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