Project Summary

Scenarios leading to the achievement of the moonshot goal in 2050

Present to 2030 2030

From now until 2030, we will conduct research and development toward the realization of an error-tolerant large-scale general-purpose optical quantum computer with an electrical signal processing system. For this purpose, we will use the method of time-domain multiplexed general-purpose optical quantum computing, which was developed by the project manager and his colleagues by applying continuous-quantity quantum teleportation, the first successful method in the world and now a world standard. The logical qubit for error tolerance is the so-called GKP qubit proposed by Gottesman, Kitaev, and Preskill. The logical qubit for error tolerance is the so-called GKP qubit proposed by Gottesman, Kitaev, and Preskill, and all the quantum gates (quantum operations) required for general-purpose quantum computation are error tolerant.

To realize our goals, we will study and develop squeezed light with sufficient bandwidth for time-domain multiplexing and sufficient level to exceed the error tolerance threshold, optical quantum computer chips for stable optical quantum computation, and superconductor photon number discriminators for arbitrary quantum state generators for generating logical qubits.

In the method of time-domain multiplexed general-purpose optical quantum computing developed by the program manager and his colleagues, clustered states are used as quantum entanglement for quantum computation. The lowest error tolerance threshold for squeezing level of squeezed light is currently considered to be achieved when GKP qubits are used as logical qubits, although a threshold of less than 10 dB has been found for GKP qubits. In this project, we aim to reduce the requirement for GKP qubits and further to decrease the threshold value for error correction with such a low error tolerance.

We are aiming to enlarge the bandwidth of the squeezed light as high as 10 THz, which is 10% of the optical carrier frequency: the band width corresponds to the clock frequency of an optical quantum computer. Currently, most of the squeezed light is generated using optical parametric oscillators with a resonator structure, but in this case, the bandwidth of the squeezed light is limited to a few hundred MHz of the resonator bandwidth. In this project, we will develop a waveguide optical parametric amplifier without any resonator structure, aiming at generating broadband squeezed light.

For the generation of GKP qubits, realization of higher-order nonlinear optical processes is required. We aim to achieve this in a measurement-induced manner using a superconductor photon number discriminator. Specifically, we generate a quantum entanglement by using squeezed light and a beam splitter, and detect a given number of photons in a part of the entanglement using a superconductor photon number discriminator to generate a GKP qubit. Here, the GKP qubit is generated at the moment when a given number of photons are detected, so in order to use it at any given time, a quantum memory is also needed. We have already started to develop a photon number discriminator, and aim to achieve an operating speed equivalent to an optical quantum computer clock frequency of 1 GHz. We have also succeeded in developing an all-optical quantum memory to some extent, and we will continue to improve its performance.

To build an error-tolerant general-purpose quantum computer means to construct a general-purpose quantum computer using only error-tolerant quantum gates (quantum operations). The codes of most logical qubits, not just GKP qubits, are called stabilizer codes, and error correction is possible only with quantum gates (Clifford gates) that perform Clifford operations (90 or 180 degree rotation on the sphere, for simplicity). In order to create an error-tolerant general-purpose quantum computer with stabilizer codes, it is necessary to perform non-Clifford operations using only Clifford gates. In order to do this, it is necessary to prepare a special state called the magic state, in which non-Clifford operations are performed on a specific state in advance, and then teleport the non-Clifford operations to an arbitrary input, using a quantum teleportation circuit consisting only of Clifford gates (quantum gate teleportation). In this method, we focus on the fact that stabilizer codes are general-purpose use for error correction of quantum information (quantum state), and by treating quantum gate errors as quantum information (quantum state) errors, we can achieve error tolerance against arbitrary quantum operation errors. In this project, we use quantum gate teleportation to make the third-order phase gates that perform non-Clifford operations error-tolerant, and realize an error-tolerant general-purpose quantum computer.

In order to put an optical quantum computer to practical use, it is necessary to convert the optical quantum computer into a chip (optical integrated circuit). For this purpose, it is necessary to reproduce the principle verification experiments that have been conducted in free space with waveguide circuits. The most important of these is quantum teleportation. This is because time-domain multiplexed general-purpose optical quantum computing refers to time-domain multiplexed quantum teleportation. We will perform quantum teleportation in waveguide circuits so that their fidelity exceeds the error tolerance threshold. Furthermore, based on this knowledge, we aim to realize a time-domain multiplexed error tolerant large-scale general-purpose optical quantum computer chip.From now until 2030, an electrical signal processing system will be required to perform measurement-induced nonlinear optical processes using photon number discriminators, and we will conduct research and development toward the realization of an error-tolerant large-scale general-purpose optical quantum computer with this system.

from 2030 to 2050. 2050

Based on the knowledge obtained in the research and development until 2030, we will conduct research and development toward the realization of an all-optical error-tolerant large-scale general-purpose optical quantum computer. With the aid of electrical circuits employed until 2030, the broadband nature of optical signals will not be fully exploited, i.e. the clock frequency of a quantum computer will be limited below several GHz due to the low clock frequency of an electrical signal processing system. Problems with fast quantum algorithms can be solved quickly, but general problems can only be expected to be as fast as a classical computer. The electrical energy consumption is still expected to be reduced drastically, it is possible to achieve higher speed by using more multiple cores compared to classical computers.

We are aiming to achieve this all-optical system between 2030 and 2050. If this is realized, all computers will be replaced by optical quantum computers with a clock frequency of 10 THz, and with multi-core technology, people will have the computer ability that is unimaginable today.