Quantum computing : 15 minutes to understand everything

This article is part of our spe­cial issue « Quan­tum : the second revo­lu­tion unfolds ». Read it here

Quan­tum com­pu­ters work with quan­tum bits, or qubits, which, unlike stan­dard com­pu­ter bits that can have a value of either 0 or 1, can be in a super­po­si­tion of both 0 and 1. This cha­rac­te­ris­tic means that quan­tum com­pu­ters could be much fas­ter than conven­tio­nal com­pu­ters for cer­tain tasks. They could also be used to solve cer­tain pro­blems that a conven­tio­nal com­pu­ter cannot.

“A quan­tum com­pu­ter will mani­pu­late many qubits in a mas­si­ve­ly super­po­sed state : 0000 plus 1111, for example, » explains Lan­dry Bre­theau. « In this ‘entan­gled’ state, seve­ral cal­cu­la­tions can be car­ried out in paral­lel. A concrete example : ima­gine that the cal­cu­la­tion, the pro­blem, is to get out of a maze. How would you go about that ? A human being or a com­pu­ter pro­gramme will try out dif­ferent pos­sible tra­jec­to­ries one by one. Each time, a dead end is rea­ched, they will then retrace their steps. In this way, they will test all the paths until they emerge from the maze. Since a quan­tum sys­tem can be in a super­po­si­tion of states, that is, it can be in seve­ral places at the same time, it can the­re­fore try to explore all the dif­ferent paths in paral­lel and escape from the maze more quickly. »

Qubits can be built from dif­ferent plat­forms or mate­rial buil­ding blocks, such as super­con­duc­ting qubits, ele­men­ta­ry par­ticles or trap­ped ions. Other methods in the pipe­line are pho­to­nic quan­tum pro­ces­sors that use light. « We use the term quan­tum com­pu­ter, but it would be more accu­rate to talk about quan­tum pro­ces­sor, because the entire cal­cu­la­tion can­not be imple­men­ted on a quan­tum com­pu­ter, only a frac­tion of it, » explains Loïc Hen­riet. « We will always need a clas­si­cal pro­ces­sor to coor­di­nate all the steps of the calculation. »

At present, a quan­tum pro­ces­sor is still in the expe­ri­men­tal stage : for one, it takes up a lot of space – for example, the one Löic Hen­riet’s team is wor­king on sits in a large box mea­su­ring 3 metres by 2 metres by 2 metres. A very high vacuum, of 10^-11 mbar, is also requi­red to place the qubits in well-defi­ned posi­tions. This cor­res­ponds to around the same pres­sure as can be found on the sur­face of the Moon.

The sophis­ti­ca­ted optics nee­ded to control the qubits consist of lasers, lenses and mir­rors. To coor­di­nate the ope­ra­tion of each of these dif­ferent pieces of equip­ment (which make up the hard­ware) and to syn­chro­nise them, on-board soft­ware is nee­ded. This soft­ware is the quan­tum pro­ces­sor’s ope­ra­ting system.

There are many areas in which a quan­tum com­pu­ter could prove more use­ful than a conven­tio­nal com­pu­ter, either in terms of com­pu­ting time or the qua­li­ty of the results obtai­ned. The best-known example is Shor’s algo­rithm, which enables large num­bers to be effi­cient­ly fac­to­ri­sed into prime fac­tors, for appli­ca­tions in cryp­to­gra­phy and com­pu­ter secu­ri­ty, for example. Quan­tum com­pu­ters will also be very good at making use of spe­cial algo­rithms to solve com­plex opti­mi­sa­tion pro­blems, such as those lin­ked to sche­du­ling, rou­ting and logis­tics. These pro­blems involve fin­ding the opti­mum solu­tion from a large num­ber of pos­si­bi­li­ties – the most famous being the « tra­vel­ling sales­man » pro­blem, which involves fin­ding the shor­test pos­sible route bet­ween seve­ral cities. Deli­ve­ry and logis­tics com­pa­nies will cer­tain­ly want to adopt quan­tum tech­no­lo­gy for this type of application.

Pro­blems lin­ked to the reac­ti­vi­ty of mole­cules will also bene­fit. « There is a lot of research going on in this area, » explains Löic Hen­riet. « With a quan­tum pro­ces­sor, we will be able to car­ry out much more effi­cient cal­cu­la­tions to deter­mine the reac­ti­vi­ty of cer­tain pro­teins, for example, which will have enor­mous appli­ca­tions for the phar­ma­ceu­ti­cal indus­try and the syn­the­sis of new drugs. We will also be able to cal­cu­late the pro­per­ties of new mate­rials of inter­est for many tech­no­lo­gi­cal fields. »

Machine lear­ning and arti­fi­cial intel­li­gence are also impor­tant appli­ca­tion areas since quan­tum com­pu­ters should be able to improve machine lear­ning algo­rithms – poten­tial­ly dra­ma­ti­cal­ly – by pro­vi­ding fas­ter and more effi­cient opti­mi­sa­tion rou­tines or by explo­ring new models and archi­tec­tures. This could be a mas­sive new mar­ket, but it will depend on buil­ding prac­ti­cal quan­tum com­pu­ters on a large scale and deve­lo­ping algo­rithms and appli­ca­tions that can take advan­tage of their unique capabilities.

For a quan­tum com­pu­ter to work, it must be « uni­ver­sal », that is, it must be capable of cor­rec­ting the errors cau­sed by the imper­fect nature of cur­rent hard­ware, which pre­vents a cal­cu­la­tion from being com­pu­ted com­ple­te­ly. The main cause of these errors is the deco­he­rence of the qubits them­selves, which des­troys the quan­tum cha­rac­ter of the qubits and returns them to the state of clas­si­cal bits. Deco­he­rence is cau­sed by the inter­ac­tion of the qubits with their envi­ron­ment. The real chal­lenge is the­re­fore to iso­late the sys­tem effec­ti­ve­ly. To do this, the qubits gene­ral­ly have to ope­rate at a tem­pe­ra­ture close to 0 K and be shiel­ded from each other and the envi­ron­ment. In addi­tion to this, quan­tum error cor­rec­tion (QEC) tech­niques can be used to achieve « fault-tole­rant quan­tum com­pu­ting ». These tech­niques involve using a large num­ber of qubits to create a « logi­cal qubit » that is much less prone to errors. 

Accor­ding to experts, a real quan­tum « advan­tage » or « supre­ma­cy » will only be achie­ved when quan­tum com­pu­ters ope­rate with a mil­lion qubits. And since the cur­rent record is still less than 100 qubits, there is still a long way to go.

While, in theo­ry, there is nothing pre­ven­ting the crea­tion of large-scale quan­tum com­pu­ters, a num­ber of major engi­nee­ring pro­blems need to be resol­ved first. « Com­pa­nies see quan­tum com­pu­ting as a stra­te­gic invest­ment and don’t want to miss out, » explains Löic Hen­riet. « It’s no lon­ger a ques­tion of if, but rather a ques­tion of when the quan­tum pro­ces­sor will become an inte­gral part of IT solutions.

We are cur­rent­ly at the cusp of a tech­no­lo­gi­cal revo­lu­tion and, in France, we have the resources to be at the heart of this, both at aca­de­mic level and at the level of com­pa­nies and start-ups. Of course, end-cus­to­mers and busi­nesses also need to be on board.

Over the last five years or so, we’ve seen a real increase in interest

That said, a quan­tum com­pu­ter, uni­ver­sal or not, will not replace your per­so­nal com­pu­ter or smart­phone any time soon, and the first cus­to­mers will cer­tain­ly be govern­ments and major cor­po­ra­tions rather than the gene­ral public, adds Lan­dry Bre­theau. « Scien­tists them­selves will also be the first users, which is why the world of quan­tum com­pu­ting is of inter­est to so many dis­ci­plines : che­mis­try, mate­rials science, bio­lo­gy and phy­sics. Each of these dis­ci­plines will deve­lop its own algo­rithms to solve very spe­ci­fic problems.

« Over the last five years or so, we’ve seen a real increase in inter­est, » he says. “Some are cal­ling it a ‘quan­tum boom’, with the crea­tion of nume­rous start-ups and major fund-rai­sing campaigns.”

Psi­Quan­tum and IonQ, which have rai­sed €600 mil­lion and €400 mil­lion res­pec­ti­ve­ly, are two outs­tan­ding examples. « In France, we have the Quan­tum Plan, announ­ced by Pre­sident Macron in ear­ly 2021, and the start-up with the most wind in its sails from a finan­cial point of view in France at the moment is Pas­qal, which has just rai­sed €100 million.

Although we pro­ba­bly won’t suc­ceed in buil­ding a ful­ly ope­ra­tio­nal and fault-tole­rant com­pu­ter in the next ten years, we can be sure that we will make unthought-of dis­co­ve­ries along the way that will be poten­tial­ly use­ful and per­haps even change the face of socie­ty, just as clas­si­cal com­pu­ting has done over the last 50 years and, more recent­ly, the Internet.

« It’s a great time to be wor­king in this field, » says Lan­dry Bre­theau. « The­re’s a lot of exci­te­ment and the sec­tor is evol­ving very quickly.”

Isabelle Dumé