Machine-Learning-Based Prediction of Quantum Magic

Deep-learning research for predicting quantum-state magic from local measurements, with tensor-network simulation for larger quantum systems. Co-First Author.

Methods & Topics

Quantum Machine LearningDeep LearningTensor NetworksQuantum Magic

Overview

This work introduces machine learning to quantum-state magic estimation. A deep-learning model learns the nonlinear relationship between local measurement features and the magic of ground states and random states.

My Contributions

  • Developed a prediction model that estimates quantum-state magic without directly evaluating complex magic measures.
  • Achieved relative accuracy above 99%.
  • Used tensor networks to simulate systems of tens of qubits and demonstrated size extrapolation from small training systems to larger quantum systems without additional training.