Google introduced a new quantum processor

Google Inc. introduced the 72-qubit quantum processor Bristlecone. Using this processor, the Google Quantum AI lab unit responsible for developing a quantum computer will test system errors and the scalability of the technology, as well as the scope of quantum simulation, optimization, and machine learning “to solve real-world problems, ” the company writes on a blog .

Квантовый процессор Google Bristlecone

The new 72-qubit Google Bristlecone quantum processor is built on the principle that allowed in the previous 9-qubit processor to show a low error rate when reading data (1%), when operating a one-qubit gate - 0.1% and when operating a two-qubit gate - 0.6 %, which, according to Google, was the best result of the company. Before using the new processor in the work, it is important to understand its capabilities: the team created a tool that checks it for errors by solving identical problems on a quantum processor and in a classic simulation. With a low number of errors, "quantum superiority" can be achieved.

Прогноз Google: График показывает снижение количества ошибок при увеличении количества кубитов в процессоре. Начиная с количества кубитов 10 в шестой степени количество ошибок будет находиться ниже уровня 10 в минус третьей степени, что сделает приемлемым использование квантовых вычислений
Google Forecast: The dependence of the number of errors on the number of qubits in the processor.
Quantum computers use quantum superposition. and quantum entanglement for data transmission and processing. One of the main tasks of quantum computers will be the strengthening of artificial intelligence . The qubits of a quantum processor are quantum analogs of bits. Two adjacent qubits have four states - both on, both off, on / off and off / on, each of them has a weight or “amplitude” that can play the role of a neuron; the third qubit in such a system allows eight neurons to be represented, and the fourth - sixteen. A change in the state of four qubits results in processing sixteen neurons at a time, while a classic computer would process these numbers one at a time.

One of the problems with a quantum computer is the number of errors that occur when computing, reading, and writing information into qubits. In June 2016, Google researchers built a 9-qubit processor, which showed high reliability. They were able to scale this development by March 2018, increasing the number of qubits to 72. In the processor, the qubits are located in two 6x6 layers on top of each other. The Google Quantum AI lab is testing development.

Схема внутреннего строения квантового процессора Google Bristlecone состоит из 72 кубитов, расположенных в шахматном порядке в два слоя, каждый 6 на 6 кубитов. Изображены в виде символов X
The Bristlecone quantum processor consists of 72 qubits, shown in the diagram (on the right) in the form of an “X”, where the points of contact between the ends of the symbol represent the connection of the qubit with its nearest “neighbors”

Currently, a number of research teams are involved in quantum computers, including IBM. In March 2017, the company announced the launch of the IBM Q project, and by June introduced two processors: 16-qubit for work in the scientific field and 17-qubit for commercial use. In 2017, IBM Research developed a 49-qubit processor.

In July 2017, a team of Russian and American scientists from Harvard University, led by co-founder of the Russian Quantum Center (RCC) Mikhail Lukin, announced the creation of a 51-qubit quantum computer.

In Russia, in March 2018, between Vnesheconombank, VEB Innovations, the Foundation for Advanced Studies (FPI), Lomonosov Moscow State University and the Digital Economy Autonomous Non-Commercial Organization An agreement was signed to develop a 50-qubit quantum computer.