Project: Quantum Computing
Quantum computers are not yet machines that can do everything. We use real hardware to find out what they can solve today, and how they should be combined with the supercomputers we already have.
What we work on
Our center of gravity is Ising machines and quantum annealing for combinatorial optimization — an Ising machine is a computer specialized in finding the best combination among an enormous number of options. We develop methods for rewriting problems into a form called QUBO, and we evaluate execution performance across multiple machines. Alongside this, we design and evaluate quantum circuits for gate-based quantum computers, the type capable of general-purpose computation.
Why it matters
Delivery routing, materials discovery, evacuation planning: when the number of combinations explodes, conventional computers can fail to solve the problem in any realistic time. Quantum and Ising computation brings a different way of attacking these problems — but it cannot be used on its own. Our research covers the whole picture, including how to make it work together with supercomputers.
Main activities
- Feasibility Study on the future HPCI (Quantum Hybrid)
- Performance evaluation of Ising machines and quantum annealing, and QUBO formulation
- Quantum-circuit design for gate-based quantum computers (presented at ISC High Performance 2026)
- Development of Qukit-learn, a shared library for quantum machine learning (available in a tutorial environment on AIST's ABCI-Q)
- Applying Ising-based clustering to materials informatics
- Joint research with the German Aerospace Center (DLR) on detecting flow separation with Ising computation
- A demonstration with Sumitomo Corporation: optimizing vehicle repositioning for EV car sharing
Keywords
Quantum annealing / Ising machines / Combinatorial optimization / Quantum circuits / Hybrid quantum-classical computing