Scalably learning one-dimensional quantum many-body Hamiltonians from dynamical data
6 affiliations
- 1Freie Universität Berlin
- 2PASQAL SAS
- 3Quantum Research Center, Technology Innovation Institute
- 4Simons Institute for the Theory of Computing, University of California at Berkeley
- 5Fraunhofer Heinrich Hertz Institute
- 6Helmholtz-Zentrum Berlin für Materialien und Energie
Abstract
The physics of a closed quantum mechanical system is governed by its Hamiltonian. However, in most practical situations, this Hamiltonian is not precisely known, and ultimately all there is are data obtained from measurements on the system. In this work, we introduce a highly scalable, data-driven approach to learning families of interacting many-body Hamiltonians from dynamical data, by bringing together techniques from gradient-based optimization from machine learning with efficient quantum state representations in terms of tensor networks. Our approach is highly practical, experimentally friendly, and intrinsically scalable to allow for system sizes of above 100 spins. In particular, we demonstrate on synthetic data that the algorithm works even if one is restricted to one simple initial state, a small number of single-qubit observables, and time evolution up to relatively short times. For the concrete example of the one-dimensional Heisenberg model our algorithm exhibits an error constant in the system size and scaling as the inverse square root of the size of the data set.
Related
- Operator Score Matching for Learning Quantum Hamiltonians
- Linear-Scaling Quantum Transport from Machine-Learning Density Functional Theory Hamiltonians
- Orbit-resolved spectra and dynamical accessibility in multi-register covering Hamiltonians
- Full-Trajectory Learning of Open Quantum Systems: Dynamical Emulation and the Limits of Hamiltonian Identifiability
- Efficient thermalization of one-dimensional noncommuting Hamiltonians via system-bath interactions with Θ(n) ancilla qubits
BibTeX
@article{wilde2026scalably,
title = {{Scalably learning one-dimensional quantum many-body Hamiltonians from dynamical data}},
author = {Frederik Wilde and Augustine Kshetrimayum and Ingo Roth and Dominik Hangleiter and Ryan Sweke and Jens Eisert},
journal = {Quantum Science and Technology},
year = {2026},
month = may,
doi = {10.1088/2058-9565/ae6fe3},
url = {https://doi.org/10.1088/2058-9565/ae6fe3},
note = {CC BY},
}
RIS
TY - JOUR
TI - Scalably learning one-dimensional quantum many-body Hamiltonians from dynamical data
AU - Frederik Wilde
AU - Augustine Kshetrimayum
AU - Ingo Roth
AU - Dominik Hangleiter
AU - Ryan Sweke
AU - Jens Eisert
PY - 2026
DA - 2026/05/19
JO - Quantum Science and Technology
DO - 10.1088/2058-9565/ae6fe3
UR - https://doi.org/10.1088/2058-9565/ae6fe3
AB - The physics of a closed quantum mechanical system is governed by its Hamiltonian. However, in most practical situations, this Hamiltonian is not precisely known, and ultimately all there is are data obtained from measurements on the system. In this work, we introduce a highly scalable, data-driven approach to learning families of interacting many-body Hamiltonians from dynamical data, by bringing together techniques from gradient-based optimization from machine learning with efficient quantum state representations in terms of tensor networks. Our approach is highly practical, experimentally friendly, and intrinsically scalable to allow for system sizes of above 100 spins. In particular, we demonstrate on synthetic data that the algorithm works even if one is restricted to one simple initial state, a small number of single-qubit observables, and time evolution up to relatively short times. For the concrete example of the one-dimensional Heisenberg model our algorithm exhibits an error constant in the system size and scaling as the inverse square root of the size of the data set.
SN - 2058-9565
ER -