Skip to content
DraconiticResearch
Search Sources API
197,357 studies31 sourcesUpdated 11 min ago
Peer-reviewedCC BYPhysical Review X

Variational Multi-Gaussian Phase-Space Bosonic Dynamics via Automatic Differentiation

Abstract

We introduce a variational method for simulating the dynamics of interacting open quantum bosonic systems deep in the quantum regime. The method is based on a multidimensional Wigner phase-space representation and employs a variational multi-Gaussian , whose accuracy is systematically controlled by the number of Gaussian components. The variational equations of motion are derived from the Dirac-Frenkel principle and evaluated efficiently by combining the analytical structure of Gaussian functions with automatic differentiation. As a key physical application, we study a driven-dissipative two-dimensional Bose-Hubbard lattice with two-boson coherent driving and two-body losses. Using our dynamical approach, we compute the finite-size scaling of the Liouvillian spectral gap—extracted from the relaxation dynamics—which vanishes in the thermodynamic limit. Our results reveal critical slowing down with dynamical exponents of the 2D quantum Ising universality class, demonstrating the power of our method to capture complex quantum dynamics in large open systems.

Related

  1. CuAD: Source-to-Source Automatic Differentiation of CUDA Kernels for Discrete Adjoints in Computational Fluid Dynamics
    Seunghyun Joo
    arXivPreprintarXiv licenceFull textComputational physicsFluid dynamicsarXiv
  2. Energy conserving variational dynamics with Gaussian wavepackets in the nonadiabatic regime
    Michael Anim Safo, Giacomo Botti, Sophya Garashchuk
    ChemRxivPreprintCC BYFull textAtomic, molecular & optical physicsChemRxiv
  3. Derivative Computation in PINNs: Automatic Differentiation, Finite Differences and Beyond
    Maciej J. Mikulski, Tadeusz Uhl
    arXivPreprintarXiv licenceFull textNumerical analysis & scientific computingComputational physicsarXiv
  4. Locating critical solutions in numerical relativity using automatic differentiation
    David Bambague, Miguel Bezares, Katy Clough +1
    arXivPreprintarXiv licenceFull textGravitation & general relativityarXiv
  5. Optimal Control of Periodic Nonequilibrium Mechanochemical Systems via Automatic Differentiation
    W. Callum Wareham, David A. Sivak
    arXivPreprintarXiv licenceFull textBiological & medical physicsStatistical & nonlinear physicsarXiv
BibTeX
@article{tosca2026variational,
  title = {{Variational Multi-Gaussian Phase-Space Bosonic Dynamics via Automatic Differentiation}},
  author = {Jacopo Tosca and Francesco Carnazza and Luca Giacomelli and Cristiano Ciuti},
  journal = {Physical Review X},
  year = {2026},
  month = sep,
  doi = {10.1103/q3m5-q44b},
  eprint = {2507.14076},
  archivePrefix = {arXiv},
  url = {https://hal.science/hal-05842833v1},
  note = {CC BY},
}
RIS
TY  - JOUR
TI  - Variational Multi-Gaussian Phase-Space Bosonic Dynamics via Automatic Differentiation
AU  - Jacopo Tosca
AU  - Francesco Carnazza
AU  - Luca Giacomelli
AU  - Cristiano Ciuti
PY  - 2026
DA  - 2026/09/24
JO  - Physical Review X
DO  - 10.1103/q3m5-q44b
UR  - https://hal.science/hal-05842833v1
AB  - We introduce a variational method for simulating the dynamics of interacting open quantum bosonic systems deep in the quantum regime. The method is based on a multidimensional Wigner phase-space representation and employs a variational multi-Gaussian , whose accuracy is systematically controlled by the number of Gaussian components. The variational equations of motion are derived from the Dirac-Frenkel principle and evaluated efficiently by combining the analytical structure of Gaussian functions with automatic differentiation. As a key physical application, we study a driven-dissipative two-dimensional Bose-Hubbard lattice with two-boson coherent driving and two-body losses. Using our dynamical approach, we compute the finite-size scaling of the Liouvillian spectral gap—extracted from the relaxation dynamics—which vanishes in the thermodynamic limit. Our results reveal critical slowing down with dynamical exponents of the 2D quantum Ising universality class, demonstrating the power of our method to capture complex quantum dynamics in large open systems.
SN  - 2160-3308
ER  -