High-Accuracy Temporal Prediction via Experimental Quantum Reservoir Computing in Correlated Spins

Area of Science:

• Quantum physics
• Machine learning
• Complex systems

Background:

• Physical reservoir computing leverages nonlinear dynamics for efficient information processing.
• Quantum reservoir computing (QRC) promises enhanced machine learning capabilities due to the complexity of quantum dynamics.
• Simulating quantum dynamics classically is computationally expensive, motivating novel QRC approaches.

Purpose of the Study:

• To present a new QRC approach utilizing correlated quantum spin systems.
• To exploit natural quantum many-body interactions for reservoir dynamics, avoiding complex quantum circuits.
• To demonstrate the potential of this method for high-performance machine learning applications.

Main Methods:

• Experimental implementation of a quantum reservoir computer based on a 9-spin quantum system.
• Utilizing natural quantum many-body interactions to generate reservoir dynamics.
• Testing the system on standard time-series benchmarks and long-term weather forecasting.

Main Results:

• The quantum reservoir system demonstrated nontrivial quantum entanglement and sufficient dynamical complexity.
• Achieved state-of-the-art performance on time-series benchmarks, reducing prediction error by 1-2 orders of magnitude.
• Outperformed large-scale classical reservoirs in long-term weather forecasting accuracy.

Conclusions:

• This work presents the first experimental demonstration of a quantum machine learning model surpassing classical models on real-world tasks.
• Correlated quantum spin systems offer a viable and powerful platform for practical quantum reservoir computing.
• The developed QRC approach shows significant promise for advancing machine learning capabilities through quantum phenomena.