We open the hood of IBM’s quantum computer, which finally begins to correct its errors.
Mathilde Fontez, editor-in-chief of the scientific magazine Epsiloon, talks to us today about the quantum computer. We’ve been waiting for this potentially revolutionary technology for a long time.
franceinfo: It seems that one of these quantum machines has turned a corner here?
Mathilde Fontez: Yes, the idea of a quantum computer is not new: it was in the 1980s that the famous American physicist, Richard Feynman, had launched, in a conference, that we could exploit the properties matter at the scale of the atom, to perform computer calculations. Because the particles have a specific behavior: for example, they can be in a state of superposition, or be linked in pairs, by a phenomenon called entanglement. And by exploiting that, by making qubits instead of bits, we could calculate faster.
It triggered a lot of work, sketches of quantum computers exist today. Most of the electronics giants have entered the race, Microsoft, IBM, Google too. And today, IBM is making the event, announcing that its quantum computer finally manages to mitigate its errors.
Errors are the main technical obstacle of the quantum computer?
It’s considered the biggest problem, yes. Because quantum systems are difficult to master – they are elusive. Quantum machines therefore introduce uncertainty at each calculation cycle. Basically, in a normal computer, a bit remains a 1 or a 0 all the time. But in a quantum computer, these values can fluctuate, which gives more computing power, but which is also very difficult to master.
IBM therefore pushed another approach: quite simply, not correcting the errors, but mitigating them, at each calculation cycle. They did the test with their 127 qubit quantum processor, it’s called Eagle. And they showed that in the end, the result is good, and the calculation speed is higher than with a conventional computer.
Could quantum computers soon come out of the labs?
This is even already the case, following this IBM approach: no need to make a perfect quantum computer, it is ultimately enough to control the error level, the noise level, as physicists say. There are already quantum computers at work, in medical research for example, labs are starting to use them to process data.
In pharma, Merck has entered into a partnership with a quantum start-up to develop, to simulate new active ingredients, new drugs. In oil exploration (to find new deposits), in banks (to improve speculation strategies), in the automotive industry (to optimize processes). In fact, the quantum computer is already somewhat there.