Citing NetKet

Citing NetKet#

Note

NetKet provides a command-line tool netket-cite to help you generate the appropriate references and acknowledgment text.

NetKet is developed by several researchers, both as part of their work and in their spare time. As with anyone in academia, careers depend heavily on citations and publication impact. Acknowledging those contributions is important, as it encourages the growth of an high-quality open-source ecosystem where researchers are rewarded for taking an active effort.

Important

If you use NetKet for academic research, or have used NetKet as inspiration or for learning in academic research, you must properly acknowledge it by citing BOTH NetKet publications. Additionally, depending on the algorithms you employed, you might have to cite additional algorithmic papers (read below).

While some researchers still pretend that software citations are not due, such behaviour is deeply irrespective of the work put in by all contributors to NetKet (and other projects) and, quite frankly, offensive.

For proper acknowledgment, include a sentence, depending on your situation, like one of the following ones:

Simulations for this work were performed using 
NetKet~\cite{netket3:2022,netket2:2019}.
% or
Simulations for this work were performed 
with codes built on top of NetKet~\cite{netket3:2022,netket2:2019}.
% or
We acknowledge the NetKet codebase as a learning 
and inspiration resource for the codes used as 
part of this work~\cite{netket3:2022,netket2:2019}.

% additionally
This software is built on top of JAX~\cite{jax2018github} 
and Flax~\cite{flax2020github}.
Click here to reveal the corresponding bibtex entries. 
@Article{netket3:2022,
    title={NetKet 3: Machine Learning Toolbox for Many-Body Quantum Systems},
    author={Filippo Vicentini and Damian Hofmann and Attila Szabó and Dian Wu and Christopher Roth and Clemens Giuliani and Gabriel Pescia and Jannes Nys and Vladimir Vargas-Calderón and Nikita Astrakhantsev and Giuseppe Carleo},
    journal={SciPost Phys. Codebases},
    pages={7},
    year={2022},
    publisher={SciPost},
    doi={10.21468/SciPostPhysCodeb.7},
    url={https://scipost.org/10.21468/SciPostPhysCodeb.7}
}

@article{netket2:2019,
    title={NetKet: A machine learning toolkit for many-body quantum systems},
    author={Carleo, Giuseppe and Choo, Kenny and Hofmann, Damian and Smith, James ET and Westerhout, Tom and Alet, Fabien and Davis, Emily J and Efthymiou, Stavros and Glasser, Ivan and Lin, Sheng-Hsuan and Mauri, Marta and Mazzola, Guglielmo and Pereira, Christian B and Vicentini, Filippo},
    journal={SoftwareX},
    volume={10},
    pages={100311},
    year={2019},
    publisher={Elsevier},
    doi={10.1016/j.softx.2019.100311},
    url={https://www.sciencedirect.com/science/article/pii/S2352711019300974}
}

@misc{jax2018github,
  title = {{{JAX}}: Composable Transformations of {{Python}}+{{NumPy}} Programs},
  author = {Bradbury, James and Frostig, Roy and Hawkins, Peter and Johnson, Matthew James and Leary, Chris and Maclaurin, Dougal and Necula, George and Paszke, Adam and VanderPlas, Jake and {Wanderman-Milne}, Skye and Zhang, Qiao},
  year = {2018}
}

@misc{flax2020github,
  title = {Flax: {{A}} Neural Network Library and Ecosystem for {{JAX}}},
  author = {Heek, Jonathan and Levskaya, Anselm and Oliver, Avital and Ritter, Marvin and Rondepierre, Bertrand and Steiner, Andreas and {van Zee}, Marc},
  year = {2024}
}

Several algorithms implemented in NetKet are derived from academic publications unrelated to NetKet itself. Those contributions should also be acknowledged. NetKet provides a command-line tool netket-cite to help you generate the appropriate references and acknowledgment text. You can also find an automatically-generated list of references below in the section Algorithm-Specific Citations

Citation Tools#

Use the netket-cite command-line tool to get the complete citation information:

# Display all relevant citations for your NetKet usage
netket-cite

# Generate a complete references.bib file
netket-cite --bib {optional_name.bib}

# From within python
>>> nk.cite(bib=True)

Additional references#

MPI Support#

If you use NetKet’s MPI functionality, please also cite the mpi4jax library that enables this feature:

@article{mpi4jax:2021,
    title={mpi4jax: Zero-copy MPI communication of JAX arrays},
    author={Häfner, Dion and Vicentini, Filippo},
    journal={Journal of Open Source Software},
    volume={6},
    number={65},
    pages={3419},
    year={2021},
    publisher={The Open Journal},
    doi={10.21105/joss.03419},
    url={https://joss.theoj.org/papers/10.21105/joss.03419}
}

Algorithm-Specific Citations#

Many algorithms implemented in NetKet are based on original research publications. When using specific features or algorithms, cite the relevant papers that introduced these methods. The citation tools help identify these references.

If using pinv_smooth_distributed solver: This work used the JAXMg distributed linear solver described in Ref. (Wiersema2026jaxmg)"JAXMg: A multi-GPU linear solver in JAX"

BibTeX entries
@misc{Wiersema2026jaxmg,
  title = {JAXMg: A multi-GPU linear solver in JAX},
  author = {Roeland Wiersema},
  year = {2026},
  eprint = {2601.14466},
  archivePrefix = {arXiv},
  primaryClass = {cs.MS},
  url = {https://arxiv.org/abs/2601.14466}
}

If using TDVPSchmitt: This work used the TDVP algorithms described in Ref. (Schmitt2020)"Quantum Many-Body Dynamics in Two Dimensions with Artificial Neural Networks"

BibTeX entries
@article{Schmitt2020,
    title={Quantum Many-Body Dynamics in Two Dimensions with Artificial Neural Networks},
    author={Markus Schmitt and Markus Heyl},
    journal={Physical Review Letters},
    volume={125},
    pages={100503},
    year={2020},
    publisher={APS},
    doi={10.1103/PhysRevLett.125.100503},
    url={https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.125.100503}
}

If using minSR, VMC_SR(use_ntk=True): This work used the efficient kernel trick for SR described in Refs. (Chen2024minsr)"Empowering deep neural quantum states through efficient optimization" (Rende2024minsr)"A simple linear algebra identity to optimize large-scale neural network quantum states"

BibTeX entries
@article{Chen2024minsr,
  title = {Empowering deep neural quantum states through efficient optimization},
  volume = {20},
  ISSN = {1745-2481},
  url = {http://dx.doi.org/10.1038/s41567-024-02566-1},
  DOI = {10.1038/s41567-024-02566-1},
  number = {9},
  journal = {Nature Physics},
  publisher = {Springer Science and Business Media LLC},
  author = {Chen,  Ao and Heyl,  Markus},
  year = {2024},
  month = jul,
  pages = {1476–1481}
}

@article{Rende2024minsr,
  title = {A simple linear algebra identity to optimize large-scale neural network quantum states},
  volume = {7},
  ISSN = {2399-3650},
  url = {http://dx.doi.org/10.1038/s42005-024-01732-4},
  DOI = {10.1038/s42005-024-01732-4},
  number = {1},
  journal = {Communications Physics},
  publisher = {Springer Science and Business Media LLC},
  author = {Rende,  Riccardo and Viteritti,  Luciano Loris and Bardone,  Lorenzo and Becca,  Federico and Goldt,  Sebastian},
  year = {2024},
  month = aug 
}

% Infidelity

If using VMC_SR with momentum != 0: This work used the SPRING optimization algorithm described in Ref. (Goldshlager2023Spring)"A Kaczmarz-inspired approach to accelerate the optimization of neural network wavefunctions"

BibTeX entries
@article{Goldshlager2023Spring,
  title = {A Kaczmarz-inspired approach to accelerate the optimization of neural network wavefunctions},
  volume = {516},
  ISSN = {0021-9991},
  url = {http://dx.doi.org/10.1016/j.jcp.2024.113351},
  DOI = {10.1016/j.jcp.2024.113351},
  journal = {Journal of Computational Physics},
  publisher = {Elsevier BV},
  author = {Goldshlager,  Gil and Abrahamsen,  Nilin and Lin,  Lin},
  year = {2024},
  month = nov,
  pages = {113351}
}

If using infidelity estimators and optimizers: This work used the Infidelity estimators and optimizers from Refs. (Sinibaldi2023Unbiasing)"Unbiasing time-dependent Variational Monte Carlo by projected quantum evolution" (Gravina2024PTVMC)"Neural Projected Quantum Dynamics: a systematic study"

BibTeX entries
@article{Sinibaldi2023Unbiasing,
  title = {Unbiasing time-dependent Variational Monte Carlo by projected quantum evolution},
  volume = {7},
  ISSN = {2521-327X},
  url = {http://dx.doi.org/10.22331/q-2023-10-10-1131},
  DOI = {10.22331/q-2023-10-10-1131},
  journal = {Quantum},
  publisher = {Verein zur Forderung des Open Access Publizierens in den Quantenwissenschaften},
  author = {Sinibaldi,  Alessandro and Giuliani,  Clemens and Carleo,  Giuseppe and Vicentini,  Filippo},
  year = {2023},
  month = oct,
  pages = {1131}
}

@article{Gravina2024PTVMC,
  title = {Neural Projected Quantum Dynamics: a systematic study},
  volume = {9},
  ISSN = {2521-327X},
  url = {http://dx.doi.org/10.22331/q-2025-07-22-1803},
  DOI = {10.22331/q-2025-07-22-1803},
  journal = {Quantum},
  publisher = {Verein zur Forderung des Open Access Publizierens in den Quantenwissenschaften},
  author = {Gravina,  Luca and Savona,  Vincenzo and Vicentini,  Filippo},
  year = {2025},
  month = jul,
  pages = {1803}
}