is stored in bits, which can be either a “0” or “1”. In quantum computing, the information is stored in qubits, which can exist as a “0” or “1” or a combination of both. This ability to exist in both states is referred to as the superposition of states and enables a quantum computer to perform multiple calculations in a massively parallel fashion. This is a very powerful capability and enables the other two properties, namely, interference and entanglement of qubits to do the computation. Why is this ability to do several compute operations in parallel so key?

One begins to realize this potential when the number of variables or permutations becomes unthinkably large and is precisely the reason why conventional computers run out of steam. The fact that a qubit can exist in a combination of the “0” and the “1” state means that 2 qubits can exist in 4 states simultaneously (00, 01, 10, 11) and 3 in 8 states (000, 001, 011, 100, 110, 101, 010, 111) so n qubits can support 2n states in parallel. This is why exponential speed up is possible.

Most people, including engineers and scientists, cannot comprehend how large 2n can become very quickly. I suggest you try to play this little game at home and you will be stunned at the power of exponential speed up. Take a chess board with 8×8 squares for a total of 64 squares. Place one grain of uncooked basmati rice in the first square. Place 2 grains in the next square and continue doubling the number of grains in each subsequent square. At the end of the first row (square number 8) you will need 128 grains and when you get to the end of the 2nd row you will need 32,768 grains that will weigh 1.5 lbs. By the time you reach the end of the third row, you will need over 8 million grains weighing nearly 400 lbs (more than all the rice in your neighborhood grocery store). By the end of the 4th row (square number 32) you will need 100,000 lbs (2 billion grains) of rice. I think you get the idea of what exponential scale-up means. Using the numbers from your 32nd square on the chess board, means that with 32 qubits you can explore 2 billion combinations in parallel. This is the power of a quantum computer.

The real challenge of getting to perfect qubits is putting these qubits in a superposition state for long enough to perform the computations. In superconducting qubits the superposition state is very vulnerable to “noise”. By noise I mean all kinds of noise including mechanical (vibration), electrical, magnetic, and thermal to name some. These computers are built with vibration isolation and shielded from electromagnetic fields and operated at near zero degrees Kelvin. Typical cryogenic fridges that house the quantum chips operate at < 0.020 degrees Kelvin (~ -459.6o F) which is colder than deep space (-455o F). There is now a large effort underway in the US and elsewhere to address this noise issue and is called the NISQ (Noisy Intermediate-State Quantum) computing initiative. The main goal is to “do more with less” ie. given that we are stuck with noisy qubits in the near term how can we do more complex computations despite the noise-related errors.

Let us now move on to the second dimension of the challenge which is the usability of quantum computers by scientists and engineers not trained in quantum physics. We have a wonderful analogy to explain this challenge. When conventional computers came on the scene a few decades ago, only a small set of folks could work with computers and they were those who could code in machine language. That was the only way to program a computer. It was not until 1957 when a high-level language called Fortran (Formula Translator) became commercially available that suddenly all engineers and scientists could use computers broadly to solve problems. That started the explosive growth and need for computers. I, myself, as a materials scientist, could use Fortran for my doctoral thesis to computationally model mass transport phenomena.

These days even higher languages like Python and Java enable a broader community to use computers with absolutely no training in computer science. The quantum computing community is exactly at this “1957” moment today. IBM has introduced Qiskit to enable those not trained in quantum physics to access and use QCs and has set the stage for mass adoption. This has opened the door for thousands to begin experimenting with QCs and begin innovating in this space.

In summary, quantum computing is becoming very real very fast. Even though the challenges are still formidable both in the quantum hardware and in the software to use QCs, the progress is very rapid and we can confidently say that the era of QC is upon us.