Pith. sign in

REVIEW 3 major objections 5 minor 46 references

For interacting fermion Hamiltonians given as one- and two-body terms, the Hartree-Fock solution can be obtained by directly minimizing the variational energy of a Slater determinant using automatic differentiation, without deriving the mea

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

AutoHF provides a general, automatic-differentiation-based Hartree–Fock solver that directly minimizes the variational energy for lattice and molecular fermion Hamiltonians.

T0 review reviewed 2026-08-02 challenge →

load-bearing objection A genuinely useful HF codebase with a correct formalism and honest benchmarks; the 'global minima with remarkable reliability' claim outruns the evidence and should be trimmed in revision. the 3 major comments →

arxiv 2607.14263 v2 pith:GEEHH5YU submitted 2026-07-15 cond-mat.str-el physics.comp-ph

AutoHF: a general Hartree-Fock solver utilizing direct energy minimization with automatic differentiation

Ryan Levy , Brandon Eskridge , Lukas Weber , Miguel A. Morales , Shiwei Zhang This is my paper
classification cond-mat.str-el physics.comp-ph
keywords Hartree-Fockdirect energy minimizationautomatic differentiationSlater determinantone-body density matrixHubbard modelgeneralized Hartree-Fockquantum chemistry
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

AutoHF is a Hartree-Fock solver that seeks to establish a single recipe: for a fermion Hamiltonian specified by one- and two-body terms, the optimal Slater determinant is found by directly minimizing the variational energy ⟨Φ|H|Φ⟩/⟨Φ|Φ⟩ with respect to the orbital coefficients, using automatic differentiation for gradients. The key step is expressing the energy through the one-body reduced density matrix, which turns any two-body Hamiltonian into a differentiable cost function and removes the need to derive the mean-field (Fock) operator by hand. The paper shows the approach handles Hubbard lattices—reproducing striped phases and noncollinear 120-degree magnetic order—and molecules like N2 and 3d transition-metal oxides, where conventional self-consistent-field codes need special initialization or stability analysis. The overall claim is that direct variational optimization plus automatic differentiation is a general, reliable route to Hartree-Fock solutions.

Core claim

On the paper's own terms, the central claim is that the Hartree-Fock ground state of an arbitrary many-fermion Hamiltonian is accessible by minimizing E_var(α) = ⟨Φ(α)|H|Φ(α)⟩/⟨Φ(α)|Φ(α)⟩, where the Slater determinant |Φ(α)⟩ is an explicit differentiable function of variational parameters. Because the expectation value of any two-body Hamiltonian reduces to a functional of the one-body reduced density matrix ρ(α) = Φ*(ΦᵀΦ*)⁻¹Φᵀ, the energy can be computed from the orbital coefficients, and automatic differentiation supplies the gradient. The paper demonstrates that a multi-start limited-memory quasi-Newton optimizer, run in parallel and selecting the lowest-energy trial, recovers the expecte

What carries the argument

The machinery has three parts: a parameterized Slater determinant (direct orbital coefficients, unitary rotation of a reference determinant, or eigenvectors of a parameterized effective Hamiltonian), the map to the one-body reduced density matrix ρ(Φ) = [Φ*(ΦᵀΦ*)⁻¹Φᵀ] which makes the variational energy a differentiable scalar function of the parameters, and automatic differentiation with a multi-start limited-memory quasi-Newton optimizer. Symmetry control is available either through the ansatz (restricted/unrestricted/generalized Hartree-Fock) or through penalty terms added to the objective; the default scheme runs many random initializations in parallel and takes the lowest converged energ

Load-bearing premise

The claim that the best trial over random initializations is the true global Hartree-Fock minimum is assumed on empirical grounds; the nonconvex energy landscape may contain local minima that the multi-start search misses, and the paper's own most difficult example required a specialized ansatz and 200 trials.

What would settle it

For a small, fully enumerated fermion system—for example a doped 4x4 or 6x6 Hubbard cluster whose mean-field stationary points can be found exhaustively by solving the Hartree-Fock equations from every symmetry-distinct initial state—compare the true lowest stationary energy with the best energy AutoHF returns across, say, 1000 random starts. A single case where the solver systematically converges above the exhaustive minimum (beyond numerical tolerance) would refute the 'global minimum with high reliability' claim. A molecular analog is N2 at intermediate bond length, where the solver must ma

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • New lattice or molecular Hamiltonians can be given a Hartree-Fock reference without writing a model-specific mean-field decoupling: the user specifies the terms and the solver minimizes the energy.
  • Restricted, unrestricted, and noncollinear (generalized) Hartree-Fock become settings rather than separate programs, and the last naturally produces spin states such as a 120-degree Néel order on the triangular lattice.
  • Custom energy penalties can select a desired symmetry sector among nearly degenerate Hartree-Fock states without changing the converged energy, enabling controlled symmetry breaking or restoration.
  • In the test cases, the direct approach matches or betters conventional self-consistent field: it recovers the symmetry-broken N2 solution that standard codes reach only through stability analysis, and it finds a lower variational energy than a published reference value for VO.
  • Molecules are handled through a Cholesky-decomposed Coulomb interaction, so the same solver spans lattice models and ab initio chemistry.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • A stress test not performed in the paper: characterize how the number of independent random starts must scale with system size and interaction strength to preserve the reported reliability; the transition-metal-oxide cases needed a specialized restricted-open-shell ansatz and hundreds of trials, suggesting the multi-start heuristic has quantitative limits.
  • Because the objective is expressed entirely through the one-body density matrix, the machinery should carry over to the extensions the paper lists as future work—Hartree-Fock-Bogoliubov and finite-temperature mean-field theory—with a free-energy functional replacing the ground-state energy.
  • For downstream many-body methods that need a trial state, this framework makes it easy to impose the exact pinning fields and symmetry constraints used in the auxiliary-field Monte Carlo step, shifting the engineering burden from mean-field algebra to ansatz design and penalty tuning.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper presents AutoHF, a Python/JAX package that performs Hartree-Fock by direct minimization of the variational energy E_var = <Phi|H|Phi>/<Phi|Phi> over Slater-determinant parameters, using automatic differentiation and generic optimizers (L-BFGS by default, with gradient descent and basin hopping also interfaced). The variational energy is expressed through the one-body reduced density matrix (Eq. 10), so the user avoids deriving a Fock operator for each Hamiltonian. The manuscript documents implemented Hamiltonian terms (Hubbard U, density-density, spin-spin, Hund's coupling, Cholesky two-body), ansaetze (SD, SD_ROT, DIAG, custom), and examples including square-lattice Hubbard stripes, a 16x16 hole-doped Hubbard model, triangular-lattice GHF, N2 dissociation, and seven 3d transition-metal oxide diatomics. The TM-oxide table reports a VO energy lower than the literature reference, stated to be independently confirmed by PySCF.

Significance. If the claims are appropriately calibrated, this is a useful open-source contribution to rapid prototyping of HF for arbitrary fermionic Hamiltonians. The core formalism is standard and correct: reducing E_var to a functional of the 1-rdm via Wick's theorem is clean and avoids per-Hamiltonian mean-field derivations. The code is publicly available, and the molecular benchmarks provide concrete validation, including the interesting VO case where AutoHF found a lower-energy solution than the reference and was confirmed by PySCF. The main weakness is that the paper oversells the reliability of global optimization: the benchmarks establish that the reported states are the best among the trials run, not that they are global minima on a nonconvex landscape.

major comments (3)
  1. [Sec. 5 and Abstract] The conclusion states that AutoHF is 'able to find global minima with remarkable reliability,' and the abstract says it 'finds the optimal Slater determinant.' The evidence does not support the global-minimum wording. E_var is nonconvex; Sec. 3.1 selects the best variational energy among parallel L-BFGS trials, which is a heuristic best-of-N local search. The examples in Sec. 4 required additional setup: a custom ROHF ansatz and batch_size=200 for TM oxides (Listing 10), a two-stage SD_ROT/CD protocol in Sec. 4, and orbital warm starts along the N2 curve (Sec. 4.2). I request either weakening the wording to 'lowest-energy solution found among the trials' or adding statistical evidence of global optimality (e.g., repeated-seed distributions, exhaustive enumeration on small systems, or comparison to known global minima).
  2. [Sec. 4.2, Table 2, Listings 9-10] The TM-oxide example is presented as evidence that AutoHF avoids 'setting up a tailored program' and 'no need to perform a stability analysis.' In the manuscript's own workflow, however, the user must implement a custom ROHF ansatz (Listing 9) and choose batch_size=200 with a tuned random reference (Listing 10). This is substantial user input and heuristic tuning, not a parameter-free automatic global search. The comparison with PySCF is still valuable, but the narrative should acknowledge the user effort and the heuristic nature of the multi-start protocol.
  3. [Sec. 4.1, Fig. 2] The 16x16 Hubbard example reports a single run per random initialization; no seed dependence or convergence statistics are shown. Given the claim of reliable global minimization, a small statistical study (e.g., 10-20 seeds on one representative system, reporting final-energy distributions and the fraction of trials reaching the best energy) would be needed. Without such data, the conclusion 'remarkable reliability' remains an extrapolation from a handful of curated successes.
minor comments (5)
  1. [Listing 5] The makeT example has a likely bug in the periodic-boundary conditions: the x-direction hopping is gated by `yperiodic` and the y-direction hopping by `xperiodic`, and the second condition uses `if i + 1 < Ny` where it should likely be `if j + 1 < Ny`. With both flags True the code produces the correct torus, but the listing is confusing and incorrect for nonperiodic edges.
  2. [Sec. 4.1.2] 'Four different random realizations' should be 'four random initializations/seeds'; a realization usually implies disorder or a physical sample, not an optimizer start.
  3. [Table 1] There is a typo 'L-BGFS' for 'L-BFGS'. Also, the default ansatz should be stated consistently: the table's description of the `ansatz` row is ambiguous, and Sec. 3.3.2 says SD_ROT is the default while the sample settings in Listing 1 use SD.
  4. [Sec. 3.3.2] The package name is spelled inconsistently as 'auto_hf' in one place; use 'autohf' throughout.
  5. [Eq. (26) and Sec. 5] The conclusion mentions 'quadratic penalty terms that vanish at the minimum.' The penalty in Eq. (26) is an L2 norm of density differences, which is indeed quadratic in the symmetry-breaking components and vanishes if the target symmetry is satisfied, but the role of the weight lambda and the condition for vanishing should be clarified.

Circularity Check

0 steps flagged

No significant circularity: the variational energy is computed from the Hamiltonian via the 1-rdm, with no fitted parameters reused as predictions.

full rationale

The derivation chain is self-contained. Eq. (6) defines E_var as the exact expectation value of the Hamiltonian in a Slater determinant; Eq. (10) evaluates it from the 1-rdm, rho(Phi) = [Phi*(Phi^T Phi*)^{-1} Phi^T], using Wick's theorem, and the cost function Eq. (9) is minimized over orbital parameters. Nothing is fitted to a data subset and then reported as a prediction: the HF energies in Table 2 are compared with independent PYSCF/literature values, and the Cholesky integrals are generated outside the solver. The only self-referential elements are validation comparisons to earlier work from the same group (e.g., Refs. 12-15), but these are benchmarks, not inputs that define the energy or constrain the variational parameters. The global-minimum claim in Sec. 5 ('able to find global minima with remarkable reliability') is an empirical/heuristic assertion about a nonconvex optimizer, not a circular reduction; the custom ROHF ansatz and batch_size=200 for transition-metal oxides are user choices that limit generality but do not make the output equal to an input. Therefore the central claim has independent content.

Axiom & Free-Parameter Ledger

3 free parameters · 6 axioms · 0 invented entities

The method introduces no new physical entities or fitted constants. It relies on standard fermionic second-quantization machinery and on an unproved optimization assumption (multi-start local minimization finds the global HF solution). Example-specific penalty weights and tolerances are user settings, not fitted parameters.

free parameters (3)
  • penalty weights lambda (Eqs. 9, 26) = 1.0 (example, Sec. 4.1.2)
    User-supplied Lagrange/penalty parameter used to select symmetric vs symmetry-broken degenerate HF states. Not fitted to data, but the symmetry-selected result depends on choosing lambda=1 'not too large'.
  • state0_scale = 0.01 (default)
    Controls the amplitude of random noise added to initial orbitals for the parallel multi-trial optimization; a user setting, not fitted.
  • chol_tol = 1e-6 (TM oxide example)
    Cholesky decomposition tolerance used in the molecular examples; energy agreement with reference values is stated to be at this tolerance.
axioms (6)
  • standard math Wick's theorem gives the many-body energy as a functional of the 1-rdm for a single Slater determinant.
    Invoked in Sec. 2 immediately after Eq. (10) to map E_var to the 1-rdm rho(alpha).
  • domain assumption The Hartree-Fock ground state is the Slater determinant that minimizes the variational energy.
    This defines HF in Sec. 2, Eqs. (2)-(6); the paper's software optimizes this variational objective.
  • standard math The non-orthonormal Slater determinant 1-rdm formula rho = Phi*(Phi^T Phi*)^{-1} Phi^T is correct.
    Used as Eq. (10), the central computational bridge between orbital parameters and energy.
  • standard math Thouless's theorem: any Slater determinant reachable by a unitary rotation e^X from a reference determinant is a valid single-determinant ansatz.
    Used in Sec. 3.3.2, Eqs. (21)-(22), to define the SD_ROT ansatz.
  • domain assumption The nonconvex optimization landscape can be searched by random restarts and L-BFGS to reach the global HF minimum.
    Assumed in Sec. 3.1 and asserted in Sec. 5; this is the paper's weakest load-bearing premise, since no proof or systematic statistics are provided.
  • domain assumption The Hamiltonians considered are expressible as one- and two-body fermion operators (kinetic, Hubbard U, density-density, spin-spin, Hund, Cholesky Coulomb).
    The solver is formulated in second quantization; all built-in terms listed in Sec. 3.2 have this quartic form.

reviewed 2026-08-02 · how reviews work

0 comments
Cite this review

Pith. "Pith review of AutoHF: a general Hartree-Fock solver utilizing direct energy minimization with automatic differentiation." pith.science (2026). https://pith.science/paper/GEEHH5YU

@misc{pith2026260714263,
  author       = {Pith},
  title        = {Pith review of: AutoHF: a general Hartree-Fock solver utilizing direct energy minimization with automatic differentiation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GEEHH5YU}},
  note         = {Machine review of arXiv:2607.14263}
}
Share X Bluesky LinkedIn Reddit HN
abstract

We present autohf, a general, easy-to-use mean-field solver for quantum many-fermion Hamiltonians. It allows the user to bypass the process of deciphering the mean-field form for each many-body Hamiltonian $H$ and thus avoid setting up a tailored program for each $H$. Rather, autohf finds the optimal Slater determinant $|\Psi\rangle$, written in terms of orbital coefficients and subject to symmetry constraints, by directly minimizing the variational energy $\langle H \rangle$. By embracing this variational approach, autohf makes use of the growing power of automatic differentiation and optimization tools developed by the machine learning community.

Figures

Figures reproduced from arXiv: 2607.14263 by Brandon Eskridge, Lukas Weber, Miguel A. Morales, Ryan Levy, Shiwei Zhang.

Figure 1
Figure 1. Figure 1: Left: Measured staggered spin density (−1) ix+i y 〈Sˆz i 〉 using a UHF autohf calculation of a Hubbard model with edge-pinning at n = 11/12 i.e. 1/12-doping), on a width-8 cylinder, with lengths 24, 48, and 72. The top (U = 1) shows the absence of long range order while the bottom (U = 4) shows consistent order. Right: Explicit spin density 〈Sˆz i 〉 of the 24 × 8 cylinders. where T = ti j is the graph of h… view at source ↗
Figure 2
Figure 2. Figure 2: Energy differences from four different random realizations of a square [PITH_FULL_IMAGE:figures/full_fig_p021_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Measured static spin structure factor S(q) using a UHF autohf (left) and GHF autohf (right) calculation of a 12×12 U = 8 Hubbard model on the triangular lattice at half-filling (n = 1) with periodic boundaries. The UHF solution cannot properly form the 120◦ order and creates a frustrated order, while the GHF solution easily obtains the correct 120◦ magnetic order. 22 [PITH_FULL_IMAGE:figures/full_fig_p022… view at source ↗
Figure 4
Figure 4. Figure 4: UHF energy for the bond stretching of N2 , using PYSCF (Naive), PYSCF with stability analysis, and autohf for solutions. For autohf, the Coulomb interaction is represented by a Cholesky decomposition. Inset: zoomed in region showing the instability as the molecule dissociates. To highlight the versatility of autohf, beyond the set of lattice-like Hamiltonian terms that are implemented by default, we includ… view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

46 extracted references · 3 canonical work pages · 1 internal anchor

  1. [1]

    J. Cai, E. Anderson, C. Wang, X. Zhang, X. Liu, W . Holtzmann, Y. Zhang, F . Fan, T . Taniguchi, K. Watanabe, Y. Ran, T . Caoet al., Signatures of fractional quantum anoma- 27 SciPost Physics Codebases Submission lous Hall states in twisted MoTe 2, Nature 622(7981), 63 (2023), doi:10.1038 /s41586- 023-06289-w

    2023

  2. [2]

    Z. Lu, T . Han, Y. Yao, A. P . Reddy , J. Yang, J. Seo, K. Watanabe, T . Taniguchi, L. Fu and L. Ju, Fractional quantum anomalous Hall effect in multilayer graphene, Nature 626(8002), 759 (2024), doi:10.1038 /s41586-023-07010-7

    2024

  3. [3]

    Qin, C.-M

    M. Qin, C.-M. Chung, H. Shi, E. Vitali, C. Hubig, U. Schollwöck, S. R. White and S. Zhang, Absence of superconductivity in the pure two-dimensional Hubbard model , Phys. Rev. X 10(3), 031016 (2020), doi:10.1103 /PhysRevX.10.031016

    2020

  4. [4]

    R. J. Bartlett and M. Musiał, Coupled-cluster theory in quantum chemistry , Rev. Mod. Phys. 79(1), 291 (2007), doi:10.1103 /RevModPhys.79.291

    2007

  5. [5]

    Carleo and M

    G. Carleo and M. Troyer, Solving the quantum many-body problem with artificial neural networks, Science 355(6325), 602 (2017), doi:10.1126 /science.aag2302

    2017

  6. [6]

    Lange, A

    H. Lange, A. Van de Walle, A. Abedinnia and A. Bohrdt, From architectures to applica- tions: a review of neural quantum states, Quantum Science and Technology9(4), 040501 (2024), doi:10.1088 /2058-9565/ad7168

    2024

  7. [7]

    Becca and S

    F . Becca and S. Sorella,Quantum Monte Carlo Approaches for Correlated Systems, Cam- bridge University Press, ISBN 9781316417041, doi:10.1017 /9781316417041 (2017)

    2017

  8. [8]

    W . M. C. Foulkes, L. Mitas, R. J. Needs and G. Rajagopal, Quantum monte carlo simulations of solids , Reviews of Modern Physics 73(1), 33–83 (2001), doi:10.1103/revmodphys.73.33

    doi:10.1103/revmodphys.73.33 2001

  9. [9]

    Zhang, J

    S. Zhang, J. Carlson and J. E. Gubernatis, Constrained path Monte Carlo method for fermion ground states , Phys. Rev. B 55(12), 7464 (1997), doi:10.1103/PhysRevB.55.7464

    doi:10.1103/physrevb.55.7464 1997

  10. [10]

    Zhang and H

    S. Zhang and H. Krakauer, Quantum Monte Carlo method using phase-free ran- dom walks with Slater determinants , Phys. Rev. Lett. 90(13), 136401 (2003), doi:10.1103/PhysRevLett.90.136401

    doi:10.1103/physrevlett.90.136401 2003

  11. [11]

    Motta and S

    M. Motta and S. Zhang, Ab initio computations of molecular systems by the auxiliary-field quantum Monte Carlo method , WIREs Computational Molecular Science 8(5), e1364 (2018), doi:10.1002 /wcms.1364

    2018

  12. [12]

    H. Xu, H. Shi, E. Vitali, M. Qin and S. Zhang, Stripes and spin-density waves in the doped two-dimensional hubbard model: Ground state phase diagram , Physical Review Research 4(1) (2022), doi:10.1103 /physrevresearch.4.013239

    2022

  13. [13]

    Xu, C.-C

    J. Xu, C.-C. Chang, E. J. Walter and S. Zhang, Spin- and charge-density waves in the hartree–fock ground state of the two-dimensional hubbard model, Journal of Physics: Con- densed Matter 23(50), 505601 (2011), doi:10.1088 /0953-8984/23/50/505601

    2011

  14. [14]

    M. Qin, H. Shi and S. Zhang, Benchmark study of the two-dimensional hubbard model with auxiliary-field quantum monte carlo method , Phys. Rev. B 94, 085103 (2016), doi:10.1103/PhysRevB.94.085103

    work page doi:10.1103/physrevb.94.085103 2016

  15. [15]

    Zheng, C.-M

    B.-X. Zheng, C.-M. Chung, P . Corboz, G. Ehlers, M.-P . Qin, R. M. Noack, H. Shi, S. R. White, S. Zhang and G. K.-L. Chan, Stripe order in the underdoped region of the two-dimensional hubbard model, Science 358(6367), 1155 (2017), doi:10.1126/science.aam7127, https: //www.science.org/doi/pdf/10.1126/science.aam7127. 28 SciPost Physics Codebases Submission

    doi:10.1126/science.aam7127 2017

  16. [16]

    Q. Sun, X. Zhang, S. Banerjee, P . Bao, M. Barbry , N. S. Blunt, N. A. Bogdanov, G. H. Booth, J. Chen, Z.-H. Cui, J. J. Eriksen, Y. Gao et al. , Recent developments in the pyscf program package , The Journal of Chemical Physics 153(2), 024109 (2020), doi:10.1063 /5.0006074, https: //pubs.aip.org/aip/jcp/article-pdf/doi/10. 1063/5.0006074/16722275/024109_1...

    2020

  17. [18]

    Q. Sun, Libcint: An efficient general integral library for gaussian ba- sis functions , Journal of Computational Chemistry 36(22), 1664 (2015), doi:https://doi.org/10.1002/jcc.23981, https: //onlinelibrary .wiley .com/doi/pdf/ 10.1002/jcc.23981

    doi:10.1002/jcc.23981 2015

  18. [19]

    E. Aprà, E. J. Bylaska, W . A. de Jong, N. Govind, K. Kowalski, T . P . Straatsma, M. Va- liev, H. J. J. van Dam, Y. Alexeev, J. Anchell, V . Anisimov, F . W . Aquinoet al., Nwchem: Past, present, and future , The Journal of Chemical Physics 152(18), 184102 (2020), doi:10.1063/5.0004997

    doi:10.1063/5.0004997 2020

  19. [20]

    G. M. J. Barca, C. Bertoni, L. Carrington, D. Datta, N. De Silva, J. E. Deustua, D. G. Fedorov, J. R. Gour, A. O. Gunina, E. Guidez, T . Harville, S. Irleet al., Recent developments in the general atomic and molecular electronic structure system, The Journal of Chemical Physics 152(15), 154102 (2020), doi:10.1063 /5.0005188

    2020

  20. [21]

    T . D. Kühne, M. Iannuzzi, M. Del Ben, V . V . Rybkin, P . Seewald, F . Stein, T . Laino, R. Z. Khaliullin, O. Schütt, F . Schiffmann, D. Golze, J. Wilhelm et al. , Cp2k: An electronic structure and molecular dynamics software package - quickstep: Efficient and accurate electronic structure calculations , The Journal of Chemical Physics 152(19), 194103 (2...

    doi:10.1063/5.0007045 2020

  21. [22]

    Giannozzi, O

    P . Giannozzi, O. Andreussi, T . Brumme, O. Bunau, M. Buongiorno Nardelli, M. Calandra, R. Car, C. Cavazzoni, D. Ceresoli, M. Cococcioni, N. Colonna, I. Carnimeoet al., Advanced capabilities for materials modelling with quantum espresso, Journal of Physics: Condensed Matter 29(46), 465901 (2017), doi:10.1088 /1361-648X/aa8f79

    2017

  22. [23]

    Giannozzi, S

    P . Giannozzi, S. Baroni, N. Bonini, M. Calandra, R. Car, C. Cavazzoni, D. Ceresoli, G. L. Chiarotti, M. Cococcioni, I. Dabo, A. Dal Corso, S. de Gironcoli et al., Quantum espresso: a modular and open-source software project for quantum simulations of materi- als, Journal of Physics: Condensed Matter 21(39), 395502 (2009), doi:10.1088 /0953- 8984/21/39/395502

    2009

  23. [24]

    Gonze, B

    X. Gonze, B. Amadon, G. Antonius, F . Arnardi, L. Baguet, J.-M. Beuken, J. Bieder, F . Bot- tin, J. Bouchet, E. Bousquet, N. Brouwer, F . Bruneval et al., The abinit project: Impact, environment and recent developments, Comput. Phys. Commun. 248, 107042 (2020)

    2020

  24. [25]

    Tancogne-Dejean, M

    N. Tancogne-Dejean, M. J. T . Oliveira, X. Andrade, H. Appel, C. H. Borca, G. Le Bre- ton, F . Buchholz, A. Castro, S. Corni, A. A. Correa, U. De Giovannini, A. Delgado et al., Octopus, a computational framework for exploring light-driven phenomena and quantum dynamics in extended and finite systems , The Journal of Chemical Physics 152(12), 29 SciPost Ph...

    doi:10.1063/1.5142502 2020

  25. [26]

    Yoshikawa and M

    N. Yoshikawa and M. Sumita, Automatic differentiation for the direct minimization approach to the hartree–fock method , The Journal of Physical Chemistry A 126(45), 8487–8493 (2022), doi:10.1021 /acs.jpca.2c05922

    2022

  26. [27]

    MESS: Modern Electronic Structure Simulations

    H. Helal and A. Fitzgibbon, Mess: Modern electronic structure simulations , doi:10.48550/ARXIV .2406.03121 (2024)

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2406.03121 2024

  27. [28]

    Tamayo-Mendoza, C

    T . Tamayo-Mendoza, C. Kreisbeck, R. Lindh and A. Aspuru-Guzik,Automatic differentia- tion in quantum chemistry with applications to fully variational hartree–fock, ACS Central Science 4(5), 559–566 (2018), doi:10.1021 /acscentsci.7b00586

    2018

  28. [29]

    M. F . Kasim, S. Lehtola and S. M. Vinko, Dqc: A python program package for differentiable quantum chemistry , The Journal of Chemical Physics 156(8) (2022), doi:10.1063/5.0076202

    doi:10.1063/5.0076202 2022

  29. [30]

    J. M. Arrazola, S. Jahangiri, A. Delgado, J. Ceroni, J. Izaac, A. Száva, U. Azad, R. A. Lang, Z. Niu, O. Di Matteo, R. Moyard, J. Soni et al., Differentiable quantum computational chemistry with pennylane, doi:10.48550 /ARXIV .2111.09967 (2021)

    Pith/arXiv arXiv doi:10.48550/arxiv.2111.09967 2021

  30. [31]

    Bradbury , R

    J. Bradbury , R. Frostig, P . Hawkins, M. J. Johnson, C. Leary , D. Maclaurin, G. Necula, A. Paszke, J. VanderPlas, S. Wanderman-Milne and Q. Zhang, JAX: composable transfor- mations of Python+NumPy programs (2018)

    2018

  31. [32]

    D. C. Liu and J. Nocedal, On the limited memory bfgs method for large scale optimization, Mathematical Programming 45(1), 503 (1989), doi:10.1007 /BF01589116

    1989

  32. [33]

    D. J. Wales and J. P . K. Doye,Global optimization by basin-hopping and the lowest energy structures of lennard-jones clusters containing up to 110 atoms , The Journal of Physical Chemistry A 101(28), 5111 (1997), doi:10.1021/jp970984n, https://doi.org/10.1021/ jp970984n

    doi:10.1021/jp970984n 1997

  33. [34]

    H. Shi, P . Rosenberg, S. Chiesa and S. Zhang,Rashba spin-orbit coupling, strong interac- tions, and the bcs-bec crossover in the ground state of the two-dimensional fermi gas, Phys. Rev. Lett.117, 040401 (2016), doi:10.1103 /PhysRevLett.117.040401

    2016

  34. [35]

    Chiciak, E

    A. Chiciak, E. Vitali, H. Shi and S. Zhang, Magnetic orders in the hole-doped three-band hubbard model: Spin spirals, nematicity, and ferromagnetic domain walls, Phys. Rev. B97, 235127 (2018), doi:10.1103 /PhysRevB.97.235127

    2018

  35. [36]

    N. H. F . Beebe and J. Linderberg, Simplifications in the generation and transformation of two-electron integrals in molecular calculations , International Journal of Quantum Chemistry 12(4), 683 (1977), doi:https: //doi.org/10.1002/qua.560120408, https: //onlinelibrary .wiley .com/doi/pdf/10.1002/qua.560120408

    doi:10.1002/qua.560120408 1977

  36. [37]

    H. Koch, A. Sánchez de Merás and T . B. Pedersen,Reduced scaling in electronic structure calculations using cholesky decompositions , The Journal of Chemical Physics 118(21), 9481 (2003), doi:10.1063 /1.1578621, https: //pubs.aip.org/aip/jcp/article-pdf/118/ 21/9481/19024657/9481_1_online.pdf

    2003

  37. [38]

    N. Xiao, X. Liu and K.-C. Toh, Constraint dissolving approaches for Riemannian optimiza- tion, arXiv preprint arXiv:2203.10319 (2022). 30 SciPost Physics Codebases Submission

    Pith/arXiv arXiv 2022

  38. [39]

    X. Hu, N. Xiao, X. Liu and K.-C. Toh, A constraint dissolving approach for nonsmooth optimization over the Stiefel manifold, arXiv preprint arXiv:2205.10500 (2022)

    Pith/arXiv arXiv 2022

  39. [40]

    Thouless, Stability conditions and nuclear rotations in the hartree-fock theory, Nuclear Physics 21, 225 (1960), doi:https: //doi.org/10.1016/0029-5582(60)90048-1

    D. Thouless, Stability conditions and nuclear rotations in the hartree-fock theory, Nuclear Physics 21, 225 (1960), doi:https: //doi.org/10.1016/0029-5582(60)90048-1

    doi:10.1016/0029-5582(60)90048-1 1960

  40. [41]

    Sun, Libcint: An efficient general integral library for gaussian basis functions , Journal of Computational Chemistry 36(22), 1664–1671 (2015), doi:10.1002 /jcc.23981

    Q. Sun, Libcint: An efficient general integral library for gaussian basis functions , Journal of Computational Chemistry 36(22), 1664–1671 (2015), doi:10.1002 /jcc.23981

    2015

  41. [42]

    Q. Sun, T . C. Berkelbach, N. S. Blunt, G. H. Booth, S. Guo, Z. Li, J. Liu, J. D. McClain, E. R. Sayfutyarova, S. Sharma, S. Wouters and G. K. Chan,PYSCF: the python-based sim- ulations of chemistry framework, WIREs Computational Molecular Science 8(1) (2017), doi:10.1002/wcms.1340

    doi:10.1002/wcms.1340 2017

  42. [43]

    Sunet al., Recent developments in the PYSCF program package, The Journal of Chemical Physics 153(2) (2020), doi:10.1063 /5.0006074

    Q. Sunet al., Recent developments in the PYSCF program package, The Journal of Chemical Physics 153(2) (2020), doi:10.1063 /5.0006074

    2020

  43. [44]

    K. T . Williams, Y. Yao, J. Li, L. Chen, H. Shi, M. Motta, C. Niu, U. Ray , S. Guo, R. J. Ander- son, J. Li, L. N. Tranet al., Direct comparison of many-body methods for realistic electronic hamiltonians, Phys. Rev. X 10, 011041 (2020), doi:10.1103 /PhysRevX.10.011041

    2020

  44. [45]

    Shi and S

    H. Shi and S. Zhang, Many-body computations by stochastic sampling in hartree-fock- bogoliubov space, Phys. Rev. B 95, 045144 (2017), doi:10.1103 /PhysRevB.95.045144

    2017

  45. [46]

    Vitali, P

    E. Vitali, P . Rosenberg and S. Zhang, Monte carlo methods in the manifold of hartree–fock–bogoliubov wave functions , The Journal of Chemical Physics 161(13), 134116 (2024), doi:10.1063/5.0228450, https://pubs.aip.org/aip/jcp/article-pdf/doi/ 10.1063/5.0228450/20194854/134116_1_5.0228450.pdf

    work page doi:10.1063/5.0228450 2024

  46. [47]

    Scholle, P

    R. Scholle, P . M. Bonetti, D. Vilardi and W . Metzner,Comprehensive mean-field analysis of magnetic and charge orders in the two-dimensional hubbard model , Phys. Rev. B 108, 035139 (2023), doi:10.1103 /PhysRevB.108.035139. 31

    2023

This paper was first reviewed by deepseek-v4-flash on August 2, 2026.