Precise measurement of quantum observables with neural-network estimators
Giacomo Torlai, Guglielmo Mazzola, Giuseppe Carleo, Antonio Mezzacapo
Abstract
This work introduces a technique to lower the overhead in measurement statistics to achieve accurate estimations of observables relevant in quantum simulations. The approach consists of integrating artificial neural networks with the measurement apparatus of a quantum simulator. By training the neural networks with unsupervised learning of single-qubit measurements, the authors demonstrate a significant decrease in the number of measurements required to reach precise estimates in quantum chemistry applications
Topics & Concepts
ObservableQuantumEstimatorArtificial neural networkComputer scienceStatistical physicsAlgorithmWork (physics)Quantum systemMathematicsQuantum phase estimation algorithmQuantum algorithmArtificial intelligenceOverhead (engineering)Observational errorPhysicsQuantum stateQuantum sensorQuantum metrologyMeasure (data warehouse)Estimation theoryQuantum measurementQuantum operationQuantum processQuantum networkUnsupervised learningQuantum Computing Algorithms and ArchitectureMachine Learning in Materials ScienceQuantum many-body systems