Learning ground state observables from quantum computing experiments
Researchers have demonstrated the practical application of machine learning on experimental quantum data from the Heisenberg XXZ model, utilizing up to 115 qubits. By training neural networks on expectation values and correlation data, they accurately predicted observables for unseen Hamiltonian parameters, even approaching the phase boundary. This work shows the potential for quantum processors to generate data for machine learning models that surpasses classical approximation capabilities. AI
IMPACT Demonstrates a novel method for using quantum computing to generate data for machine learning, potentially accelerating scientific discovery in physics.