At Avanade we are always thinking about what new technologies can do for our clients. One such promising technology is quantum computing. What is it and how can it add value for our customers?

Quantum computers are still very actively researched, and it will take some years before the power of quantum will be useful to the world. Nonetheless, we already take steps to understand what the fuzz is all about. What is this new technology? What can we do with it and where can we apply it? And most importantly, can this technology add value for our customers given their business challenges?

So, quantum computing; you may have heard a few rumors about quantum computers, but which of these are true? In this post I want to dispel some of the existing myths on quantum computers and discuss what practical problems may be solved in the future by quantum computers. I write, ‘may be solved’, because a lot is known about quantum computers theoretically, but how they will perform in practice is still an open question.

How does quantum computing work?

The fact is that quantum computers are incredibly hard to make. To explain why, I must first explain how a quantum computer works. While a classical computer works with bits, a quantum computer uses qubits. Qubits utilize a property of quantum mechanics called superposition. This means that a qubit can have a value of 0 and 1 at the same time. However, when you measure the value of the qubit, you will get either 0 or 1, with a certain probability. The probability of measuring either value depends on the operations you performed on the qubit, but it generally also depends on what you do to other qubits. These weird properties of quantum physics are what quantum computers utilize to do their calculations. To do them reliably, it is essential that these qubits do not change without instructions of the quantum computer. In contrast to classical bits, which practically never flip between 0 and 1 spontaneously, qubits are incredibly sensitive to fluctuations in the environment.The slightest perturbation to one of your qubits will mess up your entire calculation.

Current quantum computers can for the most part be divided in 3 classes:

1. Universal gate quantum computers. These are real quantum computers capable of running quantum algorithms. The ones that exist now are development machines that operate with a small number of qubits. These computers are small (50 qubits) and can only run for a short duration (0.090 millisecond) such that you can’t run practical calculations on it, however they are great for development and testing of simple circuits.
2. Quantum annealers. You may also have heard about 1000+ qubit computers. However, these are not real quantum computers. Just like a GPU is optimized to do image processing, quantum annealers have a special kind of processor that are great for solving certain optimization problems. Quantum annealers do have qubits and real quantum effects, but you don’t have the control over the qubits that a universal gate quantum computer provides. Hence quantum annealers can’t run just any quantum algorithm.
3. Quantum simulators. These are classical computers that simulate quantum effects and can therefore run quantum algorithms just like a universal gate quantum computer. Because keeping track of all states grows exponentially with the number of qubits, simulators also operate with a small number of qubits. Simulators are also great for developing and testing.

Ultimately, the goal is to have a computer that can run a quantum algorithm like a classical computer can run classical algorithms. This means it can run any kind of quantum algorithm, have enough qubits available and is able to isolate the qubits from external perturbations.

Quantum algorithms are very different from classical ones. They don’t just take input and spit out the correct answer. They are inherently stochastic, so the goal of any quantum algorithm is to give the answer (or a close enough answer) with the highest probability possible.

This immediately poses a problem: how do you know that the answer of the quantum computer is correct? That depends on the problem. For prime factorization (determining the prime numbers that divide a certain number), you can just check whether the returned value divides your input number with a classical computer. However, for optimization problems (which quantum computers are promising to solve faster than classical computers), there is no way to know the quantum computer gives the best answer without comparing it to all possibilities. The trick in that case is to prepare many instances of the final state and measure them. The answer that pops up the most times is a (nearly) optimal solution, provided the quantum algorithm was implemented correctly.

Misconceptions about quantum computers

Myth 1: Quantum computers will replace current, classical computers.

Quantum computers work completely different from classical computers, so it’s not possible to port applications for current processors to quantum processors. Quantum computing is a promising technology for a certain type of problems that are difficult to solve on classical computers. For example, all known classical algorithms for prime factorization scale exponentially with the size of the number. The quantum algorithm (Shor’s algorithm) scales polynomially, which means that for a large enough number, the quantum algorithm will be faster.

Myth 2: Real practical quantum computers are already here.

You may have seen companies claim they made a quantum computer. You may also know that quantum computers will be able to break the encryption we use to secure our data and internet traffic. So why are our data and the internet still secure? The reason is because these computers are either not real quantum computers, or they are too small to be practical. Developing these computers are important steps to take, but they are not the computers we can use to solve problems that are currently too difficult for classical computers.

What can quantum computers do for us?

It will still take some years until quantum computers will be able to solve the problems we currently cannot. However, when they are, we expect them to yield tremendous speed-ups for optimization problems that will benefit numerous markets. We also envision a big role for quantum computers in AI, e.g. machine learning.

It is impossible to explain everything there is to know about quantum computing in just one or two blogposts. But I hope you at least have an impression of what quantum computers will and will not be able to do. You may have more questions now than before you started reading, and if that’s the case: great! Our Digital Innovation Studio is happy to explore what quantum computing can do for you and your business. Just send me a message if you have any questions.

In part 2 of this blog, we will explain how to harness the power of quantum computers to solve an optimization problem.