Pith. sign in

REVIEW 5 major objections 6 minor 38 references

MatMMFuse: Multi-Modal Fusion model for Material Property Prediction

T0 review · 5 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read The paper claims that fusing a crystal graph encoder with a scientific-text encoder through multi-head cross-attention lowers formation-energy error by 40% over the graph-only model and 68% over the text-only model, while also improving…

desk verdict A plausible attention-based fusion model for crystal properties, but the headline gains rest on a leaky random split and missing baselines. read the letter →

arxiv 2505.04634 v1 pith:HDAK2JDN submitted 2025-04-30 cs.LG cs.CE

Abhiroop Bhattacharya , Sylvain G. Cloutier This is my paper · ORCID
classification cs.LGcs.CE
keywords multi-modalfusionmaterialpropertypredictioncrystalgraphneuralnetworklargelanguagemodelsSciBERTcross-attentionzero-shotformationenergy
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

This paper argues that material-property prediction improves when a model is allowed to see a crystal both as a graph of atoms and as a piece of scientific text. MatMMFuse trains a Crystal Graph Convolution Network and a pretrained SciBERT text encoder end to end, then combines their embeddings with multi-head cross-attention. On the Materials Project test split it reports a mean absolute error of 0.025 eV/atom for formation energy, roughly 40% lower than the graph-only model and 68% lower than the text-only model, with smaller gains on Fermi energy, energy above hull, and band gap. The same trained model also predicts formation energy on perovskite, chalcogenide, and JARVIS test sets without further training, beating both single-modality baselines. A sympathetic reader would take the paper's central claim to be that fusing local structural and global textual information through attention is a broadly useful recipe for materials prediction.

What carries the argument

The load-bearing mechanism is a multi-head cross-attention fusion layer. The text embedding provides the query while the graph embedding provides the keys and values, so the model decides, for each predicted property, which structure-derived features to pull out in response to the semantic content of the text description. The attended vector is passed through a feed-forward head; residual, layer-norm, and dropout components tune stability and generalization. This layer is what lets the model use both local and global information and, according to the ablations, is the component that contributes the largest performance gain.

What would settle it

Recompute the formation-energy MAE using a composition-based split of the 95,582 Materials Project structures (no chemical formula appearing in both train and test) and check that the perovskite, chalcogenide, and JARVIS test entries share no exact or near-identical compositions with the training set. If the MatMMFuse advantage over CGCNN largely disappears under that split, the reported gains are artifacts of the random split.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that cross-attention between a graph embedding and a text embedding produces a joint representation that is more accurate for several DFT-computed properties than either representation alone, and that transfers to unseen chemistry families better than the unimodal models. The authors report MAEs of 0.025 eV/atom for formation energy, 0.44 eV for Fermi energy, 0.029 eV/atom for energy above hull, and 0.31 eV for band gap, against 0.042, 0.60, 0.071, and 0.37 eV for CGCNN and 0.081, 0.59, 0.031, and 0.38 eV for SciBERT. In the zero-shot setting they report MAE reductions of roughly 10% on cubic oxide perovskites, 21% on chalcogenide perovskites, and 48% on the JARVIS subset compared with the better unimodal baseline. They attribute the gain to the attention mechanism's ability to weight local structural features and global text features such as space group and symmetry according to their relevance for the target property.

Load-bearing premise

The reported gains depend on the random 80/10/10 split of the Materials Project data being clean of near-duplicate compositions and on the zero-shot test sets not overlapping training materials; the paper does not describe composition-based splitting or duplicate removal.

Editorial extensions

If this is right

  • Formation-energy MAE drops to 0.025 eV/atom, and the model also beats both unimodal baselines on Fermi energy, energy above hull, and band gap.
  • A single trained model can be applied zero-shot to small specialized datasets such as perovskites, chalcogenides, and JARVIS without retraining, which matters when DFT labels are too expensive to collect.
  • Using a materials-specific text encoder such as MatSciBERT in place of SciBERT further lowers the JARVIS zero-shot MAE, so the choice of text encoder is a tunable lever.
  • The model keeps low training loss when the training set is reduced, indicating that the text channel partially compensates for scarce structural data.
  • The fusion's accuracy degrades sharply when the text input is corrupted, so downstream use should keep the robotic crystallographer output clean.

Reading between the lines

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

  • Because the text descriptions encode composition, symmetry, and space group, part of the zero-shot gain may come from text acting as a soft lookup of known chemistry; ablating the text to remove such phrases would reveal how much of the gain is genuinely structural.
  • The attention weights themselves could be read as a per-property attribution map, telling a researcher whether the model leaned on local bonding or on global symmetry for a given prediction.
  • The same fusion recipe should extend to other graph/text encoder pairs and to additional modalities such as diffraction patterns, provided the cross-attention's quadratic cost is acceptable.
  • The reported margins are only as clean as the data split; a composition-aware split would test whether the 40% improvement survives removal of near-duplicate train/test structures.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 6 minor

Summary. The paper proposes MatMMFuse, a multi-modal fusion model that combines a CGCNN graph encoder with a SciBERT text encoder using multi-head cross-attention, trained end-to-end on Materials Project data for formation energy, band gap, energy above hull, and Fermi energy prediction. The authors report an MAE of 0.025 eV/atom for formation energy on a held-out test set, a 40% improvement over vanilla CGCNN and 68% over vanilla SciBERT, and they report zero-shot results on cubic oxide perovskites, chalcogenides, and a JARVIS subset. Ablation studies examine alternate BERT and GNN encoders, fusion-module components, training-data size, and text-input corruption. The paper includes a public code repository and pseudocode for the proposed framework.

Significance. If validated, the paper would make a modest but useful contribution: a simple cross-attention fusion of structural and textual modalities for materials property prediction, with a plausible mechanism for combining local graph information with global symmetry/text information. The release of code, the inclusion of pseudocode, and the ablation coverage are strengths, and the explicit limitations section is honest. However, the current evaluation protocol does not yet establish the central claims: the random split is likely to leak near-duplicate compositions between train and test, the zero-shot benchmarks are not shown to be disjoint from training data, no concatenation baseline is tested despite the paper's stated motivation, and all results come from single runs with no uncertainty quantification.

major comments (5)
  1. [Section 3.1, Table 1] Section 3.1 describes only an 80/10/10 random split of 95,582 Materials Project structures, with no composition-based splitting, deduplication, or overlap filtering. Because the Materials Project contains multiple entries with identical or nearly identical reduced formulas (different magnetic orderings, DFT settings, polymorphs), this protocol can place near-duplicate compositions in both training and test sets; the reported in-domain MAE improvements in Table 1 may therefore reflect memorization of composition-specific shortcuts rather than generalization. Please re-evaluate using a split by unique reduced formula (or an explicit similarity threshold) and report test-set overlap statistics.
  2. [Section 4.2, Table 2] Section 4.2 and Table 2 report 'zero-shot' MAEs for perovskites, chalcogenides, and JARVIS, but the paper does not quantify how many training-set compositions overlap with these external sets. Given that Materials Project contains many ABO3, AB(S,Se)3, and JARVIS-like compounds, the zero-shot gains over the vanilla models are not yet supported; report the exact intersection between the training set and each external benchmark, and re-run after removing overlapping entries.
  3. [Section 2.3, Section 4.3.3] The contribution is presented as a cross-attention fusion that improves over static concatenation, but Sections 2.3 and 4.3.3 provide no concatenation baseline. Add a baseline that concatenates the two embeddings and feeds them to the same predictor (or an existing concatenation model such as CrysMMNet) under identical training conditions; without it, the claimed benefit of the multi-head attention module is not demonstrated.
  4. [Section 4.1, Tables 1 and 2] All quantitative comparisons in Sections 4.1 and 4.2 are based on single runs with no error bars or significance testing. The headline 40% improvement over CGCNN could be within run-to-run variance; report mean ± standard deviation over at least three random seeds for every model and dataset.
  5. [Table 3, Section 4.3.1] Table 3 is internally inconsistent with Table 2: the SciBERT column lists 1.28 eV/atom for perovskites and 0.98 for chalcogenides, whereas Table 2 reports SciBERT values of 2.84 and 1.44 and the proposed model values of 1.28 and 1.05 for those datasets. Either the columns or the entries are mislabeled; this undermines the encoder-ablation conclusions in Section 4.3.1.
minor comments (6)
  1. [Abstract, Section 1, Section 5] The reported improvement over CGCNN for formation energy is 40% in the abstract and Section 4.1, but 35% in the introduction and conclusion; make these numbers consistent.
  2. [Equation (1)] Equation (1) and the surrounding text contain a notation error: the text says 'the concatenation of h(l)i and h(j)i' but should read h(l)i and h(l)j; please correct the superscripts.
  3. [Section 4.1] The text says the model predicts 'four important material properties' but then lists only three (formation energy, Fermi energy, band gap), omitting energy above hull, and labels Fermi energy as Eg; clarify the list and symbols.
  4. [Section 4.2, Tables 2 and 3] State explicitly which target property is being predicted in the zero-shot tables (presumably formation energy per atom) and specify how the JARVIS subset was chosen and whether it was filtered to the same target.
  5. [Section 4.3.4, Figure 8] The robustness-to-training-size experiment plots only training loss; report test MAE as a function of training-set size to support the claim of robustness.
  6. [Section 1, Section 4.1] The introduction states that MatMMFuse 'performs in line with state of the art models,' but no state-of-the-art baselines (e.g., MEGNet, SchNet, Wrenformer, or recent fusion models) are included in Table 1; add such comparisons or soften the claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: MatMMFuse's claims are empirical benchmark results, not derivations that reduce to fitted inputs or self-citations.

full rationale

MatMMFuse is an empirical machine-learning paper. The claimed improvements (lower MAE for formation energy, band gap, energy above hull, and Fermi energy) come from training an end-to-end model on a fixed Materials Project split and comparing against vanilla CGCNN and SciBERT baselines on held-out and external datasets. I inspected the architecture and training description: the CGCNN encoder, SciBERT text encoder, and multi-head cross-attention fusion are standard components; no parameter is fitted to the test set and then renamed as a prediction, and no equation defines the target in terms of the output. The comparisons with vanilla models are legitimate baselines, not circular constructions. The paper cites prior work for CGCNN and SciBERT, and these are external, well-established models; there is no load-bearing self-citation chain or imported uniqueness theorem. The main evaluation weakness is that Section 3.1 describes only an "80%,10%,10% train, validation and test split" of 95,582 Materials Project structures without composition-based deduplication, so near-duplicate compositions may leak between train and test and inflate the in-domain and zero-shot numbers. That is a data-split and benchmarking concern, and it is correctly classified as a correctness risk rather than circularity. Under the hard rules, circularity requires exhibiting a specific reduction of a claimed result to its inputs; no such reduction exists in this paper, so the appropriate finding is no significant circularity (score 0).

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claim rests on standard ML training and evaluation assumptions: the data split is representative, the external datasets are comparable, and the baselines are fair. The model introduces no new physical entities and has no hand-fitted scientific constants; its parameters are learned from data. The main fragility is the random split and the unstated hyperparameters.

free parameters (4)
  • Learning rate and optimizer hyperparameters = not reported
    AdamW with cosine warmup is mentioned but the learning rate, weight decay, and warmup steps are not given; these affect the final MAE.
  • Batch size and number of training epochs = not reported
    Not stated in the paper or pseudocode; the results depend on the training budget.
  • Fusion module hyperparameters = not reported
    Number of attention heads, projection dimensions, and dropout rate in the fusion layer are not specified.
  • Robocrystallographer text generation settings = defaults assumed
    The text descriptions that drive the SciBERT branch are generated with Robocrystallographer, but no parameters are given; the zero-shot results depend on this text.
assumptions (4)
  • domain assumption Random 80/10/10 split of Materials Project structures yields a test set that is not contaminated by near-duplicate compositions.
    No composition-based splitting or duplicate removal is mentioned in Section 3.1; random splits of materials databases typically overestimate generalization.
  • domain assumption SciBERT text embeddings of Robocrystallographer descriptions carry global structural information (space group, symmetry) that is complementary to the CGCNN graph embedding.
    The paper's motivation in Sections 1 and 2.3 rests on this complementarity; no quantitative evidence of complementarity is supplied.
  • domain assumption The DFT-computed property values in the external zero-shot datasets are directly comparable to Materials Project values.
    Perovskite, chalcogenide, and JARVIS datasets may use different DFT codes or settings; the paper does not discuss calibration in Section 4.2.
  • domain assumption Baseline CGCNN and SciBERT were trained under conditions comparable to MatMMFuse (same data, same splits, same hyperparameter budget).
    No details are given for the baselines in Section 4.1; unfair baselines would explain the observed improvement.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MatMMFuse: Multi-Modal Fusion model for Material Property Prediction." pith.science (2026). https://pith.science/paper/HDAK2JDN

@misc{pith2026250504634,
  author       = {Pith},
  title        = {Pith review of: MatMMFuse: Multi-Modal Fusion model for Material Property Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HDAK2JDN}},
  note         = {Machine review of arXiv:2505.04634}
}
read the original abstract

The recent progress of using graph based encoding of crystal structures for high throughput material property prediction has been quite successful. However, using a single modality model prevents us from exploiting the advantages of an enhanced features space by combining different representations. Specifically, pre-trained Large language models(LLMs) can encode a large amount of knowledge which is beneficial for training of models. Moreover, the graph encoder is able to learn the local features while the text encoder is able to learn global information such as space group and crystal symmetry. In this work, we propose Material Multi-Modal Fusion(MatMMFuse), a fusion based model which uses a multi-head attention mechanism for the combination of structure aware embedding from the Crystal Graph Convolution Network (CGCNN) and text embeddings from the SciBERT model. We train our model in an end-to-end framework using data from the Materials Project Dataset. We show that our proposed model shows an improvement compared to the vanilla CGCNN and SciBERT model for all four key properties: formation energy, band gap, energy above hull and fermi energy. Specifically, we observe an improvement of 40% compared to the vanilla CGCNN model and 68% compared to the SciBERT model for predicting the formation energy per atom. Importantly, we demonstrate the zero shot performance of the trained model on small curated datasets of Perovskites, Chalcogenides and the Jarvis Dataset. The results show that the proposed model exhibits better zero shot performance than the individual plain vanilla CGCNN and SciBERT model. This enables researchers to deploy the model for specialized industrial applications where collection of training data is prohibitively expensive.

Figures

Figures reproduced from arXiv: 2505.04634 by the authors.

Figure 1
Figure 1. The figure provides an overview of MatMMFuse. The CGCNN model generates a struc [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. The scatter plot presents the actual versus predicted values for (a) Formation Energy [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. The t-SNE plot of the embedding from the embedding for the test dataset from Materials [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: The scatter plot presents a comparison between the actual and the predicted values of [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: The plot compares the performance of different BERT models for encoding the text [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: The plot compares the performance of different GNN models for encoding the lattice [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: The waterfall chart shows the effect of adding individual components to improve the [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: The chart shows that MatMMFuse is relatively robust to reduction in training data. [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: The figure (9a) and figure (9b) shows the effect of the corruption of the input text on the [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 22 canonical work pages

  1. [1]

    Convolutional neural network of atomic surface structures to predict binding energies for high-throughput screening of catalysts

    Seoin Back, Junwoong Yoon, Nianhan Tian, Wen Zhong, Kevin Tran, and Zachary W Ulissi. Convolutional neural network of atomic surface structures to predict binding energies for high-throughput screening of catalysts. The journal of physical chemistry letters, 10 0 (15): 0 4401--4408, 2019

    work page 2019

  2. [2]

    Chalcogenide perovskites (abs3; a= ba, ca, sr; b= hf, sn): An emerging class of semiconductors for optoelectronics

    Pooja Basera and Saswata Bhattacharya. Chalcogenide perovskites (abs3; a= ba, ca, sr; b= hf, sn): An emerging class of semiconductors for optoelectronics. The Journal of Physical Chemistry Letters, 13 0 (28): 0 6439--6446, 2022

    work page 2022

  3. [3]

    Atom-centered symmetry functions for constructing high-dimensional neural network potentials

    J \"o rg Behler. Atom-centered symmetry functions for constructing high-dimensional neural network potentials. The Journal of chemical physics, 134 0 (7), 2011

    work page 2011

  4. [4]

    Scibert: A pretrained language model for scientific text

    Iz Beltagy, Kyle Lo, and Arman Cohan. Scibert: A pretrained language model for scientific text. arXiv preprint arXiv:1903.10676, 2019

    arXiv 1903

  5. [5]

    Graph networks as a universal machine learning framework for molecules and crystals

    Chi Chen, Weike Ye, Yunxing Zuo, Chen Zheng, and Shyue Ping Ong. Graph networks as a universal machine learning framework for molecules and crystals. Chemistry of Materials, 31 0 (9): 0 3564--3572, 2019

    2019

  6. [6]

    A critical review of machine learning of energy materials

    Chi Chen, Yunxing Zuo, Weike Ye, Xiangguo Li, Zhi Deng, and Shyue Ping Ong. A critical review of machine learning of energy materials. Advanced Energy Materials, 10 0 (8): 0 1903242, 2020

    work page 2020

  7. [7]

    The joint automated repository for various integrated simulations (jarvis) for data-driven materials design

    Kamal Choudhary, Kevin F Garrity, Andrew CE Reid, Brian DeCost, Adam J Biacchi, Angela R Hight Walker, Zachary Trautt, Jason Hattrick-Simpers, A Gilad Kusne, Andrea Centrone, et al. The joint automated repository for various integrated simulations (jarvis) for data-driven materials design. npj computational materials, 6 0 (1): 0 173, 2020

    work page 2020

  8. [8]

    Crysmmnet: multimodal representation for crystal property prediction

    Kishalay Das, Pawan Goyal, Seung-Cheol Lee, Satadeep Bhattacharjee, and Niloy Ganguly. Crysmmnet: multimodal representation for crystal property prediction. In Uncertainty in Artificial Intelligence, pp.\ 507--517. PMLR, 2023

    work page 2023

Show all 38 references
  1. [9]

    Comparing molecules and solids across structural and alchemical space

    Sandip De, Albert P Bart \'o k, G \'a bor Cs \'a nyi, and Michele Ceriotti. Comparing molecules and solids across structural and alchemical space. Physical Chemistry Chemical Physics, 18 0 (20): 0 13754--13769, 2016

    2016

  2. [10]

    The nomad laboratory: from data sharing to artificial intelligence

    Claudia Draxl and Matthias Scheffler. The nomad laboratory: from data sharing to artificial intelligence. Journal of Physics: Materials, 2 0 (3): 0 036001, 2019

    2019

  3. [11]

    Crystal structure representations for machine learning models of formation energies

    Felix Faber, Alexander Lindmaa, O Anatole Von Lilienfeld, and Rickard Armiento. Crystal structure representations for machine learning models of formation energies. International Journal of Quantum Chemistry, 115 0 (16): 0 1094--1101, 2015

    2015

  4. [12]

    Benchmarking graph neural networks for materials chemistry

    Victor Fung, Jiaxin Zhang, Eric Juarez, and Bobby G Sumpter. Benchmarking graph neural networks for materials chemistry. npj Computational Materials, 7 0 (1): 0 84, 2021

    2021

  5. [13]

    Robocrystallographer: automated crystal structure text descriptions and analysis

    Alex M Ganose and Anubhav Jain. Robocrystallographer: automated crystal structure text descriptions and analysis. MRS Communications, 9 0 (3): 0 874--881, 2019

    2019

  6. [14]

    A new model for learning in graph domains

    Marco Gori, Gabriele Monfardini, and Franco Scarselli. A new model for learning in graph domains. In Proceedings. 2005 IEEE International Joint Conference on Neural Networks, 2005., volume 2, pp.\ 729--734. IEEE, 2005

    2005

  7. [15]

    Structure and properties of perovskite oxides

    Tatsumi Ishihara. Structure and properties of perovskite oxides. Perovskite Oxide for Solid Oxide Fuel Cells, pp.\ 1--16, 2009

    2009

  8. [16]

    14 examples of how llms can transform materials science and chemistry: a reflection on a large language model hackathon

    Kevin Maik Jablonka, Qianxiang Ai, Alexander Al-Feghali, Shruti Badhwar, Joshua D Bocarsly, Andres M Bran, Stefan Bringuier, L Catherine Brinson, Kamal Choudhary, Defne Circi, et al. 14 examples of how llms can transform materials science and chemistry: a reflection on a large...

    2023

  9. [17]

    The materials project: A materials genome approach to accelerating materials innovation, apl mater

    Anubhav Jain, Shyue Ping Ong, Geoffroy Hautier, Wei Chen, William Davidson Richards, Stephen Dacek, Shreyas Cholia, Dan Gunter, David Skinner, Gerbrand Ceder, et al. The materials project: A materials genome approach to accelerating materials innovation, apl mater. Applied Phy...

    2013

  10. [18]

    Is bert really robust? a strong baseline for natural language attack on text classification and entailment

    Di Jin, Zhijing Jin, Joey Tianyi Zhou, and Peter Szolovits. Is bert really robust? a strong baseline for natural language attack on text classification and entailment. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pp.\ 8018--8025, 2020

    2020

  11. [19]

    Albert: A lite bert for self-supervised learning of language representations

    Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, and Radu Soricut. Albert: A lite bert for self-supervised learning of language representations. arXiv preprint arXiv:1909.11942, 2019

    1909 arXiv

  12. [20]

    Cast: Cross attention based multimodal fusion of structure and text for materials property prediction

    Jaewan Lee, Changyoung Park, Hongjun Yang, Sungbin Lim, and Sehui Han. Cast: Cross attention based multimodal fusion of structure and text for materials property prediction. arXiv preprint arXiv:2502.06836, 2025

    2025 arXiv

  13. [21]

    Evaluating graph neural networks for link prediction: Current pitfalls and new benchmarking

    Juanhui Li, Harry Shomer, Haitao Mao, Shenglai Zeng, Yao Ma, Neil Shah, Jiliang Tang, and Dawei Yin. Evaluating graph neural networks for link prediction: Current pitfalls and new benchmarking. Advances in Neural Information Processing Systems, 36, 2024

    2024

  14. [22]

    Hybrid-llm-gnn: integrating large language models and graph neural networks for enhanced materials property prediction

    Youjia Li, Vishu Gupta, Muhammed Nur Talha Kilic, Kamal Choudhary, Daniel Wines, Wei-keng Liao, Alok Choudhary, and Ankit Agrawal. Hybrid-llm-gnn: integrating large language models and graph neural networks for enhanced materials property prediction. Digital Discovery, 2025

    2025

  15. [23]

    Robert--a romanian bert model

    Mihai Masala, Stefan Ruseti, and Mihai Dascalu. Robert--a romanian bert model. In Proceedings of the 28th International Conference on Computational Linguistics, pp.\ 6626--6637, 2020

    2020

  16. [24]

    Unimat: Unifying materials embeddings through multi-modal learning

    Janghoon Ock, Joseph Montoya, Daniel Schweigert, Linda Hung, Santosh K Suram, and Weike Ye. Unimat: Unifying materials embeddings through multi-modal learning. arXiv preprint arXiv:2411.08664, 2024

    2024 arXiv

  17. [25]

    Toward predicting intermetallics surface properties with high-throughput dft and convolutional neural networks

    Aini Palizhati, Wen Zhong, Kevin Tran, Seoin Back, and Zachary W Ulissi. Toward predicting intermetallics surface properties with high-throughput dft and convolutional neural networks. Journal of chemical information and modeling, 59 0 (11): 0 4742--4749, 2019

    2019

  18. [26]

    Automatic differentiation in pytorch

    Adam Paszke, Sam Gross, Soumith Chintala, Gregory Chanan, Edward Yang, Zachary DeVito, Zeming Lin, Alban Desmaison, Luca Antiga, and Adam Lerer. Automatic differentiation in pytorch. Neural Information Processing Systems, 2017

    2017

  19. [27]

    Materials design and discovery with high-throughput density functional theory: the open quantum materials database (oqmd)

    James E Saal, Scott Kirklin, Muratahan Aykol, Bryce Meredig, and Christopher Wolverton. Materials design and discovery with high-throughput density functional theory: the open quantum materials database (oqmd). Jom, 65: 0 1501--1509, 2013

    2013

  20. [28]

    Distilbert, a distilled version of bert: smaller, faster, cheaper and lighter

    Victor Sanh, Lysandre Debut, Julien Chaumond, and Thomas Wolf. Distilbert, a distilled version of bert: smaller, faster, cheaper and lighter. arXiv preprint arXiv:1910.01108, 2019

    1910 arXiv

  21. [29]

    The graph neural network model

    Franco Scarselli, Marco Gori, Ah Chung Tsoi, Markus Hagenbuchner, and Gabriele Monfardini. The graph neural network model. IEEE transactions on neural networks, 20 0 (1): 0 61--80, 2008

    2008

  22. [30]

    Recent advances and applications of machine learning in solid-state materials science

    Jonathan Schmidt, M \'a rio RG Marques, Silvana Botti, and Miguel AL Marques. Recent advances and applications of machine learning in solid-state materials science. npj Computational Materials, 5 0 (1): 0 83, 2019

    2019

  23. [31]

    u tt, Huziel E Sauceda, P-J Kindermans, Alexandre Tkatchenko, and K-R M \

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

    2018

  24. [32]

    Stacked debert: All attention in incomplete data for text classification

    Gwenaelle Cunha Sergio and Minho Lee. Stacked debert: All attention in incomplete data for text classification. Neural Networks, 136: 0 87--96, 2021

    2021

  25. [33]

    Visualizing data using t-sne

    Laurens Van der Maaten and Geoffrey Hinton. Visualizing data using t-sne. Journal of machine learning research, 9 0 (11), 2008

    2008

  26. [34]

    Crystal graph convolutional neural networks for an accurate and interpretable prediction of material properties

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

    2018

  27. [35]

    Predicting the band gaps of inorganic solids by machine learning

    Ya Zhuo, Aria Mansouri Tehrani, and Jakoah Brgoch. Predicting the band gaps of inorganic solids by machine learning. The journal of physical chemistry letters, 9 0 (7): 0 1668--1673, 2018

    2018

  28. [36]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  29. [37]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  30. [38]

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.