A collaborative study from University College London shows that integrating quantum computing with artificial intelligence improves long-term predictions of complex physical systems, notably fluid dynamics.

Future work aims to scale to larger datasets, apply the method to real-world problems, and develop a provable theoretical framework for the approach.

The method first uses a quantum computer to extract invariant statistical patterns, which then guide training of an AI model running on a classical supercomputer, enabling significant memory efficiency.

Quantum effects like entanglement and superposition allow the quantum component to compactly represent complex system states, driving the performance gains.

Experiments were conducted on a 20-qubit IQM quantum computer connected to powerful classical resources at Germany’s Leibniz Supercomputing Centre, operating at low temperatures.

Results indicate practical quantum advantage with potential applications in climate forecasting, blood flow modeling, molecular interactions, and wind-farm optimization, signaling broader impacts for scientific simulations.

The hybrid quantum-informed AI model achieves about 20% higher accuracy and maintains stability over longer prediction horizons than conventional AI models.

Funding and support come from UCL, EPSRC, IQM Quantum Computers, and the Leibniz Supercomputing Centre, highlighting strong institutional collaboration.