In this project, we develop quantum-classical hybrid algorithms by combining quantum computation and classical machine learning. In particular, we aim to boost the performance of noisy and intermediate-scale quantum (NISQ) computers available by the state-of-the-art technology, with the use of post-processing by machine learning such as neural networks. The goal of this project is application of the developed algorithms to important problems in material science such as quantum condensed-matter physics and quantum chemistry.

Quantum Artificial Intelligence (QAI) is considered to be a promising quantum computing application in the present era of noisy-intermediate scale quantum computers. There are two aims in this research project: first, we focus on a QAI algorithm to coherently learn quantum data using quantum simulations of quantum field theory as benchmarks; second, we develop techniques to design quantum circuits and optimize quantum gates, tailored for specific high energy physics applications.

To develop quantum information science and technology beyond the NISQ (Noisy Intermediate-Scale Quantum) era, we establish new programming methods that implement higher-order functions in quantum computers based on higher-order quantum operations. We are developing novel quantum algorithms and quantum applications for quantum simulation and quantum sensors based on this method. We also explore the new frontier of quantum computation by improving our understanding of the spacetime structure of quantum physics in terms of information processing and clarifying characteristic properties of parallelizability and causal structures of quantum programming.

Quantum chemistry on quantum computers has attracted much attention as a promising target of near-term quantum devices. Quantum chemistry solves the time-independent Schrödinger equation of electrons. On the other hand, time-dependent Schrödinger equation (TDSE) on quantum computers is far less investigated. One needs to solve TDSE to describe the light-matter interactions, which, however, suffers from the problem of combinatorial explosion peculiar to quantum many-body systems. In this research, we will implement a hybrid quantum/classical simulator of multielectron dynamics on IBM Quantum. Furthermore, we will extend the hybrid simulator to treat nonadiabatic dynamics where both electrons and nuclei are treated quantum mechanically. Using the simulator, we will demonstrate the first successful application of the real quantum computer to nontrivial multielectron and/or nonadiabatic dynamics such as dissociative ionization in a hydrogen molecule. This will pave the way for future realization, e.g., of accurate simulations of biologically relevant photoreactions, which would be difficult with a classical computer.

Quantum computation is now attracting attention in various fields, although there are fundamental issues yet to be solved for its practical application. In this project, by exploiting the spintronics and AI physics we have established so far, we will develop q-bit drivers and efficient calibration methods that are indispensable for improving the performance of quantum computers.

Superconducting quantum circuits are being actively investigated for quantum information processing. In a quantum transducer, electromagnetic waves in the microwave region handled by superconducting quantum circuits are converted into light and vice versa. The quantum transducer provides a quantum interconnect between superconducting quantum computers, and also to control them optically. Therefore, the optical interface using quantum transducers will be a core technology for the expansion of quantum technologies such as large scale quantum computers and a quantum internet. In this research, we will develop a quantum transducer that connects light and microwaves via acoustic quanta such as vibrations of objects and elastic waves. By combining high performance optical waveguide, superconducting technology and optomechanical technology, we aim to achieve highly efficient optical microwave conversion.

Quantum states formed in relatively simple circuit dynamics of a superconducting processor are subject to external noise sources and unwanted inter-qubit interactions. Quantum states in a multi-junction circuit possess symmetry which can be used to protect qubits from noise effects. In this research, we will leverage the added circuit degrees of freedom of the multi-junction qubit to enhance the qubit performance. We will also explore novel two-qubit gates and gate decompositions using higher energy levels of the qubit, through which we aim to extend the computational capacity of current noisy superconducting processors.