Quantum computer

How does a quantum computer work?

Date:
Changed on 29/03/2021
Quantum computers are expected to be able to solve all sorts of problems, some of which are still unsolvable, in record time. But how will they go about doing this? And what is it that makes quantum computers so unique...and so sought-after?
qubits quantique
© Wikimedia.org – D-Wave Systems Inc.

Bits versus qubits

Before we take a look at the difference between a standard computer and a quantum computer, you need to understand two concepts from quantum physics:

  • Quantum superposition. Superposition is the counter-intuitive capacity quantum objects, such as electrons, have that allows them to exist in more than one “state” simultaneously. The object is literally in both states at the same time, without splitting itself across the two. Making a measurement will destroy this superposition, and it is only then that we can say whether or not the object is in the lower or higher state.
  • Measurement in quantum mechanics. You cannot measure two things which are not compatible with each other, such as the position and the speed of a molecule. As a consequence, you need to know what it is you are measuring, in order to avoid affecting the result. This is what’s known as back action: measuring an object’s quantum state could irreversibly change the state of the object.

Standard or conventional computers are programmed using bits as data units, each bit capable of storing either a 0 or a 1. But the limitations of these computers come to the fore when they are given a problem with multiple variables. In such cases, the computer has to perform a new calculation each time a change is made to a variable. Each calculation is a unique path leading towards a unique result.

Quantum computers, on the other hand, are based on the concepts outlined above. In accordance with the laws of quantum mechanics, quantum computers use qubits, which can represent a combination of 0s and 1s at the same time, following the superposition principle. Each qubit exists in multiple states of 0 or 1, simultaneously. Quantum computers use the entanglement between qubits and probabilities relating to superpositions to perform a series of operations, in such a way that certain probabilities are increased (for the correct answers) and others decreased, or even reduced to zero (for the wrong answers).

A revolution is on the way - in multiple sectors

It is this capacity to take a vast number of different paths that makes quantum computers much faster than standard computers. Quantum computers will not replace current systems, but instead will be used for incredibly complex problems where eliminating such a wide range of possibilities will help to save a considerable amount of time.

One commonly-cited example is solving certain algorithms much more quickly, through which it could be possible to factorise into prime numbers, to solve the travelling salesman problem (finding the shortest possible route), to search through databases or to simulate complex differential equations

A quantum computer would also be able to invent drugs, to simulate new materials, to solve logistical problems or to develop accurate simulations of chemical reactions which still seem a mystery to us.

We don't yet know all of the problems that could be solved through the use of quantum computing, which is something researchers are currently working on.

Technical obstacles to overcome

But before we can all reap the benefits of quantum computing, a number of obstacles still have to be overcome, the biggest of which is quantum decoherence. Quantum computers are much quicker and much more efficient than standard computers, using superposed and entangled states, which are much more sensitive to their environment than classic states. The more qubits are added to a system, the greater the number of parallel operations, which also increases the processing power. It is estimated that close to 300 perfectly tangled and superposed qubits could map all information in the universe, dating back to the Big Bang.

However, when the environment interacts with the qubits (which is needed in order for quantum measurement to take place), this uncontrollably changes their quantum states. This is what is known as quantum decoherence, and can be caused by various aspects of the environment: changes in magnetic or electrical fields, radiation from nearby hot objects or uncontrolled interactions between qubits.

Decoherence affects superposition and disturbs the quantum processing of information, leading to errors in quantum processing systems. Where a standard computer would be highly reliable, a quantum computer would make an error in every 1,000 operations (at best).

Verbatim

Information on computers is constantly contaminated with errors. But in the quantum world, the way in which measurement influences variables makes measuring these errors extremely difficult

Auteur

Zaki Leghtas

Poste

researcher - Quantic project team (Inria Paris)

Challenges for the future

But how can these obstacles be overcome? In order for a quantum computer to successfully perform viable calculations and to revolutionise a whole range of sectors, multiple qubits have to be produced, controlled and measured; quantum logic gates have to be made; and algorithms that would benefit from quantum acceleration have to be developed.

Video on the algorithmic interests of a quantum computer © Collège de France in partnership with Inria (in French).

Qubits can only retain their quantum properties for a limited period of time before errors begin to affect the processing mechanism. This is what is known as coherence length. In order to reduce the risk of errors in calculations performed by a quantum computer, the coherence length of the qubits has to be long enough to calculate mathematical problems.

Researchers are currently working to develop algorithms for reducing errors in order to increase this coherence length, using a quantum error correction code (the first of which was developed by Peter Shor). This can be used to encode a logical qubit into several physical qubits, in such a way that errors become computationally tractable. Logic gates also have to be produced in order to ensure operations run more smoothly.

In any case, a general purpose quantum computer would require hundreds of millions of qubits connected in a coherent way. The few quantum machines that have been developed are not yet capable of managing the requisite number of qubits that would allow them to move up to the next level. “Based on the available information, it is still too early to predict when a sufficiently large quantum computer could be produced”, explained The National Academies of Sciences, Engineering and Medicine in 2019 in a report entitled Quantum Computing: Progress and Prospects.

In the meantime, in France, a quantum computing ecosystem is beginning to establish itself: scientists, manufacturers, start-ups and political decision makers have already joined the race for quantum acceleration, working on the deployment of quantum infrastructure and the means by which they will be able to take advantage of the possibilities on offer through computer architectures.

The main players in the world of quantum computing