close

Steps towards Quantum Computing

First steps towards Quantum Computing

I’m going to start by asking you to keep a strong hold on to your common sense – that is, until I ask you to let it go.

There’s one thing that your IT team will definitely agree that is common sense, and that’s using encryption technology – for example an RSA asymmetric public/private key for your important data and emails. This is considered to be pretty strong encryption, as it is based on the mathematical problem of prime factorization. Creating the encryption essentially involves taking two large integers and multiplying them. The answer is the public key, which is published. Your private key, which you use to read the encrypted data, is one of those two prime numbers.

Therefore, you’d think that if you have the solution to the multiplication, it would be easy to figure out the private key. Well, it is possible, but calculating the correct prime factors of a number with 617 decimal digits (2,048-bit key) would take a regular PC around six quadrillion years. A supercomputer could find the solution much faster, in a relatively speedy 21 billion years, or one and a half times the age of the universe.

But what if a totally new type of computing came along to exponentially accelerate the speed of breaking the code? This is eventually going to happen, thanks to quantum computing. Quantum computers are based on a type of science entirely unlike anything else, as they allow us to test – essentially all possibilities simultaneously. How does this work? Well, here’s where you need to forget common sense for a while.

Imagine you are skiing down a hill, and there’s a tree in front of you. You need to decide whether to go to right or left, as going through the middle is not generally a smart option. With quantum computing you don’t have to do one or the other, you CAN do both and go left and right – this is the strange but amazing duality of quantum physics.

Here’s the science bit

To understand how this is possible, we’re going to look at the physics of light – which Christiaan Huygens suggested moved in waves (1670). This was proven 100 years later by Thomas Young, who used a double slit experiment to demonstrate that light produces diffraction and interference (observed via an alternating sequence of bright and dark spots behind the double-slit), essentially just like the patterns made when you throw a stone into a pond and watch the waves.

Some 100 years later, in 1905, Einstein shared his quantum theory of light – the idea that light moves as tiny packages of energy, or quanta. This starts getting into the duality of light – it behaves like both waves and particles. This theory was taken a step further by Louis de Broglie in the 1920s. He postulated that, if light behaves like particles, then particles can also behave like waves. This astonishing idea was eventually proven using a similar experiment to Thomas Young – instead of light, physicists fired electrons at the slits. The resulting diffraction pattern was the same as when light was fired at it. De Broglie could also determine the wavelength of the electrons – something so small it can only be observed on a microscopic scale.

Erwin Schrödinger later took this even further and defined an equation that describes the evolution over time of physical systems affected by quantum mechanics – this is essentially a wave equation that determines the probabilities of the physical quantities of the system. Quantum theory is not deterministic like traditional physics, because in quantum theory, everything has a probability. The solutions to the equation are waves of probabilities. With very significant consequences.

This is where quantum physics starts to get rather strange

We’re going to talk about quantum tunneling. If particles are moving in smooth waves towards a wall, then the wave function of any particle that hits the wall will drop smoothly rather than abruptly, leaving a small tail on the other side – and leading to a small but finite probability that it went through the wall. This strange (and rare) event is called quantum tunneling.

Also strange is the concept of superposition. When you combine two different waves (wave a and wave b), then the result is also a solution of the wave equation. But as we are talking about probability waves, in a quantum sense the system is then both in state a and b at the same time. This superposition capability is one of the key enablers of quantum theory and quantum computation.

The problem is you can’t observe or measure a system in a quantum state. This is because any interaction with the real world causes a system to leave the quantum state and collapse into a classical (observable) one. When we measure, we essentially change it – forcing it into a classic state. This was what Schrödinger explained with his, I hasten to emphasize, thought experiment related to a cat in a box. The theoretical cat was in its box along with a flask of poison that would break open in the case of a certain radioactive decay that occurs with the probability of 50%. In its quantum state, the cat could be described as being simultaneously in two states: alive and dead. When you open the box to check on the status of the cat, the quantum state is destroyed. Then the cat is either dead or alive – and certainly not both. But until then…

Another peculiar effect is that when particles created together are in the same quantum state, they are bound or entangled with one another through a shared wave function. Even if you separate them, they stay entangled. This entanglement means that if you change one, the other will reflect the change. And this can have very fast effects over distances. Over any distance.

This combination of superposition and entanglement catapults us forwards in terms of computer processing speeds, not as a result of accelerating the computation, but rather by computing all possibilities at the same time. A quantum bit is not in the state of zero or 1 – it can be both simultaneously, which means everything happens at the same time. Consequently, you can have many bits that represent each possibility at once – which essentially gives you parallel computing on steroids. The challenge is to keep the computer in that state long enough to do the calculation. This is obviously very desirable, just really hard to build. But when we figure it out, then that will be the end of certain types of encryption used today.

We’re obviously some time away from being able to deliver true quantum computing, so RSA encryption and its private keys are safe for now. But, what if the focus was just on delivering speed rather than true quantum computing? Fujitsu has actually succeeded in making this work by taking inspiration from quantum mechanics and implementing this in hardware that uses quantum-inspired tricks to deliver extremely fast processing speeds on (more or less) normal digital circuits.

This result is the Fujitsu Digital Annealer, which is designed to accelerate the processing of computational problems in the area of combinatorial optimization. That is, essentially finding the best possible solution from a finite set of possibilities – for example the shortest or cheapest trips, the most effective routing, traffic management or scheduling. It does this by mimicking quantum tunneling and undertaking a simultaneous search and evaluation of candidate states in the optimization process – offering some of the benefits of quantum computing. The advantage is that it is far easier to build and run using standard technology and needs much less energy.

For example, look at the classic travelling salesman problem, which requires you to identify the shortest possible connecting path between all defined stops. This is handled by the Fujitsu Digital Annealer some 17,000 times faster than classical simulated annealing, making it highly suitable for deployments such as optimizing logistics. By contrast, Moore’s Law would take 14 chip generations, some 25 years of development, to achieve the same, so this is a really significant step forward.

This new approach has potentially significant benefits to offer to multiple industries – from disaster management, apps for the Internet of Things, and from traffic optimization to molecular design and made-to-measure drugs. And all this is just around the corner. Fujitsu will commercialize this new technology starting in 2018 – so now it is time to recapture your common sense and imagine what quantum-inspired computing could do for your business.

Watch my keynote on Quantum Computing from Fujitsu Forum 2017 in Munich

Tags: , , , , , ,

No Comments

Leave a reply

Post your comment
Enter your name
Your e-mail address

Before you submit your comment you must solve the following arithmetic function! * Time limit is exhausted. Please reload CAPTCHA.

Story Page