Protocol using static single-qubit fields achieves Heisenberg-limited learning of geometrically local many-body Hamiltonians without requiring fast or trusted multi-qubit gates, with field strength independent of system size.

#QuantumControl #HamiltonianLearning #Research

I was intrigued about the problem of #HamiltonianLearning, and could not believe at first that it could be so difficult. It turned out that the true challenge lies in making the method robust to render in applicable in practical situations.

Novel eigenphase engineering technique enables Heisenberg-limited learning of sparse k-local Hamiltonians with near-maximal step size—matching best known total evolution time while removing precision-dependent bottlenecks in quantum control.

#QuantumAlgorithms #QuantumControl #HamiltonianLearning

Hamiltonian Learning via Inverse Physics-Informed Neural Networks
Jie Liu, Xin Wang
Paper
Details
#HamiltonianLearning #PhysicsInformedNNs #MachineLearningPhysics

Efficient protocols now enable learning of Hamiltonian parameters in bosonic quantum systems with logarithmic sample complexity—using only heterodyne measurements, dramatically reducing observations needed.

#HamiltonianLearning #QuantumStates #QuantumInformation

New algorithm learns unknown quantum Hamiltonians without quantum controls or ancillas, achieving optimal standard quantum limit scaling—crucial for practical in situ device characterization and calibration on near-term quantum platforms.

#QuantumAlgorithms #QuantumMetrology #HamiltonianLearning

Hybrid Bayesian MCMC—combining reversible-jump MCMC & parallel tempering—recovers nuclear spin hyperfine couplings near semiconductor spin-defects using ~10x less data than prior methods, validated on experimental diamond NV-center measurements.

#HamiltonianLearning #QuantumSensing #News