Quantum

The Quantum Reference Architecture aims to provide HUN-REN Science Cloud users with a unified, easy-to-use application layer for exploring and conducting research on the available quantum computing resources. Through hands-on examples and runnable notebooks, the architecture demonstrates how classical and quantum tools can be integrated to solve optimization, machine-learning and algorithmic problems.

• The system is accessed through a JupyterLab-based interface, where users can learn to build and run quantum circuits, work with basic algorithms and tackle optimization problems – for example QAOA-based problems, TSP and QUBO or Ising formulations – using prepared, runnable notebooks. The examples cover the most widely used quantum SDKs, including IBM Qiskit, D-Wave Leap, Amazon Braket and PennyLane, and Spark integration is also available for larger data-processing workloads.
• A key focus area is quantum machine learning, with quantum autoencoders, quantum-kernel methods, QSVM, quantum neural networks and hybrid quantum-classical neural network architectures all available, along with comparisons against classical baseline models.

• The examples can be run on real quantum hardware – such as IBM Quantum, D-Wave or Amazon Braket – as well as on CPU- and GPU-accelerated simulators, making full-featured experimentation possible even without QPU access.