Machine Learning Hamiltonians are Accurate Energy-Force Predictors

ChangYoung Park; Chanhui Lee; Honghui Kim; Nayoung Kim; Sehui Han; Seongjun Yun; Seongsu Kim; Sungbin Lim; Sungsoo Ahn; Yoonho Kim

REVIEW 3 major objections 4 minor 9 references

Under direct evaluation, a Hamiltonian-predicting model (QHFlow2) matches leading force-field accuracy on forces and beats it up to 20× on energies — making predicted Hamiltonians viable energy–force predictors.

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 ↗

T0 review · deepseek-v4-flash

2026-08-02 22:24 UTC pith:HB5DRYEH

load-bearing objection The direct-evaluation benchmark is a real contribution and the model results look strong, but the force-evaluation pipeline is underspecified in exactly the place the headline claim rests. the 3 major comments →

arxiv 2602.16897 v2 pith:HB5DRYEH submitted 2026-02-18 physics.comp-ph cond-mat.mtrl-sci

Machine Learning Hamiltonians are Accurate Energy-Force Predictors

Seongsu Kim , Chanhui Lee , Yoonho Kim , Seongjun Yun , Honghui Kim , Nayoung Kim , Changyoung Park , Sehui Han

show 2 more authors

Sungbin Lim Sungsoo Ahn

This is my paper

classification physics.comp-ph cond-mat.mtrl-sci MSC 68T0781V55 PACS 71.15.Mb31.15.Ew07.05.Mh

keywords machine learning HamiltoniansKohn–Sham Hamiltonian predictionenergy and force predictionequivariant neural networksflow matchingMD17/rMD17 benchmarkQH9 benchmarkinteratomic potential comparison

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.

Machine-learning Hamiltonian models predict the quantum-chemical matrix that encodes a molecule's electronic structure, but they have been judged almost entirely by how well they reconstruct that matrix — not by whether the energies and forces derived from it are physically accurate. This paper asks whether such models can work as direct energy–force predictors, and it builds a benchmark that pushes each predicted Hamiltonian through the Roothaan–Hall equation into a density, a Kohn–Sham energy, and analytic forces. Within that benchmark, the paper's new model, QHFlow2 — a flow-matching network with an SO(2)-equivariant backbone and a two-stage edge update — is the first Hamiltonian model to reach the force accuracy of a leading interatomic-potential model on MD17/rMD17, while reporting up to 20× lower energy errors; on QH9 its energy errors are up to 20× smaller than a leading machine-learning force field. It also reports 40% lower Hamiltonian reconstruction error than the previous best Hamiltonian model with roughly half the parameters, and shows that Hamiltonian, energy, and force errors fall together as model size and data grow. If the benchmark is sound, the payoff is a single model that supplies accurate interatomic forces and, from the same forward pass, orbital energies, densities, and reactivity descriptors — quantities force-field models cannot provide. The authors note the current scope is restricted closed-shell molecules under a fixed functional and basis, with evaluation cost still dominated by the density-functional step.

Core claim

The paper's central claim is that an accurate predicted Hamiltonian, evaluated directly and without self-consistent refinement, yields total energies and atomic forces that are competitive with the best machine-learning interatomic potentials. The pipeline is: predict the Kohn–Sham Hamiltonian Ĥ in an atom-centered orbital basis from the molecular geometry; solve the Roothaan–Hall equation ĤC = SCϵ against the fixed reference overlap matrix S with a simple occupation rule; build the density matrix and evaluate the Kohn–Sham energy functional; take forces as analytic nuclear gradients. On the MD17/rMD17 trajectories, QHFlow2 reports force mean absolute errors between 0.34 and 1.35 meV/Å acros

What carries the argument

The load-bearing object is the predicted Kohn–Sham Hamiltonian matrix Ĥ in an atom-centered orbital basis, and the identity that carries the argument is the Roothaan–Hall equation ĤC = SCϵ — the generalized eigenproblem that, with the geometry-fixed overlap matrix S, converts the predicted matrix into orbital coefficients, a density matrix, a Kohn–Sham energy, and analytic forces. The architecture that makes the prediction accurate enough for this shortcut is QHFlow2's combination of an SO(2)-equivariant message-passing backbone — efficient edge updates carried out in local reference frames while keeping an SO(3)-equivariant feature structure — with a two-stage pair update: pairwise features

Load-bearing premise

The central comparison assumes that evaluating the Kohn–Sham energy functional — and especially its analytic nuclear gradients — on a density built from a predicted, non-self-consistent Hamiltonian is a well-defined, correctly implemented prescription for physical energies and forces; the paper specifies the pipeline in Section 3.1 but defers the gradient details to a formula in Appendix A.1 that assumes SCF stationarity, leaving the non-stationary case unspecified.

What would settle it

Take the released code and several MD17 test molecules: for a fixed predicted Hamiltonian, compute the reported analytic forces and compare them with finite differences of the reported energy under the identical non-self-consistent prescription; disagreement beyond numerical tolerance would show the force numbers do not measure what they claim. Second check: retrain the interatomic-potential baselines on the exact QH9-stable-id split used for QHFlow2 — the paper's QH9 comparison uses published numbers from a different split — and see whether the up-to-20× energy margin survives a strict head-t

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

If this is right

Hamiltonian models become directly usable as energy–force predictors: a predicted matrix alone, without any SCF refinement, yields MD-grade forces and accurate energies, giving MLH models a concrete accuracy reference to be held to.
A single model now provides both interatomic forces and electronic-structure observables — occupied orbital energies, HOMO–LUMO gaps, densities, and reactivity descriptors — from the same forward pass, which interatomic-potential models cannot do.
Hamiltonian, energy, and force errors decrease together with model size and training data, so the measured scaling curves imply that more parameters and data will continue to improve downstream simulation accuracy.
Accurate predicted Hamiltonians also accelerate iterative DFT: used as initial guesses, they cut SCF iterations by roughly half or more on QH9 relative to the default start, so the same model can speed up both direct evaluation and converged calculations.
The released MD17/rMD17 benchmark with recomputed, internally consistent Hamiltonians gives future work a controlled, split-matched setting for comparing Hamiltonian predictors against interatomic potentials.

Where Pith is reading between the lines

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

The implied displacement is bigger than the headline: if predicted Hamiltonians are this accurate without self-consistency, then SCF iteration — the main cost of DFT — becomes partially optional for energy and force estimation; a natural test is to measure how energy and force error degrade as the predicted Hamiltonian is deliberately perturbed away from self-consistency.
The reactivity-stretching demonstration (frontier-orbital energies under bond distortion) points to a use case force-field models cannot serve: replacing DFT in reaction-coordinate and descriptor screening. A concrete extension is full reaction-path scans comparing predicted barrier heights and HOMO–LUMO trends against DFT references across many bonds.
The scaling curves invite an extrapolation the authors do not make: if Hamiltonian error continues to fall as a power law in parameters and data, the direct-evaluation route could overtake retrained interatomic potentials at scale, because it produces forces, energies, and electronic structure from one model — worth testing on larger molecular diversity than the six MD17 molecules.
The fixed-functional, fixed-basis training and closed-shell scope suggest a cheap transfer check: test the direct-evaluation pipeline across basis sets (e.g., def2-SVP versus def2-TZVP) to see whether Hamiltonian accuracy transfers or whether conditioning on the basis is required — the authors list this as open future work.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

The direct-evaluation benchmark is a real contribution and the model results look strong, but the force-evaluation pipeline is underspecified in exactly the place the headline claim rests.

read the letter

Two things to know. First, this paper makes a useful shift in how ML Hamiltonians are judged: instead of only matrix-reconstruction metrics, they compute energies and forces from the predicted H through a fixed pipeline and compare to MLIP baselines. That benchmark is genuinely new. Second, the central comparison—QHFlow2 reaching NequIP-level forces—depends on an analytic-gradient implementation the manuscript does not specify. If the forces are computed from a non-self-consistent density without the density-matrix response terms, the numbers may not mean what the abstract says.

What's good: the architecture work is credible. QHFlow2 uses an SO(2) backbone with a two-stage pair update, and the MD17/rMD17 experiments are controlled—they retrained NequIP, GemNet-T, DimeNet and the Hamiltonian baselines on the same splits and labels. The scaling plots are clean. The 40% Hamiltonian error reduction with fewer parameters is a concrete, reproducible-looking claim, and the code is linked. The paper is clearly written and does not oversell the SCF acceleration results.

Where it's soft. The force evaluation is the load-bearing part. Section 3.1 says they solve Roothaan–Hall with H_pred and S, form D, then evaluate E_KS and analytic gradients. Appendix A.1 gives the standard Hellmann–Feynman/Pulay expression but never says how the response of D is handled when D is not SCF-stationary. PySCF's RKS gradient assumes stationarity. If they call grad.RKS on a non-converged density, the output is not the gradient of E[D]. They need to state whether they include response terms or whether they re-converge SCF (which would change what 'direct' means). I did not see this in the main text or appendix; the linked code might resolve it, but a referee should check.

Secondary issues: the QH9 comparison to MACE/Equiformer uses published QM9 numbers from different DFT settings; that's fine as a reference point but shouldn't be read as a controlled comparison. HELM is missing from the main Hamiltonian tables. And Table 12 claims prior models 'fail' without the initial Hamiltonian but doesn't report their energy/force errors—just dashes.

Net: the benchmark and architecture are worth taking seriously. The paper deserves a serious referee, but the referee needs to pin down the gradient implementation before the headline claim can be trusted. I'd want a revision with an explicit derivation or a code inspection.

Referee Report

3 major / 4 minor

Summary. The paper proposes QHFlow2, an SO(2)-equivariant flow-matching model for predicting Kohn–Sham Hamiltonians, and introduces a benchmark that computes total energies and analytic forces directly from predicted Hamiltonians. The main claims are that on MD17/rMD17 QHFlow2 is the first Hamiltonian model to reach NequIP-level force accuracy while achieving up to 20× lower energy MAE, and on QH9 it reduces energy error by up to 20× compared to MACE. Additional contributions include a two-stage pair update, a recomputed rMD17 benchmark for controlled comparison, scaling studies, and SCF-initialization experiments.

Significance. If the direct-evaluation pipeline is implemented correctly and the comparisons are fair, the paper would provide strong evidence that MLH models are practically viable energy–force predictors, not just Hamiltonian reconstruction tools. The paper's controlled MD17/rMD17 setup—retraining all baselines on the same splits and recomputing reference Hamiltonians—is a methodological strength, as are the scaling and SCF-cycle experiments. The architecture improvements are clearly described and appear to give substantial accuracy gains. However, the central force-accuracy claim depends on an unspecified non-SCF analytic-gradient implementation, and the QH9 energy comparison against published QM9 numbers is not apples-to-apples. These issues need to be resolved before the headline claims can be accepted.

major comments (3)

[Section 3.1 and Appendix A.1] The force evaluation is not specified for a non-self-consistent density matrix. After solving the Roothaan–Hall equation with a predicted Hamiltonian, the density matrix D is not SCF-stationary, so the gradient of E_KS[rho] contains terms proportional to dD/dR that do not vanish. Appendix A.1 gives only the standard Hellmann–Feynman/Pulay expression for SCF-converged densities and never describes how the density-matrix response is treated. Standard PySCF gradient routines assume SCF stationarity; applying them to a non-SCF D would give forces that are not the true gradient of E_KS[rho]. The authors must specify whether they (a) rerun SCF before differentiation, (b) solve coupled-perturbed response equations for the non-SCF D, or (c) use only Hellmann–Feynman terms, and they should numerically verify against finite differences. This is the load-bearing concern because the headline 'NequIP
[Section 4.1 and Table 15 / Figure 3] The QH9 energy comparisons against MACE, Equiformer, and EquiformerV2 use published QM9 results from different DFT settings, QH9 labels being B3LYP/def2-SVP, while the published QM9 numbers likely use a different functional/basis combination. Reporting energy MAE in meV and calculating ratios such as '20× lower than MACE' from these numbers is not valid without correcting for the reference electronic-structure settings or retraining/evaluating the MLIPs on QH9. The abstract's QH9 claim is therefore not supported as stated. The authors should either recompute MLIP baselines under the same DFT setup or clearly present the QH9 comparison as cross-references that are illustrative, not quantitative.
[Section G.1, Table 12] The text states that prior Hamiltonian predictors 'fail to produce accurate energies when the initial Hamiltonian is removed,' but Table 12 reports no energy or force values for SchNOrb, PhiSNet, QHNet, or HELM—only Hamiltonian and orbital metrics, with dashes for downstream quantities. This claim is therefore unsupported by the presented data. To make this point, the authors need to run those baselines through the same downstream evaluation pipeline without an initial Hamiltonian, or soften the claim accordingly.

minor comments (4)

[Throughout] Typos and inconsistent naming: 'Traininig', 'comparion', 'Dimenet', 'Nequip', 'W ANet', 'V on Lilienfeld' (the latter appears in a citation). A careful proofread is needed.
[Section 3.1, Eq. (3)] The occupation rule uses a simple eigenvalue ordering without considering near-degeneracies or level crossings. The text should clarify how orbital ordering is resolved for the predicted Hamiltonian, particularly for molecules with near-degenerate frontier orbitals.
[Section E.1] The description of rMD17 says 'we recompute reference quantities with PySCF at the PBE/def2-SVP level,' and in the main text the authors claim consistency with the original rMD17 labels. If the original rMD17 labels were generated with ORCA, the two are not necessarily identical; please clarify whether the benchmark uses only the recomputed labels or a mix of original and recomputed labels.
[Table 14] The comparison between MLIPs and QHFlow2 on energy MAE is striking (e.g., 0.003 meV vs 0.33 meV for NequIP on ethanol). Since energy MAE is dominated by per-atom reference-energy shifts, the authors should state whether both models are evaluated on exactly the same energy target normalization; otherwise the large apparent advantage might partly reflect different target definitions.

Circularity Check

0 steps flagged

No significant circularity: downstream energy/force metrics are computed through a fixed physical pipeline from predicted Hamiltonians, not fitted or derived from the metrics themselves.

full rationale

QHFlow2 is trained to minimize the flow-matching objective in Eq. (7), which regresses predicted Hamiltonians against reference Hamiltonians; energies and forces are never used as training targets. The downstream evaluation in Section 3.1 solves the Roothaan–Hall equation with the predicted Hamiltonian and overlap matrix, forms a density matrix, and evaluates the KS energy functional and analytic gradients. This is a deterministic post-processing pipeline, so the reported energy/force errors are empirical consequences of Hamiltonian accuracy, not quantities fitted into the model. The only self-citation of note is Kim et al. (2025) for QHFlow and the TE prior; QHFlow is used as a baseline, and the flow-matching/TE prior choices are training and architectural components, not invoked as proof of the headline claim. There is no imported uniqueness theorem, no ansatz justified solely by a same-author citation, and no renaming of a known result. The paper's Section 5 honestly states limitations (fixed functional/basis, closed-shell restriction, solver-dominated runtime), and Appendix A.1 leaves the non-SCF analytic-gradient implementation underspecified; that is a reproducibility/correctness concern and does not constitute circularity because no reported prediction is equivalent to its input by construction. The scaling analyses in Figures 5–6 are independent empirical checks. Score 0.

Axiom & Free-Parameter Ledger

3 free parameters · 4 axioms · 0 invented entities

The model itself has learned weights (not enumerated), but the central claims also rest on hand-chosen hyperparameters (cutoff, prior, steps) and on the DFT pipeline being treated as ground truth. No new physical entities are introduced.

free parameters (3)

SO(2) cutoff radius = 5.0 Å (default)
Graph neighborhood cutoff for the backbone and pair initialization; ablation shows off-diagonal Hamiltonian error is sensitive to this choice, and the reference-frame cue gref is introduced to reduce that sensitivity (Table 10).
TE prior radial distribution = LogNormal(1, 0.1)
The rotation-invariant flow-matching prior samples radial magnitudes from LogNormal(1,0.1); chosen by hand and not derived from data.
Flow sampling steps = 3
Number of ODE steps at inference; ablation shows 1–3 steps are sufficient, but 3 is the default for reported metrics.

axioms (4)

domain assumption Kohn-Sham DFT with the chosen functional/basis (PBE or B3LYP, def2-SVP) provides ground-truth Hamiltonians, energies, and forces.
All reference labels come from DFT; the benchmark inherits DFT's approximations.
domain assumption The density matrix D = C diag(o) C^T from the generalized eigenproblem with the predicted Hamiltonian approximates the ground-state density well enough for energy/force evaluation.
Invoked in Section 3.1; the energy functional is variational, so this is an upper-bound energy, but the gradient w.r.t. nuclei is not guaranteed to be the physical force.
standard math The Hamiltonian matrix transforms as an intertwiner under SO(3), enabling SO(3)-equivariant prediction.
Section B.3, Eq. (45); standard representation theory.
standard math Conditional flow matching (Lipman et al., 2023) provides a valid training objective for Hamiltonian prediction.
Used in Section 3.2, Eq. (7); prior method adopted as-is.

pith-pipeline@v1.3.0-alltime-deepseek · 31947 in / 16377 out tokens · 130759 ms · 2026-08-02T22:24:08.703726+00:00 · methodology

0 comments

read the original abstract

Recently, machine learning Hamiltonian (MLH) models have gained traction as fast approximations of electronic structures such as orbitals and electron densities, while also enabling direct evaluation of energies and forces from their predictions. However, despite their physical grounding, existing Hamiltonian models are evaluated mainly by reconstruction metrics, leaving it unclear how well they perform as energy-force predictors. We address this gap with a benchmark that computes energies and forces directly from predicted Hamiltonians. Within this framework, we propose QHFlow2, a state-of-the-art Hamiltonian model with an SO(2)-equivariant backbone and a two-stage edge update. QHFlow2 achieves $40\%$ lower Hamiltonian error than the previous best model with fewer parameters. Under direct evaluation on MD17/rMD17, it is the first Hamiltonian model to reach NequIP-level force accuracy while achieving up to $20\times$ lower energy MAE. On QH9, QHFlow2 reduces energy error by up to $20\times$ compared to MACE. Finally, we demonstrate that QHFlow2 exhibits consistent scaling behavior with respect to model capacity and data, and that improvements in Hamiltonian accuracy effectively translate into more accurate energy and force computations.

Figures

Figures reproduced from arXiv: 2602.16897 by ChangYoung Park, Chanhui Lee, Honghui Kim, Nayoung Kim, Sehui Han, Seongjun Yun, Seongsu Kim, Sungbin Lim, Sungsoo Ahn, Yoonho Kim.

**Figure 1.** Figure 1: QHFlow2 overall workflow. Given the molecular structure M, flow time t, and an intermediate Hamiltonian state Ht, QHFlow2 applies an SO(2) backbone, a two-stage pairwise update, and construct Hamiltonian via tensor expansion for energy and force evaluation. (a) Pipeline overview. (b) Detailed architecture of the matrix encode, the SO(2) backbone, and the two-stage pair update. SO(2)-equivariant message-pas… view at source ↗

**Figure 2.** Figure 2: Energy and force accuracy under direct evaluation on the MD benchmark. We report mean absolute errors (MAE) of total energy (top) and forces (bottom) computed from predicted Hamiltonians on six molecular systems. gray bars denote MLIP baselines, green bars denote prior Hamiltonian predictors, and blue bars denote QHFlow2. All methods use the same data splits and evaluation setup. Overall, the results show … view at source ↗

**Figure 4.** Figure 4: Relative reduction of SCF cycles on QH9. We report the SCF cycle ratio (lower is better): SCF iterations to converge when initializing with a predicted Hamiltonian Hˆ , normalized by the MinAO baseline (100%). All SCF/DFT settings are fixed; only the initialization is changed. Stronger predictors reduce SCF cycles and approach REF, which initializes with the dataset Hamiltonian H (data limit), while RC ini… view at source ↗

**Figure 5.** Figure 5: Data scaling on the MD benchmark. (a) Energy and force MAE versus training set size for QHFlow2-m and MLIP baselines on salicylic acid and aspirin. (b) Hamiltonian MAE versus training set size for QHFlow2-m. (c) Occupied orbital energy MAE versus training set size for QHFlow2-m. Improved Hamiltonian accuracy consistently accompanies improved orbital, energy, and force accuracy. pacity on QH9 and includes a… view at source ↗

**Figure 7.** Figure 7: Inference speed and memory scaling on QH9. (left) inference speed (samples/s) versus parameter count. (right) peak inference GPU memory (GB) versus parameter count. posed QHFlow2, a scalable MLH model that combines an SO(2) backbone with a two-stage pair update that reduces cutoff sensitivity. Across datasets, improved Hamiltonian accuracy consistently translated into lower downstream energy and force err… view at source ↗

**Figure 8.** Figure 8: Robustness of Frontier Molecular Orbital (FMO) Predictions under Reactive Bond Stretching. The plots illustrate the Mean Absolute Error (MAE) of calculated HOMO and LUMO energies from Hamiltonian predictors (i.e. QHFlow2 and SPHNet), relative to DFT value as the reactive bond length is distorted from 0.95× to 1.15× the equilibrium length. (a) Performance on the reactive ester carbonyl bond (C11=O12) of Asp… view at source ↗

discussion (0)

Reference graph

Works this paper leans on

9 extracted references · 1 canonical work pages

[42]

URL https://www.sciencedirect.com/ science/article/pii/S2589004224008952

doi: https://doi.org/10.1016/j.isci.2024.109673. URL https://www.sciencedirect.com/ science/article/pii/S2589004224008952. 29 Weigend, F. and Ahlrichs, R. Balanced basis sets of split valence, triple zeta valence and quadruple zeta valence quality for h to rn: Design and assessment of accuracy. Physical Chemistry Chemical Physics, 7(18):3297–3305,

arXiv 2024
[205]

16 Gong, X., Li, H., Zou, N., Xu, R., Duan, W., and Xu, Y

Elsevier, 1994. 16 Gong, X., Li, H., Zou, N., Xu, R., Duan, W., and Xu, Y . General framework for e (3)-equivariant neural network representation of density functional theory hamiltonian. Nature Communications, 14(1):2848, 2023. 1, 2 Hall, G. G. The molecular orbital theory of chemical va- lency viii. a method of calculating ionization potentials. Proceed...

arXiv 1994
[1951]

3 Ruddigkeit, L., Van Deursen, R., Blum, L

doi: https://doi.org/10.1103/revmodphys.23.69. 3 Ruddigkeit, L., Van Deursen, R., Blum, L. C., and Reymond, J.-L. Enumeration of 166 billion organic small molecules in the chemical universe database gdb-17.Journal of chemical information and modeling, 52(11):2864–2875,

doi:10.1103/revmodphys.23.69
[2000]

P., Simm, G., Ortner, C., and Cs´anyi, G

13 Batatia, I., Kovacs, D. P., Simm, G., Ortner, C., and Cs´anyi, G. Mace: Higher order equivariant message passing neural networks for fast and accurate force fields.Ad- vances in neural information processing systems, 35: 11423–11436, 2022. 2, 5 Batzner, S., Musaelian, A., Sun, L., Geiger, M., Mailoa, J. P., Kornbluth, M., Molinari, N., Smidt, T. E., an...

Pith/arXiv arXiv 2022
[2005]

Elsevier, 2012

21 Wigner, E.Group theory: and its application to the quantum mechanics of atomic spectra, volume 5. Elsevier, 2012. 15 Wu, X., Sun, Q., Pu, Z., Zheng, T., Ma, W., Yan, W., Yu, X., Wu, Z., Huo, M., Li, X., Ren, W., Gong, S., Zhang, Y ., and Gao, W. Enhancing gpu-acceleration in the python- based simulations of chemistry framework, 2024. URL https://arxiv....

Pith/arXiv arXiv 2012
[2012]

An undulatory theory of the mechanics of atoms and molecules.Physical review, 28(6):1049, 1926

5, 21 Schr¨odinger, E. An undulatory theory of the mechanics of atoms and molecules.Physical review, 28(6):1049, 1926. 12 Sch¨utt, K., Gastegger, M., Tkatchenko, A., M ¨uller, K.-R., and Maurer, R. Unifying machine learning and quantum chemistry with a deep neural network for molecular wave- functions.Nature Communications, 10(1), 2019. 1, 2, 5, 21 Slater...

Pith/arXiv arXiv 1926
[2017]

Deep-learning density functional theory hamiltonian for efficient ab initio electronic-structure cal- culation.Nature Computational Science, 2(6):367–377,

23 Li, H., Wang, Z., Zou, N., Ye, M., Xu, R., Gong, X., Duan, W., and Xu, Y . Deep-learning density functional theory hamiltonian for efficient ab initio electronic-structure cal- culation.Nature Computational Science, 2(6):367–377,
[2021]

Recent advances in machine learning- assisted multiscale design of energy materials

29 Mortazavi, B. Recent advances in machine learning- assisted multiscale design of energy materials. Advanced Energy Materials, 15(9):2403876, 2025. doi: https://doi.org/10.1002/aenm.202403876. URL https://advanced.onlinelibrary.wiley. com/doi/abs/10.1002/aenm.202403876. 29 Mortazavi, B., Zhuang, X., Rabczuk, T., and Shapeev, A. V . Atomistic modeling of...

work page doi:10.1002/aenm.202403876 2025
[2022]

Enhancing the scalability and applicability of kohn-sham hamiltonians for molecular systems

1, 2 Li, Y ., Xia, Z., Huang, L., Wei, X., Harshe, S., Yang, H., Luo, E., Wang, Z., Zhang, J., Liu, C., et al. Enhancing the scalability and applicability of kohn-sham hamiltonians for molecular systems. InThe Thirteenth International Conference on Learning Representations, 2025. 1, 2, 5 Liao, Y .-L. and Smidt, T. Equiformer: Equivariant graph attention t...

2025

[1] [42]

URL https://www.sciencedirect.com/ science/article/pii/S2589004224008952

doi: https://doi.org/10.1016/j.isci.2024.109673. URL https://www.sciencedirect.com/ science/article/pii/S2589004224008952. 29 Weigend, F. and Ahlrichs, R. Balanced basis sets of split valence, triple zeta valence and quadruple zeta valence quality for h to rn: Design and assessment of accuracy. Physical Chemistry Chemical Physics, 7(18):3297–3305,

arXiv 2024

[2] [205]

16 Gong, X., Li, H., Zou, N., Xu, R., Duan, W., and Xu, Y

Elsevier, 1994. 16 Gong, X., Li, H., Zou, N., Xu, R., Duan, W., and Xu, Y . General framework for e (3)-equivariant neural network representation of density functional theory hamiltonian. Nature Communications, 14(1):2848, 2023. 1, 2 Hall, G. G. The molecular orbital theory of chemical va- lency viii. a method of calculating ionization potentials. Proceed...

arXiv 1994

[3] [1951]

3 Ruddigkeit, L., Van Deursen, R., Blum, L

doi: https://doi.org/10.1103/revmodphys.23.69. 3 Ruddigkeit, L., Van Deursen, R., Blum, L. C., and Reymond, J.-L. Enumeration of 166 billion organic small molecules in the chemical universe database gdb-17.Journal of chemical information and modeling, 52(11):2864–2875,

doi:10.1103/revmodphys.23.69

[4] [2000]

P., Simm, G., Ortner, C., and Cs´anyi, G

13 Batatia, I., Kovacs, D. P., Simm, G., Ortner, C., and Cs´anyi, G. Mace: Higher order equivariant message passing neural networks for fast and accurate force fields.Ad- vances in neural information processing systems, 35: 11423–11436, 2022. 2, 5 Batzner, S., Musaelian, A., Sun, L., Geiger, M., Mailoa, J. P., Kornbluth, M., Molinari, N., Smidt, T. E., an...

Pith/arXiv arXiv 2022

[5] [2005]

Elsevier, 2012

21 Wigner, E.Group theory: and its application to the quantum mechanics of atomic spectra, volume 5. Elsevier, 2012. 15 Wu, X., Sun, Q., Pu, Z., Zheng, T., Ma, W., Yan, W., Yu, X., Wu, Z., Huo, M., Li, X., Ren, W., Gong, S., Zhang, Y ., and Gao, W. Enhancing gpu-acceleration in the python- based simulations of chemistry framework, 2024. URL https://arxiv....

Pith/arXiv arXiv 2012

[6] [2012]

An undulatory theory of the mechanics of atoms and molecules.Physical review, 28(6):1049, 1926

5, 21 Schr¨odinger, E. An undulatory theory of the mechanics of atoms and molecules.Physical review, 28(6):1049, 1926. 12 Sch¨utt, K., Gastegger, M., Tkatchenko, A., M ¨uller, K.-R., and Maurer, R. Unifying machine learning and quantum chemistry with a deep neural network for molecular wave- functions.Nature Communications, 10(1), 2019. 1, 2, 5, 21 Slater...

Pith/arXiv arXiv 1926

[7] [2017]

Deep-learning density functional theory hamiltonian for efficient ab initio electronic-structure cal- culation.Nature Computational Science, 2(6):367–377,

23 Li, H., Wang, Z., Zou, N., Ye, M., Xu, R., Gong, X., Duan, W., and Xu, Y . Deep-learning density functional theory hamiltonian for efficient ab initio electronic-structure cal- culation.Nature Computational Science, 2(6):367–377,

[8] [2021]

Recent advances in machine learning- assisted multiscale design of energy materials

29 Mortazavi, B. Recent advances in machine learning- assisted multiscale design of energy materials. Advanced Energy Materials, 15(9):2403876, 2025. doi: https://doi.org/10.1002/aenm.202403876. URL https://advanced.onlinelibrary.wiley. com/doi/abs/10.1002/aenm.202403876. 29 Mortazavi, B., Zhuang, X., Rabczuk, T., and Shapeev, A. V . Atomistic modeling of...

work page doi:10.1002/aenm.202403876 2025

[9] [2022]

Enhancing the scalability and applicability of kohn-sham hamiltonians for molecular systems

1, 2 Li, Y ., Xia, Z., Huang, L., Wei, X., Harshe, S., Yang, H., Luo, E., Wang, Z., Zhang, J., Liu, C., et al. Enhancing the scalability and applicability of kohn-sham hamiltonians for molecular systems. InThe Thirteenth International Conference on Learning Representations, 2025. 1, 2, 5 Liao, Y .-L. and Smidt, T. Equiformer: Equivariant graph attention t...

2025