Pith. sign in

REVIEW 2 major objections 5 minor 49 references

Hot-Ham predicts DFT Hamiltonians with the best reported mean absolute errors by replacing Clebsch–Gordan tensor products with Gaunt tensor products, plus one final SO(2) layer to restore antisymmetric parity.

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

2026-08-05 05:48 UTC pith:474KO3WF

load-bearing objection Hot-Ham is a genuinely useful efficiency combination for equivariant Hamiltonian models, with strong benchmark numbers and an honest but untested design assumption that needs an ablation. the 2 major comments →

arxiv 2509.04875 v1 pith:474KO3WF submitted 2025-09-05 physics.comp-ph cond-mat.mtrl-sci

Hot-Ham: an accurate and efficient E(3)-equivariant machine-learning electronic structures calculation framework

Zhixin Liang , Yunlong Wang , Chi Ding , Junjie Wang , Hui-Tian Wang , Dingyu Xing , Jian Sun This is my paper
classification physics.comp-ph cond-mat.mtrl-sci
keywords E(3)-equivariant neural networkmachine-learning HamiltonianGaunt tensor productDFT Hamiltonian predictionSO(2) convolutionelectronic structuremessage passing neural networktwisted moiré materials
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.

Hot-Ham is an E(3)-equivariant message-passing network that predicts density-functional-theory Hamiltonians from atomic positions and species. Its central claim is that the expensive Clebsch–Gordan tensor product can be replaced by Gaunt tensor products, computed in Fourier or sphere-grid bases with fast Fourier transforms, while a single final SO(2)-convolution layer restores the antisymmetric tensor components that GTP cannot generate. On standard benchmarks this yields the best reported Hamiltonian mean absolute errors of 0.08, 0.12, and 0.15 meV for monolayer graphene, monolayer MoS2, and bilayer graphene, using fewer than two million parameters. The method also generalizes to twist angles, heterostructure sizes, and crystal allotropes not seen in training, so the practical payoff is DFT-quality electronic structure at a small fraction of the cost.

Core claim

Hot-Ham establishes that a Hamiltonian-learning equivariant network can get its high-order equivariant features from the Gaunt tensor product rather than full Clebsch–Gordan coupling, provided one CGTP step is kept at the last layer. The Gaunt tensor product expresses the product of two spherical-harmonic features as a pointwise multiplication in a Fourier or sphere-grid basis, reducing the cost of the tensor product from O(L^6) to O(L^3) or O(L^2 log^2 L). Because Gaunt coefficients vanish for odd l1+l2+l3, GTP cannot produce antisymmetric tensors, so Hot-Ham introduces those exactly once, through an SO(2) convolution after local coordinate alignment. The resulting features are read out as

What carries the argument

The central object is the Gaunt tensor product (GTP): instead of directly summing Clebsch–Gordan coefficients, it multiplies the spherical functions associated with two equivariant features in a spectral or grid basis and converts the product back to spherical-harmonic coefficients, enabling FFT acceleration. GTP carries the bulk of the message passing. The complementary mechanism is a local-coordinate SO(2) convolution in the final layer, which is the only place antisymmetric tensors enter; it rotates features so filters become m=0, turning the CGTP into a cheaper 2D operation. Together they enforce E(3) equivariance while keeping the feature set at O(L^3) or O(L^2 log^2 L).

Load-bearing premise

The design relies on the assumption that one last layer of the expensive kind of coupling is enough to supply every parity-odd quantity the Hamiltonian needs, and the paper does not test whether more such layers would change the answer.

What would settle it

Train Hot-Ham with zero, one, or several final parity-restoring layers on a dataset where the sign of hopping terms under inversion matters (for example, twisted bilayers) and compare Hamiltonian errors; if the one-layer model is not as accurate as the multi-layer model, the central design assumption fails.

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

Share X Bluesky LinkedIn Reddit HN

If this is right

  • Hamiltonian prediction for large systems—the paper demonstrates inference on structures with over 10^4 atoms—becomes practical on a single CPU or GPU node.
  • Models trained on small, untwisted supercells transfer to twisted multilayers, arbitrary moiré mismatches, and allotropes outside the training set, so band structures of hypothetical materials can be screened without new DFT runs.
  • The parameter count stays under 2M while beating larger models, which lowers training cost and improves data efficiency.
  • Because the Hamiltonian is predicted directly, band structures, DOS, and derived quantities follow by diagonalization, and the paper notes forces and electron–phonon couplings become accessible via automatic differentiation.

Where Pith is reading between the lines

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

  • The same GTP-plus-one-CGTP design pattern should apply to other equivariant targets that need odd-parity high-order features, such as spin–orbit-coupled or magnetic Hamiltonians, where the single SO(2) layer may need to be revisited.
  • The strongest unproven spot is the sufficiency of one final antisymmetric layer; an ablation that moves CGTP layers earlier or adds more of them would directly test whether the parity bottleneck is real.
  • The complexity reduction is asymptotic and assumes the FFT/grid operations dominate; for practical L values, constant-factor overhead and memory layout may determine which GTP variant wins, so the O(L^2 log^2 L) advantage may not always show in wall-clock time.
  • If the approach generalizes beyond LCAO-basis Hamiltonians, it could extend the same efficiency gain to other quantum-operator learning tasks; the paper only targets the LCAO basis.

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

2 major / 5 minor

Summary. The paper introduces Hot-Ham, an E(3)-equivariant message-passing neural network for predicting DFT Hamiltonians. The architecture combines Gaunt tensor products (GTP), which are efficient but generate only symmetric-parity features, with a single final SO(2)-convolution/CGTP layer intended to introduce the antisymmetric-parity tensors needed by the Hamiltonian readout. The authors report O(L^6)-to-O(L^3) or O(L^2 log^2 L) complexity reductions, benchmark on monolayer graphene, monolayer MoS2, and bilayer graphene, and demonstrate transferability to twisted multilayer MoS2, graphene/h-BN heterostructures, and phosphorus allotropes.

Significance. If the reported accuracy and efficiency hold, Hot-Ham is a meaningful advance: it addresses a real computational bottleneck in equivariant Hamiltonian learning, and the combination of GTP with local-coordinate SO(2) convolution is a plausible design. The generalization tests to unseen twists, heterostructures, and allotropes are valuable and go beyond standard benchmarks. The complexity reduction from O(L^6) to O(L^3) or O(L^2 log^2 L) is well motivated and supported by timing experiments. The paper also reports competitive or superior MAEs with fewer parameters than prior models, which, if reproducible, would make Hot-Ham practically useful for large-scale electronic-structure prediction.

major comments (2)
  1. [Equivariant message passing neural network] The central architectural assumption is stated but not verified: 'Since only symmetric tensors can be generated by GTP, we need to introduce antisymmetric tensors at least once through CGTP. Given that including antisymmetric tensors in intermediate layers would result in about double computational costs, we choose to perform CGTP in the last layer through SO(2) convolution.' This single final CGTP/SO(2) layer must generate all antisymmetric tensor components required by Eq. (1) for every parity p3 = (-1)^{l1+l2}. The paper gives no expressivity argument and no ablation. If some antisymmetric irreps are structurally missing from this restricted final bilinear map, the model would systematically mispredict those Hamiltonian blocks, and the reported MAEs would be dataset-specific rather than generally accurate. I recommend adding (i) an ablation with antisymmetric tensors in one or more in
  2. [Benchmark accuracy / Table 1] The 'state-of-the-art accuracy' claim is based on single best MAE values compared with the best previously published values from other models. Table 1 reports no error bars, and the Hot-Ham numbers appear to be from a single training run. Since the margins over DeepH-2 and DeePTB are small (e.g., 0.08 vs 0.12 meV for graphene, 0.12 vs 0.14 meV for MoS2), it is unclear whether these differences are statistically significant. Please report the mean and standard deviation over multiple random seeds, and, where feasible, retrain the baselines on the same train/validation/test splits used for Hot-Ham, or at least state the exact split used. Without this, the comparison is suggestive but not conclusive.
minor comments (5)
  1. [Introduction] Typo: 'graph neutral networks' should be 'graph neural networks'. Also, 'refered' (Equivariant Hamiltonian section) should be 'referred', and 'traning' (Multilayer twisted MoS2 section) should be 'training'.
  2. [Readout] The Hermitian enforcement is written as H̃_ij = 1/2 (H_ij + H_ji^⋇). The symbol '⋇' is nonstandard; presumably this is a conjugate transpose or dagger. Please use standard notation, e.g., H_ji^†.
  3. [Fig. 1c] The complexity labels in the figure (O(L^2), O(L^2 log^2 L)) are somewhat detached from the plotted curves. Please make clear which curve corresponds to which complexity, or move the labels to the caption/legend.
  4. [References] Reference 6 is incomplete: it lists authors and title but no journal/arXiv identifier. Please complete it. Also check for formatting issues like 'V ASPKIT' in the Applications section.
  5. [Data availability] No statement is provided about code or data availability. Given the empirical nature of the paper, a code release or at least a detailed data-availability statement would substantially aid reproducibility; this is not a correctness issue but would strengthen the contribution.

Circularity Check

0 steps flagged

No significant circularity: the accuracy and efficiency claims are empirical ML benchmarks and algebraic complexity results, not self-referential derivations.

full rationale

Hot-Ham's derivation chain is an empirical ML construction, not a closed-form derivation from its own outputs. The Hamiltonian readout (Eq. 1) is a fixed Wigner-Eckart decomposition with Clebsch-Gordan coefficients; the network learns the irreducible tensor coefficients x_{ij,l3,m3,p3} by minimizing a loss against DFT Hamiltonians on held-out structures. The accuracy claims (0.08/0.12/0.15 meV on monolayer graphene, monolayer MoS2, and bilayer graphene) are evaluated on external public datasets and on out-of-distribution twisted/heterostructure/allotrope examples, so no test quantity is used to set a training parameter. The complexity reduction from O(L^6) to O(L^3) or O(L^2 log^2 L) follows from the algebraic properties of SO(2) convolution and Gaunt tensor products and is not a circular restatement. The choice to perform the only antisymmetry-generating CGTP in the final layer is an architectural assumption; it is not claimed to be forced by a theorem, and it is validated empirically rather than by definition. The only self-citation (DeepH-2, ref. 13) is used as a comparison baseline and architectural precedent, not as load-bearing justification for Hot-Ham's accuracy or efficiency. Consequently there is no circular step.

Axiom & Free-Parameter Ledger

2 free parameters · 5 axioms · 0 invented entities

The model depends on standard group-theoretic decomposition of Hamiltonians, the equivariance properties of GTP, and the quality of DFT reference data. No new physical entities are introduced. The neural network is a high-dimensional fit, which is acknowledged by the method itself.

free parameters (2)
  • Trainable MPNN weights (linear layers, MLPs, gates, normalization parameters) = Parameter counts per model: 0.9M (graphene), 1.9M (MoS2)
    Optimized with AdamW to minimize the L1+L2 Hamiltonian loss on training structures.
  • Hyperparameters (number of layers, channels, cutoff radius, Chebyshev/RBF basis, learning rate, epochs) = Not stated in main text; deferred to supplementary
    Chosen by hand and validation; they affect accuracy and the complexity claims.
axioms (5)
  • standard math LCAO Hamiltonian blocks can be expanded via Clebsch-Gordan coefficients as in Eq. (1) (Wigner-Eckart theorem).
    Used to construct the readout; standard angular momentum coupling.
  • domain assumption Atomic orbital parity is p=(-1)^l, so Hamiltonian block parity is p3=(-1)^{l1+l2}.
    This fixes which tensor parities the network must produce; physical convention for LCAO basis.
  • domain assumption The Gaunt tensor product is E(3)-equivariant and faster than CGTP, but only generates tensors with parity (-1)^l; it lacks antisymmetric parity.
    Borrowed from refs 22, 23 and 35; central to the architecture's efficiency and to the need for a final SO(2) layer.
  • ad hoc to paper GTP(sphere-grid) FFT and GTP(2D-FB) implementations are numerically accurate for the products used.
    The paper relies on these implementations for accuracy and timing; no numerical error analysis is given.
  • domain assumption DFT Hamiltonians computed with OpenMX/PBE and the specified PAO basis are the ground-truth targets.
    All MAEs are measured against this reference; if the reference is inadequate, conclusions transfer to it.

pith-pipeline@v1.4.0-alltime-deepseek-medium · 14067 in / 13711 out tokens · 129891 ms · 2026-08-05T05:48:29.413158+00:00 · methodology

0 comments
Cite this review

Pith. "Pith review of Hot-Ham: an accurate and efficient E(3)-equivariant machine-learning electronic structures calculation framework." pith.science (2026). https://pith.science/paper/474KO3WF

@misc{pith2026250904875,
  author       = {Pith},
  title        = {Pith review of: Hot-Ham: an accurate and efficient E(3)-equivariant machine-learning electronic structures calculation framework},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/474KO3WF}},
  note         = {Machine review of arXiv:2509.04875}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

The combinations of machine learning with ab initio methods have attracted much attention for their potential to resolve the accuracy-efficiency dilemma and facilitate calculations for large-scale systems. Recently, equivariant message passing neural networks (MPNNs) that explicitly incorporate symmetry constraints have demonstrated promise for interatomic potential and density functional theory (DFT) Hamiltonian predictions. However, the high-order tensors used to represent node and edge information are coupled through the Clebsch-Gordan tensor product (CGTP), leading to steep increases in computational complexity and seriously hindering the performance of equivariant MPNNs. Here, we develop High-order Tensor machine-learning Hamiltonian (Hot-Ham), an E(3) equivariant MPNN framework that combines two advanced technologies local coordinate transformation and Gaunt tensor product (GTP) to efficiently model DFT Hamiltonians. These two innovations significantly reduce the complexity of tensor products from O(L^6) to O(L^3) or O(L^2 log^2 L) for the max tensor order L, and enhance the performance of MPNNs. Benchmarks on several public datasets demonstrate its state-of-the-art accuracy with relatively few parameters, and the applications to multilayer twisted moir\'e systems, heterostructures and allotropes showcase its generalization ability and high efficiency. Our Hot-Ham method provides a new perspective for developing efficient equivariant neural networks and would be a promising approach for investigating the electronic properties of large-scale materials systems.

Figures

Figures reproduced from arXiv: 2509.04875 by Chi Ding, Dingyu Xing, Hui-Tian Wang, Jian Sun, Junjie Wang, Yunlong Wang, Zhixin Liang.

Figure 1
Figure 1. Figure 1: Illustration of Gaunt tensor product (GTP) method and comparison of [PITH_FULL_IMAGE:figures/full_fig_p022_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The architecture of Hot-Ham. a The overall architecture of Hot-Ham. Atom species and relative positions are embedded into initial node and edge features. After several iterative updates in convolution layers, the final node and edge features are used to construct Hamiltonians. b the embedding layer. The scalar information including atom species 𝑍𝑖 , 𝑍𝑗 and interatomic distance |r𝑖𝑗| is embedded through the… view at source ↗
Figure 3
Figure 3. Figure 3: Application of Hot-Ham to multilayer MoS2. a Hamiltonian MAE for bilayer, trilayer, and double bilayer under different twist angles. b The wall time for DFT calculations and our model’s inference at CPU and GPU. c structure of double bilayer MoS2 containing 1524 atoms with twist angle 𝜃 = 5.09° . d band structure for the structure in c predicted by DFT (OpenMX) and Hot-Ham. (a) (b) 20 15 10 5 1.4 1.0 0.6 0… view at source ↗
Figure 4
Figure 4. Figure 4: Performance of Hot-Ham on graphene/h-BN heterostructures. a 16 × 16 supercell cell of a Bernal arrangement graphene/h-BN heterostructure. b the band structure of a. c comparison of band eigenvalues within -1.5~1.5 eV for the 100 heterostructures with various mismatch and twisted angles. The MAE of their Hamiltonians is also displayed. d structure of graphene/h-BN containing 1022 atoms (518 C atoms, 252 B/N… view at source ↗
Figure 5
Figure 5. Figure 5: Predicted electronic structure on phosphorus allotropes. [PITH_FULL_IMAGE:figures/full_fig_p027_5.png] 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

49 extracted references · 33 canonical work pages · 1 internal anchor

  1. [1]

    Slater, J. C. & Koster, G. F. Simplified LCAO Method for the Periodic Potential Problem. Phys. Rev. 94, 1498–1524 (1954)

    work page 1954

  2. [2]

    Wang, Z. et al. Machine learning method for tight -binding Hamiltonian parameterization from ab-initio band structure. npj Comput Mater 7, 11 (2021)

    work page 2021

  3. [3]

    & Feng, J

    Gu, Q., Zhang, L. & Feng, J. Neural network representation of electronic structure from ab initio molecular dynamics. Science Bulletin 67, 29–37 (2022)

    work page 2022

  4. [4]

    Gu, Q. et al. Deep learning tight -binding approach for large -scale electronic simulations at finite temperatures with ab initio accuracy. Nat Commun 15, 6772 (2024)

    work page 2024

  5. [5]

    T., Gastegger, M., Tkatchenko, A., Müller, K.-R

    Schütt, K. T., Gastegger, M., Tkatchenko, A., Müller, K.-R. & Maurer, R. J. Unifying machine learning and quantum chemistry with a deep neural network for molecular wavefunctions. Nat Commun 10, 5024 (2019)

    work page 2019

  6. [6]

    Unke, O. T. & Bogojeski, M. SE(3) -equivariant prediction of molecular wavefunctions and electronic densities. 17 / 27

    work page

  7. [7]

    Yu, H., Xu, Z., Qian, X., Qian, X. & Ji, S. Efficient and Equivariant Graph Networks for Predicting Quantum Hamiltonian. in Proceedings of the 40th International Conference on Machine Learning 40412–40424 (PMLR, 2023)

    work page 2023

  8. [8]

    Zhang, L. et al. Equivariant analytical mapping of first principles Hamiltonians to accurate and transferable materials models. npj Comput Mater 8, 158 (2022)

    work page 2022

  9. [9]

    Nigam, J., Willatt, M. J. & Ceriotti, M. Equivariant representations for molecular Hamiltonians and N -center atomic -scale properties. The Journal of Chemical Physics 156, 014115 (2022)

    work page 2022

  10. [10]

    Li, H. et al. Deep-Learning Density Functional Theory Hamiltonian for Efficient ab initio Electronic-Structure Calculation. Nat Comput Sci 2, 367–377 (2022)

    work page 2022

  11. [11]

    Gong, X. et al. General framework for E(3) -equivariant neural network representation of density functional theory Hamiltonian. Nat Commun 14, 2848 (2023)

    work page 2023

  12. [12]

    & Xiang, H

    Zhong, Y ., Yu, H., Su, M., Gong, X. & Xiang, H. Transferable equivariant graph neural networks for the Hamiltonians of molecules and solids. npj Comput Mater 9, 182 (2023)

    work page 2023

  13. [13]

    Wang, Y . et al. DeepH-2: Enhancing deep -learning electronic structure via an equivariant local -coordinate transformer. Preprint at http://arxiv.org/abs/2401.17015 (2024)

    Pith/arXiv arXiv 2024

  14. [14]

    K., Zhang, L

    Zhouyin, Z., Gan, Z., Pandey, S. K., Zhang, L. & Gu, Q. Learning local equivariant representations for quantum operators. Preprint at http://arxiv.org/abs/2407.06053 (2024)

    Pith/arXiv arXiv 2024

  15. [15]

    C., Hagenbuchner, M

    Scarselli, F., Gori, M., Tsoi, A. C., Hagenbuchner, M. & Monfardini, G. The Graph Neural Network Model. IEEE Transactions on Neural Networks 20, 61–80 (2009). 16.Battaglia, P. W. et al. Relational inductive biases, deep learning, and graph networks. Preprint at https://doi.org/10.48550/arXiv.1806.01261 (2018)

    Pith/arXiv arXiv doi:10.48550/arxiv.1806.01261 2009

  16. [17]

    Kipf, T. N. & Welling, M. Semi -Supervised Classification with Graph Convolutional Networks. Preprint at https://doi.org/10.48550/arXiv.1609.02907 (2017). 18 / 27

    Pith/arXiv arXiv doi:10.48550/arxiv.1609.02907 2017

  17. [18]

    Zhou, J. et al. Graph Neural Networks: A Review of Methods and Applications. Preprint at https://doi.org/10.48550/arXiv.1812.08434 (2021)

    Pith/arXiv arXiv doi:10.48550/arxiv.1812.08434 2021

  18. [19]

    F., Vinyals, O

    Gilmer, J., Schoenholz, S., Riley, P. F., Vinyals, O. & Dahl, G. E. Neural Message Passing for Quantum Chemistry. in (2017)

    work page 2017

  19. [20]

    & Zitnick, C

    Passaro, S. & Zitnick, C. L. Reducing SO(3) Convolutions to SO(2) for Efficient Equivariant GNNs. Preprint at http://arxiv.org/abs/2302.03655 (2023)

    Pith/arXiv arXiv 2023

  20. [21]

    P., Simm, G

    Batatia, I., Kovács, D. P., Simm, G. N. C., Ortner, C. & Csányi, G. MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fields. Preprint at http://arxiv.org/abs/2206.07697 (2023)

    Pith/arXiv arXiv 2023

  21. [22]

    Xin, H. et al. Fast and accurate spherical harmonics products. ACM Trans. Graph. 40, 1–14 (2021)

    work page 2021

  22. [23]

    & Krishnapriyan, A

    Luo, S., Chen, T. & Krishnapriyan, A. S. Enabling Efficient Equivariant Operations in the Fourier Basis via Gaunt Tensor Products. Preprint at http://arxiv.org/abs/2401.10216 (2024)

    Pith/arXiv arXiv 2024

  23. [24]

    Projector augmented -wave method | Phys. Rev. B. https://journals.aps.org/prb/abstract/10.1103/PhysRevB.50.17953

    doi:10.1103/physrevb.50.17953

  24. [25]

    & Furthmüller, J

    Kresse, G. & Furthmüller, J. Efficient iterative schemes for ab initio total -energy calculations using a plane-wave basis set. Phys. Rev. B 54, 11169–11186 (1996)

    1996

  25. [26]

    & Furthmüller, J

    Kresse, G. & Furthmüller, J. Efficiency of ab -initio total energy calculations for metals and semiconductors using a plane-wave basis set. Computational Materials Science 6, 15–50 (1996)

    work page 1996

  26. [27]

    & Kino, H

    Ozaki, T. & Kino, H. Numerical atomic basis orbitals from H to Kr. Phys. Rev. B 69, 195113 (2004)

    work page 2004

  27. [28]

    Variationally optimized atomic orbitals for large -scale electronic structures

    Ozaki, T. Variationally optimized atomic orbitals for large -scale electronic structures. Phys. Rev. B 67, 155108 (2003)

    work page 2003

  28. [29]

    Li, P. et al. Large-scale ab initio simulations based on systematically improvable atomic basis. Computational Materials Science 112, 503–517 (2016)

    work page 2016

  29. [30]

    Lin, P., Ren, X., Liu, X. & He, L. Ab initio electronic structure calculations based on numerical atomic orbitals: Basic fomalisms and recent progresses. WIREs 19 / 27 Computational Molecular Science 14, e1687 (2024)

    work page 2024

  30. [31]

    & Smidt, T

    Liao, Y .-L. & Smidt, T. Equiformer: Equivariant Graph Attention Transformer for 3D Atomistic Graphs. Preprint at http://arxiv.org/abs/2206.11990 (2023)

    Pith/arXiv arXiv 2023

  31. [32]

    Batzner, S. et al. E(3)-equivariant graph neural networks for data -efficient and accurate interatomic potentials. Nat Commun 13, 2453 (2022)

    work page 2022

  32. [33]

    Musaelian, A. et al. Learning Local Equivariant Representations for Large -Scale Atomistic Dynamics. Preprint at http://arxiv.org/abs/2204.05249 (2022)

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  33. [34]

    https://pubs.aip.org/aapt/ajp/article- abstract/28/4/408/1036645/Group-Theory-and-Its-Application-to-the- Quantum?redirectedFrom=fulltext

    Group Theory and Its Application to the Quantum Mechanics of Atomic Spectra | American Journal of Physics | AIP Publishing. https://pubs.aip.org/aapt/ajp/article- abstract/28/4/408/1036645/Group-Theory-and-Its-Application-to-the- Quantum?redirectedFrom=fulltext

    work page

  34. [35]

    & Smidt, T

    Xie, Y ., Daigavane, A., Kotak, M. & Smidt, T. The Price of Freedom: Exploring Tradeoffs between Expressivity and Computational Efficiency in Equivariant Tensor Products. 36.Weiler, M., Geiger, M., Welling, M., Boomsma, W. & Cohen, T. 3D Steerable CNNs: Learning Rotationally Equivariant Features in V olumetric Data. Preprint at http://arxiv.org/abs/1807.0...

    Pith/arXiv arXiv 2018

  35. [37]

    Precise control of the interlayer twist angle in large scale MoS2 homostructures

    Liao, M. Precise control of the interlayer twist angle in large scale MoS2 homostructures

    work page

  36. [38]

    Zhang, S. et al. Domino-like stacking order switching in twisted monolayer – multilayer graphene. Nat. Mater. 21, 621–626 (2022)

    work page 2022

  37. [39]

    & Koshino, M

    Moon, P. & Koshino, M. Electronic properties of graphene/hexagonal-boron-nitride moire´ superlattice. PHYSICAL REVIEW B (2014)

    work page 2014

  38. [40]

    Wang, E. et al. Gaps induced by inversion symmetry breaking and second - generation Dirac cones in graphene/hexagonal boron nitride. Nature Phys 12, 1111– 1115 (2016)

    work page 2016

  39. [41]

    Electrical properties and applications of graphene, hexagonal boron nitride (h-BN), and graphene/h-BN heterostructures

    Wang, J. Electrical properties and applications of graphene, hexagonal boron nitride (h-BN), and graphene/h-BN heterostructures. Materials Today Physics (2017)

    work page 2017

  40. [42]

    & Geng, W

    Wang, V ., Xu, N., Liu, J.-C., Tang, G. & Geng, W. -T. V ASPKIT: A user-friendly 20 / 27 interface facilitating high -throughput computing and analysis using V ASP code. Computer Physics Communications 267, 108033 (2021)

    work page 2021

  41. [43]

    Jain, A. et al. Commentary: The Materials Project: A materials genome approach to accelerating materials innovation. APL Materials 1, 011002 (2013)

    2013

  42. [44]

    Zhang, L. et al. Structure and Properties of Violet Phosphorus and Its Phosphorene Exfoliation. Angew Chem Int Ed 59, 1074–1080 (2020)

    work page 2020

  43. [45]

    Wang, J. et al. MAGUS: machine learning and graph theory assisted universal structure searcher. National Science Review 10, nwad128 (2023)

    work page 2023

  44. [46]

    Fan, Z. et al. Linear scaling quantum transport methodologies. Physics Reports 903, 1–69 (2021)

    work page 2021

  45. [47]

    & Yuan, S

    Li, Y ., Zhan, Z., Kuang, X., Li, Y . & Yuan, S. TBPLaS: A tight-binding package for large-scale simulation. Computer Physics Communications 285, 108632 (2023)

    work page 2023

  46. [48]

    P., Burke, K

    Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 77, 3865–3868 (1996)

    1996

  47. [49]

    Larsen, A. H. et al. The atomic simulation environment —a Python library for working with atoms. J. Phys.: Condens. Matter 29, 273002 (2017)

    work page 2017

  48. [50]

    & Hutter, F

    Loshchilov, I. & Hutter, F. Decoupled Weight Decay Regularization. Preprint at https://doi.org/10.48550/arXiv.1711.05101 (2019)

    Pith/arXiv arXiv doi:10.48550/arxiv.1711.05101 2019

  49. [51]

    J., Kale, S

    Reddi, S. J., Kale, S. & Kumar, S. On the Convergence of Adam and Beyond. Preprint at https://doi.org/10.48550/arXiv.1904.09237 (2019). Acknowledgments We gratefully acknowledge the financial support from the National Natural Science Foundation of China (grant number. 12125404, T2495231, 123B2049), the Basic Research Program of Jiangsu (Grant BK20233001, ...

    Pith/arXiv arXiv doi:10.48550/arxiv.1904.09237 1904