What is my quantum computer good for? Quantum capability learning with physics-aware neural networks

2024-06-09

Neural Information Processing Systems

AI Summary

Current NISQ devices suffer from hardware-induced noise, causing quantum program failures and lacking fast, reliable methods for assessing execution capability. Method: This paper proposes a physics-aware neural network that explicitly incorporates physical error mechanisms—particularly gate error propagation—into its architecture, integrating graph neural networks (GNNs) with approximate error propagation modeling to overcome the limitations of black-box models that violate quantum physical constraints. Contribution/Results: The model performs end-to-end regression to predict program-level success probability. Evaluated on both simulation and real-device experimental data, it achieves ~50% lower mean absolute error than a CNN baseline, significantly improving prediction accuracy and cross-hardware generalization. By embedding domain-specific physical priors, it establishes an interpretable, scalable paradigm for characterizing quantum hardware capability.

Technology Category

• Machine Learning: Quantum Machine Learning
• Search and Optimization: Learning to Search
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Security and Privacy: Large-scale security measurements

Abstract

Quantum computers have the potential to revolutionize diverse fields, including quantum chemistry, materials science, and machine learning. However, contemporary quantum computers experience errors that often cause quantum programs run on them to fail. Until quantum computers can reliably execute large quantum programs, stakeholders will need fast and reliable methods for assessing a quantum computer's capability-i.e., the programs it can run and how well it can run them. Previously, off-the-shelf neural network architectures have been used to model quantum computers' capabilities, but with limited success, because these networks fail to learn the complex quantum physics that determines real quantum computers' errors. We address this shortcoming with a new quantum-physics-aware neural network architecture for learning capability models. Our architecture combines aspects of graph neural networks with efficient approximations to the physics of errors in quantum programs. This approach achieves up to $sim50%$ reductions in mean absolute error on both experimental and simulated data, over state-of-the-art models based on convolutional neural networks.

Problem

• Assessing quantum computer capabilities
• Modeling quantum physics errors
• Improving neural network accuracy

Innovation

• Quantum-physics-aware neural network
• Graph neural networks integration
• Efficient quantum error approximations