Pith. sign in

REVIEW 4 major objections 6 minor 57 references

Transformer-Enhanced Variational Autoencoder for Crystal Structure Prediction

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read TransVAE-CSP claims that an equivariant dot-product attention encoder plus per-dataset radial-basis-function choice improves crystal reconstruction and generation over previous generative models on carbon_24, perov_5, and mp_20.

desk verdict A plausible CDVAE encoder upgrade with real reconstruction wins on two datasets, but the abstract overclaims and the single-seed, early-loss RBF selection makes the central comparison untrustworthy. read the letter →

arxiv 2502.09423 v1 pith:7WX752K2 submitted 2025-02-13 cond-mat.mtrl-sci cs.AI

Ziyi Chen , Yang Yuan , Siming Zheng , Jialong Guo , Sihan Liang , Yangang Wang , Zongguo Wang This is my paper · ORCID
classification cond-mat.mtrl-scics.AI
keywords crystalstructurepredictionvariationalautoencoderequivariantattentionradialbasisfunctionsgenerativematerialsdesigndiffusiondecoderperiodicE(3)symmetry
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

TransVAE-CSP is a generative model that aims to learn the distribution of stable crystal structures so that a variational autoencoder can both reconstruct known crystals and generate new ones from scratch. The paper argues that two changes to the CDVAE (Crystal Diffusion Variational Autoencoder) recipe carry the improvement: replacing the graph-convolutional encoder with a Transformer based on equivariant dot-product attention, and choosing the radial basis function used for distance expansion separately for each dataset (Gaussian, Bessel, or a hybrid). The reported results show higher reconstruction match rates than CDVAE on carbon_24 and mp_20, competitive generation coverage and validity on all three benchmarks, and the best density and energy distribution distances on mp_20. This matters because crystal structure prediction is a bottleneck in materials discovery, and a model that reconstructs and generates structures more faithfully can narrow candidate lists before expensive DFT screening.

What carries the argument

The load-bearing mechanism is an equivariant dot-product attention layer in the encoder: each atom carries features organized in irreducible representations (rotation-covariant geometric channels), pairwise messages are built by depth-wise tensor products of spherical harmonics with radial-basis-expanded distances, and attention weights come from scaled dot products between query and key features. This replaces the graph convolution used by CDVAE and is what preserves E(3) equivariance while letting atoms attend over local environments. The second mechanism is adaptive distance expansion: interatomic distances are expanded by a Gaussian, Bessel, or weighted-concatenation hybrid radial basis function, and the variant is selected per dataset by comparing training loss after 200 iterations. The diffusion-based decoder and predictor are inherited from CDVAE.

What would settle it

Train all three radial-basis variants to full convergence on each of the three datasets and compare final reconstruction match rates and generation metrics; if the variant selected by the 200-iteration loss curves is not the best at convergence, the adaptive-selection claim is unsupported.

Watch

Extended reading notes

Core claim

The central claim is that a VAE whose encoder is an E(3)-equivariant dot-product attention Transformer, paired with a diffusion-style decoder, learns a latent distribution over crystal structures that is more faithful than the graph-convolutional encoder used by CDVAE. On the paper's own numbers, the model reconstructs 80.75% of carbon_24 test structures and 71.14% of mp_20 test structures (versus 55.22% and 45.43% for CDVAE), and reaches 98.19% on perov_5 with the lowest RMSE among listed models. In ab initio generation it reports 100% structural validity on perov_5 and mp_20, 100% coverage recall on carbon_24, and the best density Earth mover's distance on perov_5 among all baselines. The paper attributes these gains to better symmetry and periodicity capture through the equivariant encoder and to a per-dataset choice of distance expansion.

Load-bearing premise

The per-dataset choice of distance-expansion function is based on which variant has the lowest loss after only 200 training iterations (Section 4.3 and Appendix B.4), and the paper assumes this early ranking predicts final reconstruction and generation quality, with no repeated-seed statistics reported for the chosen variant.

Editorial extensions

If this is right

  • If the reported numbers hold, equivariant attention encoders are a direct upgrade over graph convolutional encoders in VAE-based crystal generation, not just in property prediction.
  • Dataset-specific distance expansion becomes a cheap and effective tuning lever, since the best RBF differs by dataset and no single representation dominates.
  • A VAE with a diffusion decoder can match diffusion-only generators on validity and coverage while retaining the reconstruction and latent-space consistency checks a VAE provides.
  • On mp_20, better composition validity and density and energy distribution distances mean the generated distribution is closer to stable inorganic materials, which is the practical target for materials screening.
  • The model's latent space is a standard normal, so it can be sampled directly for ab initio generation without requiring composition or atom-count inputs.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A testable extension is to replace the 200-iteration RBF selection with full-convergence or early-stopping validation; if the cheap selection rule holds across datasets, it becomes a practical hyperparameter curriculum for representation choice.
  • The carbon_24 coverage gap (100% recall versus 78.62% precision) suggests the model generates structures that are diverse but not all near the test distribution, and checking the generated carbon allotropes against known metastable motifs would show whether this is a sampling-precision issue rather than a validity issue.
  • Because the encoder is equivariant, the same attention block could transfer to conditional generation tasks such as generating structures with specified composition or lattice parameters, which the paper names as future work.
  • The hybrid RBF's weight-scaling scheme could be studied as a continuous interpolation between Bessel and Gaussian behavior, potentially making the per-dataset choice differentiable and trainable rather than a discrete selection.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper proposes TransVAE-CSP, a variational autoencoder for crystal structure prediction that builds on CDVAE by replacing the encoder with a transformer based on equivariant dot-product attention and by introducing an 'adaptive distance expansion' that selects among Gaussian, Bessel, and hybrid radial basis functions for each dataset. The authors evaluate the model on the carbon_24, perov_5, and mp_20 datasets, reporting reconstruction match rates and RMSE in Table 1 and generation validity, coverage, and property statistics in Table 3. The abstract claims the model outperforms existing methods, but a close reading of the tables shows that the claimed superiority is not uniform across metrics and datasets, and the experimental protocol lacks repeated seeds and error bars.

Significance. If the reported improvements were robust, the equivariant dot-product attention encoder would be a useful contribution to crystal generative modeling, and the per-dataset choice of RBF could serve as a practical guideline. The paper provides a detailed architecture description and follows established evaluation benchmarks, which are valuable. However, the central claim of overall superiority is currently unsupported by the paper's own numbers, and the evaluation protocol does not separate genuine gains from selection or seed noise. The potential significance is therefore real but contingent on a more rigorous and more carefully qualified experimental demonstration.

major comments (4)
  1. [Abstract and §4.2.3, Table 3] The abstract's claim that TransVAE-CSP 'outperforms existing methods' is contradicted by the paper's own results: in Table 3, on carbon_24, TransVAE-CSP achieves COV-P 78.62 versus DiffCSP's 97.27 and dE 1.744 versus 0.082, and on MP_20 its delem of 0.7348 is worse than DiffCSP's 0.3398. The text in §4.2.3 saying that 'other indicators also achieve excellent performance' does not reconcile these deficits with the blanket claim. The claim must be reformulated as per-metric, per-dataset comparisons, or supporting statistical evidence must be provided.
  2. [§4.3 and Appendix B.4] The RBF variant for each dataset is selected by inspecting loss curves over only 200 training iterations, and the single chosen variant is then used to produce the final benchmark numbers in Tables 1 and 3, with no repeated seeds or error bars. The paper provides no evidence that the early-convergence ranking at 200 iterations predicts final reconstruction or generation quality; with a single run, the reported improvements could be selection or seed noise. The authors should report results over multiple seeds with mean and variance, and either justify the 200-iteration selection criterion or treat the RBF choice as a hyperparameter selected on the validation set rather than the test set.
  3. [§3.1 and Eq. (H(x))] The 'adaptive distance expansion' is not adaptive in the sense the name suggests: it is a per-dataset choice among three fixed RBF functions, as stated in §4.3. Moreover, the hybrid RBF formula H(x) = Φ(x) L k · Ψ(x) is ambiguous because the concatenation operation 'L' and the role of the scalar weight k are not formally defined; if the weight is intended to be a learned parameter, the optimization of k is never described. This ambiguity undermines the reproducibility of the proposed representation module.
  4. [§4.1.3 and Table 1] The reconstruction RMSE of TransVAE-CSP on MP_20 (0.0377) is larger than CDVAE's (0.0356), and the match rate on perov_5 is lower than FTCP (98.19 versus 99.34). These differences are small and could be within run-to-run noise, but without error bars or significance tests the statement that 'other indicators surpass the baseline models to varying extents' is not verifiable and should be supported by repeated runs or omitted.
minor comments (6)
  1. [Abstract and Introduction] The abstract and introduction contain grammatical errors ('who learns', 'pervious work'); these should be corrected.
  2. [§3.3] The notation A = (a0, a1, ..., a_{n-1}) ∈ E^N is confusing because the set E is not defined as the set of element types; furthermore, L = (a,b,c,d,e) ∈ R^6 does not match the usual representation of a lattice as a 3×3 matrix, so the meaning of these five parameters should be spelled out.
  3. [Appendix A.2, Eq. (5)] The Bessel RBF is defined with J_v and a cosine factor, but the parameters v and ω are not specified; please state their values or explain how they are set.
  4. [§3.2, Eq. (1)] Equation (1) uses the notation ⊗_DTP and refers to a 'DTP network' without a precise definition; please provide a definition or a clear reference to the original formulation.
  5. [References] The reference to Oubari et al. is dated 2012, but the cited workshop paper appears to be from 2021; please verify and correct the year.
  6. [Appendix B.4] The text refers to Fig. 5 for the RBF comparison curves, but the figure is not included in the presented version; the figure should be provided so that the selection rationale can be checked.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: TransVAE-CSP is an empirical benchmark study whose reconstruction and generation claims are evaluated on held-out test sets, with no derivation step reducing to its own inputs.

full rationale

This is an empirical machine-learning paper rather than a derivation chain. The central claims are reconstructed and generated crystal structures evaluated on held-out test splits against external baselines and metrics. The paper explicitly states that it builds on CDVAE ('Building on the work of CDVAE... our research focuses on optimizing crystal structure representation and the encoder network'), which is a normal and transparent incremental-development statement rather than a circular reduction. The only mild selection concern is that the radial basis function variant is chosen per dataset after comparing loss curves over 200 training iterations (Section 4.3 and Appendix B.4), and the final benchmark tables then report the chosen variant; however, this is standard hyperparameter/model selection on training behavior, not a fitted parameter being relabeled as a prediction, and the reported test metrics still require the model to actually reconstruct and generate structures. Moreover, the tables include multiple cases where TransVAE-CSP does not beat a baseline (e.g., carbon_24 COV-P 78.62 vs. DiffCSP 97.27 and dE 1.744 vs. 0.082; MP_20 delem 0.7348 vs. 0.3398), so the headline claim is not forced by construction. The only self-citation to the authors' prior work (Yuan et al. 2024, cited for three-body interaction representations) appears in related work and is not load-bearing for the model's design or conclusions. No quoted equation or argument reduces by the paper's own definitions to its inputs; the absence of repeated seeds is a reproducibility concern, not circularity.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central claims rest on the CDVAE pipeline, the equivariant attention design, the per-dataset RBF selection, and the benchmark data. No new physical entities are introduced. The main free parameters are the per-dataset RBF choice and the unspecified hybrid weighting, both tuned empirically and central to the claimed improvements. The most fragile assumption is that early training-loss comparisons justify the final RBF selection.

free parameters (5)
  • Per-dataset RBF basis function = Perov5: Bessel; Carbon24: Hybrid; MP20: Gaussian
    The authors compare training loss curves for Gaussian, Bessel, and hybrid RBFs and select the one with the best early convergence per dataset (Section 4.1.3, Section 4.3). This choice affects the input features and the final reported metrics.
  • Hybrid RBF weight k = Not reported
    The hybrid formula H(x) = Phi(x) L k * Psi(x) includes a weight k and a concatenation operation, but the value of k and how it is 'adaptively adjusted' are not specified. The weighting changes the feature scaling and thus the model behavior.
  • Loss weights and KL beta = lambda_Ac=1, lambda_L=10, lambda_N=1, lambda_X=10, lambda_A=1; beta=0.01 (Perov5, MP20), beta=0.03 (Carbon24)
    These hyperparameters are set by hand following CDVAE and control the balance between reconstruction, prediction, and latent regularization in the total loss (Appendix B.3).
  • Cutoff radius and max neighbors = Cutoff 6/6/10 Angstrom; max neighbors 20/20/50 for Perov5, Carbon24, MP20
    These graph construction parameters are chosen per dataset and determine which atomic pairs are included in the distance expansion and attention (Appendix B.2).
  • Training epochs and learning rate = Epochs 3500/4000/1500; lr 1e-4
    Per-dataset training length and learning rate are chosen empirically; these affect convergence and final performance (Appendix B.2).
assumptions (5)
  • standard math RBF and spherical harmonic expansions provide adequate geometric features for crystal representation
    The method relies on the established approximation properties of Gaussian and Bessel RBFs and spherical harmonics to encode interatomic distances and angles (Section 3.1, Section 3.2).
  • domain assumption The CDVAE decoder, noise schedule, and loss decomposition form a valid generative foundation
    The paper builds directly on CDVAE (Xie et al., 2021) and adopts its predictor, denoising decoder, and loss terms without re-deriving them (Section 3.3, Appendix B.3). If CDVAE's generative assumptions are flawed, the reported improvements inherit those flaws.
  • domain assumption Equivariant dot-product attention from Equiformer and SE(3)-Transformer preserves the required E(3) symmetry in the latent space
    The encoder replaces CDVAE's GCN with an equivariant transformer; the paper assumes this architecture maintains translation, rotation, and inversion equivariance as claimed by the cited works (Section 3.2).
  • ad hoc to paper Early (200-iteration) training loss ranking predicts final generation and reconstruction performance
    The RBF selection is based on loss convergence over only 200 iterations (Section 4.3, Appendix B.4). The paper does not show that this early ranking correlates with final test metrics, yet it reports the selected variant's test results as the model's performance.
  • domain assumption Dataset labels and coordination numbers computed by CrystalNN are reliable
    The benchmarks (Perov5, Carbon24, MP20) are taken from prior work with DFT-derived stability labels, and coordination numbers are computed with Pymatgen's CrystalNN (Appendix B.1). Errors in these inputs would propagate to the learned representation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Transformer-Enhanced Variational Autoencoder for Crystal Structure Prediction." pith.science (2026). https://pith.science/paper/7WX752K2

@misc{pith2026250209423,
  author       = {Pith},
  title        = {Pith review of: Transformer-Enhanced Variational Autoencoder for Crystal Structure Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7WX752K2}},
  note         = {Machine review of arXiv:2502.09423}
}
read the original abstract

Crystal structure forms the foundation for understanding the physical and chemical properties of materials. Generative models have emerged as a new paradigm in crystal structure prediction(CSP), however, accurately capturing key characteristics of crystal structures, such as periodicity and symmetry, remains a significant challenge. In this paper, we propose a Transformer-Enhanced Variational Autoencoder for Crystal Structure Prediction (TransVAE-CSP), who learns the characteristic distribution space of stable materials, enabling both the reconstruction and generation of crystal structures. TransVAE-CSP integrates adaptive distance expansion with irreducible representation to effectively capture the periodicity and symmetry of crystal structures, and the encoder is a transformer network based on an equivariant dot product attention mechanism. Experimental results on the carbon_24, perov_5, and mp_20 datasets demonstrate that TransVAE-CSP outperforms existing methods in structure reconstruction and generation tasks under various modeling metrics, offering a powerful tool for crystal structure design and optimization.

Figures

Figures reproduced from arXiv: 2502.09423 by the authors.

Figure 1
Figure 1. TransVAE-CSP, based on the Variational Autoencoder (VAE) paradigm, excels in both crystal structure reconstruction and ab initio generation tasks. Experimental results show exceptional performance across evaluation metrics. Compared to previous work(Xie et al., 2021), we have introduced innovations and conducted verifications in crystal structure representation and the encoder network. MatterGen(Zeni et al., 2025), … view at source ↗
Figure 2
Figure 2. Overview of TransVAE-CSP. Training VAE model: Given M = (A, N, X, L), it captures feature information through the Embedding layer and the equivariant attention encoder to obtain the latent space variable z, which is utilized as a condition to guide the output of both the Predictor and the denoising Decoder. The loss function comprises three parts:LP red, LDec, LKL. Generation: The variable z is sampled from a multi-… view at source ↗
Figure 3
Figure 3. The technique of equivariant dot product attention net￾work. This is the core network of Transformer. “DTP” stands for depth-wise tensor product, L denotes addition, N denotes multiplication.P denotes scatter operation. 3.2. Equivariant Dot Product Attention Network The encoder utilizes an adapted SE(3)-Transformer net￾work (Fuchs et al., 2020), with the core structure being the equivariant dot product attention net… view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Embedding block. Compared to pervious work(Liao & Smidt, 2023), our adaptation work is highlighted in red, and RBF refers to the feature vector obtained after the radial basis function is expanded based on the distance. Embedding together, the feature embedding of the …
Figure 5
Figure 5. Figure 5: The Comparison of different RBF on Perov 5, Car￾bon 24, MP 20. Figure a) illustrates the curve generated by train￾ing on Perov 5. It is evident that the model utilizing the Bessel RBF exhibits the most effective training results. Similarly, b) Car￾bon 24 - Hybrid RBF, …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

57 extracted references · 35 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    J., and Welling, M

    Brandstetter, J., Hesselink, R., van der Pol, E., Bekkers, E. J., and Welling, M. Geometric and physical quantities improve e(3) equivariant message passing. In International Conference on Learning Representations, 2022

    work page 2022

  3. [3]

    Buhmann, M. D. Radial basis functions. Acta numerica, 9: 0 1--38, 2000

    work page 2000

  4. [4]

    E., Landis, D

    Castelli, I. E., Landis, D. D., Thygesen, K. S., Dahl, S., Chorkendorff, I., Jaramillo, T. F., and Jacobsen, K. W. New cubic perovskites for one-and two-photon water splitting using the computational materials repository. Energy & Environmental Science, 5 0 (10): 0 9034--9043, 2012 a

    work page 2012

  5. [5]

    E., Olsen, T., Datta, S., Landis, D

    Castelli, I. E., Olsen, T., Datta, S., Landis, D. D., Dahl, S., Thygesen, K. S., and Jacobsen, K. W. Computational screening of perovskite metal oxides for optimal solar light capture. Energy & Environmental Science, 5 0 (2): 0 5814--5819, 2012 b

    work page 2012

  6. [6]

    and Ong, S

    Chen, C. and Ong, S. P. A universal graph deep learning interatomic potential for the periodic table. Nature Computational Science, 2 0 (11): 0 718--728, 2022

    work page 2022

  7. [7]

    Chen, C., Ye, W., Zuo, Y., Zheng, C., and Ong, S. P. Graph networks as a universal machine learning framework for molecules and crystals. Chemistry of Materials, 31 0 (9): 0 3564--3572, 2019

    work page 2019

  8. [8]

    and DeCost, B

    Choudhary, K. and DeCost, B. Atomistic line graph neural network for improved materials property predictions. npj Computational Materials, 7 0 (1): 0 185, 2021

    work page 2021

Show all 57 references
  1. [9]

    J., Yildirim, B., Jain, A., and Cole, J

    Court, C. J., Yildirim, B., Jain, A., and Cole, J. M. 3-d inorganic crystal structure generation and property prediction via representation learning. Journal of Chemical Information and Modeling, 60 0 (10): 0 4518--4535, 2020

    2020

  2. [10]

    Generative adversarial networks (gan) based efficient sampling of chemical composition space for inverse design of inorganic materials

    Dan, Y., Zhao, Y., Li, X., Li, S., Hu, M., and Hu, J. Generative adversarial networks (gan) based efficient sampling of chemical composition space for inverse design of inorganic materials. npj Computational Materials, 6 0 (1): 0 84, 2020

    2020

  3. [11]

    W., Butler, K

    Davies, D. W., Butler, K. T., Jackson, A. J., Skelton, J. M., Morita, K., and Walsh, A. Smact: Semiconducting materials by analogy and chemical theory. Journal of Open Source Software, 4 0 (38): 0 1361, 2019

    2019

  4. [12]

    Stable computations with gaussian radial basis functions

    Fornberg, B., Larsson, E., and Flyer, N. Stable computations with gaussian radial basis functions. SIAM Journal on Scientific Computing, 33 0 (2): 0 869--892, 2011

    2011

  5. [13]

    Se (3)-transformers: 3d roto-translation equivariant attention networks

    Fuchs, F., Worrall, D., Fischer, V., and Welling, M. Se (3)-transformers: 3d roto-translation equivariant attention networks. Advances in neural information processing systems, 33: 0 1970--1981, 2020

    1970

  6. [14]

    T., and G \"u nnemann, S

    Gasteiger, J., Giri, S., Margraf, J. T., and G \"u nnemann, S. Fast and uncertainty-aware directional message passing for non-equilibrium molecules. In Machine Learning for Molecules Workshop, NeurIPS, 2020

    2020

  7. [15]

    Gemnet: Universal directional graph neural networks for molecules

    Gasteiger, J., Becker, F., and G \"u nnemann, S. Gemnet: Universal directional graph neural networks for molecules. Advances in Neural Information Processing Systems, 34: 0 6790--6802, 2021

    2021

  8. [16]

    Symmetry-adapted generation of 3d point sets for the targeted discovery of molecules

    Gebauer, N., Gastegger, M., and Sch \"u tt, K. Symmetry-adapted generation of 3d point sets for the targeted discovery of molecules. Advances in neural information processing systems, 32, 2019

    2019

  9. [17]

    Geiger, M., Smidt, T., M., A., Miller, B. K., Boomsma, W., Dice, B., Lapchevskyi, K., Weiler, M., Tyszkiewicz, M., Batzner, S., Madisetti, D., Uhrin, M., Frellsen, J., Jung, N., Sanborn, S., Wen, M., Rackers, J., Rød, M., and Bailey, M. e3nn/e3nn: 2022-04-13, April 2022

    2022

  10. [18]

    Generative adversarial nets

    Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., and Bengio, Y. Generative adversarial nets. Advances in neural information processing systems, 27, 2014

    2014

  11. [19]

    Denoising diffusion probabilistic models

    Ho, J., Jain, A., and Abbeel, P. Denoising diffusion probabilistic models. Advances in neural information processing systems, 33: 0 6840--6851, 2020

    2020

  12. [20]

    M., and Bengio, Y

    Hoffmann, J., Maestrati, L., Sawada, Y., Tang, J., Sellier, J. M., and Bengio, Y. Data-driven approach to encoding and decoding 3-d crystal structures. arXiv preprint arXiv:1909.00949, 2019

    1909 arXiv

  13. [21]

    and Kohn, W

    Hohenberg, P. and Kohn, W. Inhomogeneous electron gas. Physical review, 136 0 (3B): 0 B864, 1964

    1964

  14. [22]

    G., Vignac, C., and Welling, M

    Hoogeboom, E., Satorras, V. G., Vignac, C., and Welling, M. Equivariant diffusion for molecule generation in 3d. In International conference on machine learning, pp.\ 8867--8887. PMLR, 2022

    2022

  15. [23]

    Howard, A. G. Mobilenets: Efficient convolutional neural networks for mobile vision applications. arXiv preprint arXiv:1704.04861, 2017

    2017 arXiv

  16. [24]

    P., Hautier, G., Chen, W., Richards, W

    Jain, A., Ong, S. P., Hautier, G., Chen, W., Richards, W. D., Dacek, S., Cholia, S., Gunter, D., Skinner, D., Ceder, G., et al. Commentary: The materials project: A materials genome approach to accelerating materials innovation. APL materials, 1 0 (1), 2013

    2013

  17. [25]

    H., Khaji, N., and Noorzad, A

    Javaran, S. H., Khaji, N., and Noorzad, A. First kind bessel function (j-bessel) as radial basis function for plane dynamic analysis using dual reciprocity boundary element method. Acta mechanica, 218 0 (3): 0 247--258, 2011

    2011

  18. [26]

    Crystal structure prediction by joint equivariant diffusion

    Jiao, R., Huang, W., Lin, P., Han, J., Chen, P., Lu, Y., and Liu, Y. Crystal structure prediction by joint equivariant diffusion. Advances in Neural Information Processing Systems, 36: 0 17464--17497, 2023

    2023

  19. [27]

    Effect of invariance requirements on the elastic strain energy of crystals with application to the diamond structure

    Keating, P. Effect of invariance requirements on the elastic strain energy of crystals with application to the diamond structure. Physical Review, 145 0 (2): 0 637, 1966

    1966

  20. [28]

    H., Aspuru-Guzik, A., and Jung, Y

    Kim, S., Noh, J., Gu, G. H., Aspuru-Guzik, A., and Jung, Y. Generative adversarial networks for crystal structure prediction. ACS central science, 6 0 (8): 0 1412--1420, 2020

    2020

  21. [29]

    Kingma, D. P. Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114, 2013

    2013 arXiv

  22. [30]

    D., and Vecchi, M

    Kirkpatrick, S., Gelatt Jr, C. D., and Vecchi, M. P. Optimization by simulated annealing. science, 220 0 (4598): 0 671--680, 1983

    1983

  23. [31]

    Clebsch--gordan nets: a fully fourier space spherical convolutional neural network

    Kondor, R., Lin, Z., and Trivedi, S. Clebsch--gordan nets: a fully fourier space spherical convolutional neural network. Advances in Neural Information Processing Systems, 31, 2018

    2018

  24. [32]

    and Smidt, T

    Liao, Y.-L. and Smidt, T. Equiformer: Equivariant graph attention transformer for 3d atomistic graphs. In The Eleventh International Conference on Learning Representations, 2023

    2023

  25. [33]

    General invariance and equilibrium conditions for lattice dynamics in 1d, 2d, and 3d materials

    Lin, C., Ponc \'e , S., and Marzari, N. General invariance and equilibrium conditions for lattice dynamics in 1d, 2d, and 3d materials. npj Computational Materials, 8 0 (1): 0 236, 2022

    2022

  26. [34]

    M., Opahle, I., Zhang, Y., Samathrakis, I., Shen, C., Gutfleisch, O., and Zhang, H

    Long, T., Fortunato, N. M., Opahle, I., Zhang, Y., Samathrakis, I., Shen, C., Gutfleisch, O., and Zhang, H. Constrained crystals deep convolutional generative adversarial network for the inverse design of crystal structures. npj Computational Materials, 7 0 (1): 0 66, 2021

    2021

  27. [35]

    O., Oganov, A

    Lyakhov, A. O., Oganov, A. R., Stokes, H. T., and Zhu, Q. New developments in evolutionary structure prediction algorithm uspex. Computer Physics Communications, 184 0 (4): 0 1172--1182, 2013

    2013

  28. [36]

    S., Aykol, M., Cheon, G., and Cubuk, E

    Merchant, A., Batzner, S., Schoenholz, S. S., Aykol, M., Cheon, G., and Cubuk, E. D. Scaling deep learning for materials discovery. Nature, 624 0 (7990): 0 80--85, 2023

    2023

  29. [37]

    S., Sanchez-Lengeling, B., Gregoire, J

    Noh, J., Kim, J., Stein, H. S., Sanchez-Lengeling, B., Gregoire, J. M., Aspuru-Guzik, A., and Jung, Y. Inverse design of solid-state materials via a continuous representation. Matter, 1 0 (5): 0 1370--1384, 2019

    2019

  30. [38]

    Crystalgan: learning to discover crystallographic structures with generative adversarial networks

    Nouira, A., Sokolovska, N., and Crivello, J.-C. Crystalgan: learning to discover crystallographic structures with generative adversarial networks. arXiv preprint arXiv:1810.11203, 2018

    2018 arXiv

  31. [39]

    A Binded VAE for Inorganic Material Generation

    Oubari, F., de Mathelin, A., D \'e catoire, R., and Mougeot, M. A Binded VAE for Inorganic Material Generation . In NeurIPS 2021 Workshop on Deep Generative Models and Downstream Applications , Online, France, December 2012. doi:10.48550/arXiv.2112.09570

    2021 doi arXiv

  32. [40]

    Pickard, C. J. Airss data for carbon at 10gpa and the c+ n+ h+ o system at 1gpa. (No Title), 2020

    2020

  33. [41]

    Pickard, C. J. and Needs, R. Ab initio random structure searching. Journal of Physics: Condensed Matter, 23 0 (5): 0 053201, 2011

    2011

  34. [42]

    Ren, Z., Noh, J., Tian, S. I. P., Oviedo, F., Xing, G., Liang, Q., Aberle, A. G., Liu, Y., Li, Q., Jayavelu, S., Hippalgaonkar, K., Jung, Y., and Buonassisi, T. Inverse design of crystals using generalized invertible crystallographic representation. ArXiv, abs/2005.07609, 2020

    2005 arXiv

  35. [43]

    Ren, Z., Tian, S. I. P., Noh, J., Oviedo, F., Xing, G., Li, J., Liang, Q., Zhu, R., Aberle, A. G., Sun, S., Wang, X., Liu, Y., Li, Q., Jayavelu, S., Hippalgaonkar, K., Jung, Y., and Buonassisi, T. An invertible crystallographic representation for general inverse design of inor...

    2022 doi

  36. [44]

    u tt, K. T., Sauceda, H. E., Kindermans, P.-J., Tkatchenko, A., and M \

    Sch \"u tt, K. T., Sauceda, H. E., Kindermans, P.-J., Tkatchenko, A., and M \"u ller, K.-R. Schnet--a deep learning architecture for molecules and materials. The Journal of Chemical Physics, 148 0 (24), 2018

    2018

  37. [45]

    and Ermon, S

    Song, Y. and Ermon, S. Generative modeling by estimating gradients of the data distribution. Advances in neural information processing systems, 32, 2019

    2019

  38. [46]

    and Fabritiis, G

    Th \"o lke, P. and Fabritiis, G. D. Equivariant transformers for neural network based molecular potentials. In International Conference on Learning Representations, 2022

    2022

  39. [47]

    E., Kearnes, S

    Thomas, N., Smidt, T. E., Kearnes, S. M., Yang, L., Li, L., Kohlhoff, K., and Riley, P. F. Tensor field networks: Rotation- and translation-equivariant neural networks for 3d point clouds. ArXiv, abs/1802.08219, 2018

    2018 arXiv

  40. [48]

    Se (3)-equivariant prediction of molecular wavefunctions and electronic densities

    Unke, O., Bogojeski, M., Gastegger, M., Geiger, M., Smidt, T., and M \"u ller, K.-R. Se (3)-equivariant prediction of molecular wavefunctions and electronic densities. Advances in Neural Information Processing Systems, 34: 0 14434--14447, 2021

    2021

  41. [49]

    Attention is all you need

    Vaswani, A. Attention is all you need. Advances in Neural Information Processing Systems, 2017

    2017

  42. [50]

    N., Kaiser, L

    Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L. u., and Polosukhin, I. Attention is all you need. In Guyon, I., Luxburg, U. V., Bengio, S., Wallach, H., Fergus, R., Vishwanathan, S., and Garnett, R. (eds.), Advances in Neural Informatio...

    2017

  43. [51]

    Weiler, M., Geiger, M., Welling, M., Boomsma, W., and Cohen, T. S. 3d steerable cnns: Learning rotationally equivariant features in volumetric data. Advances in Neural Information Processing Systems, 31, 2018

    2018

  44. [52]

    C., Zhao, X., Umemoto, K., Wentzcovitch, R., and Ho, K.-M

    Wu, S., Ji, M., Wang, C.-Z., Nguyen, M. C., Zhao, X., Umemoto, K., Wentzcovitch, R., and Ho, K.-M. An adaptive genetic algorithm for crystal structure prediction. Journal of Physics: Condensed Matter, 26 0 (3): 0 035402, 2013

    2013

  45. [53]

    and Grossman, J

    Xie, T. and Grossman, J. C. Crystal graph convolutional neural networks for an accurate and interpretable prediction of material properties. Physical review letters, 120 0 (14): 0 145301, 2018

    2018

  46. [54]

    Crystal diffusion variational autoencoder for periodic material generation

    Xie, T., Fu, X., Ganea, O.-E., Barzilay, R., and Jaakkola, T. Crystal diffusion variational autoencoder for periodic material generation. arXiv preprint arXiv:2110.06197, 2021

    2021 arXiv

  47. [55]

    Yang, S., Cho, K., Merchant, A., Abbeel, P., Schuurmans, D., Mordatch, I., and Cubuk, E. D. Scalable diffusion for materials generation. In The Twelfth International Conference on Learning Representations, 2024

    2024

  48. [56]

    Tripartite interaction representation algorithm for crystal graph neural networks

    Yuan, Y., Chen, Z., Feng, T., Xiong, F., Wang, J., Wang, Y., and Wang, Z. Tripartite interaction representation algorithm for crystal graph neural networks. Scientific Reports, 14 0 (1): 0 24881, 2024

    2024

  49. [57]

    A generative model for inorganic materials design

    Zeni, C., Pinsler, R., Z \"u gner, D., Fowler, A., Horton, M., Fu, X., Wang, Z., Shysheya, A., Crabb \'e , J., Ueda, S., et al. A generative model for inorganic materials design. Nature, pp.\ 1--3, 2025

    2025

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.