Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

Energy & Force Regression on DFT Trajectories is Not Enough for Universal Machine Learning Interatomic Potentials

T0 review · 3 major / 6 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read This paper argues that energy and force regression on DFT trajectories cannot alone produce universal machine learning interatomic potentials, and that the field must shift to higher-accuracy data, simulation-informed metrology, and…

desk verdict A coherent, honest position paper arguing MLIPs should look beyond DFT trajectories; the priority shift to CCSD(T) data is plausible but not demonstrated. read the letter →

arxiv 2502.03660 v1 pith:Q3AFQ2M6 submitted 2025-02-05 cond-mat.mtrl-sci cs.AIcs.LG

Santiago Miret , Kin Long Kelvin Lee , Carmelo Gonzales , Sajid Mannan , N. M. Anoop Krishnan This is my paper · ORCID
classification cond-mat.mtrl-scics.AIcs.LG
keywords machinelearninginteratomicpotentialsdensityfunctionaltheorycoupledclustermoleculardynamicsdevice-scalesimulationMLIPmetrologytrainingdataqualityuniversal
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

The paper argues that universal machine learning interatomic potentials (MLIPs) — fast models that predict atomic energies and forces for molecular dynamics — are currently trained by regressing energies and forces on density functional theory (DFT) trajectories and judged by error metrics, and that this recipe is not enough for reliable device-scale simulation. It claims that DFT labels carry known inaccuracies and code-to-code inconsistencies that propagate into MLIP predictions, so model improvements are gated by data quality. It therefore recommends three research priorities: generating training data with more accurate methods such as coupled cluster theory, building metrology that tests MLIPs in realistic MD simulations against experimental properties, and engineering computationally efficient inference workflows. If these directions are right, progress depends less on model architecture and more on the quality of reference data and evaluation.

What carries the argument

The argument is carried by an accuracy ladder of quantum chemistry: DFT at roughly O($N^{3}$) to O($N^{5}$) cost as the current label source, coupled cluster theory, specifically CCSD(T) at O($N^{7}$), as the 'gold standard' reference, and full configuration interaction at O(N!) as the exact nonrelativistic limit. This ladder frames data quality as the limiting factor: a MLIP trained on DFT can at best reproduce DFT, and its prediction errors inherit DFT's systematic errors and code-specific biases. The same ladder anchors the paper's central recommendation to bootstrap high-volume, lower-accuracy DFT data into low-volume, higher-accuracy CCSD(T) data.

What would settle it

Train the same MLIP architecture on matching structures using PBE-DFT labels and CCSD(T) labels, then run identical MD simulations and compare predicted experimental properties such as lattice parameters, thermal expansion, or elastic moduli; if the DFT-trained model matches the higher-accuracy model and experiment, the paper's load-bearing premise is falsified. A cheaper check is to compute energies and forces for a small set of molecules with two DFT codes using the same functional and see whether code-to-code differences are negligible relative to model error.

Watch

Extended reading notes

Core claim

The central claim is that the standard MLIP recipe — regress energy and forces from DFT trajectories, report energy and force MAEs, and scale up architecture and dataset size — has a ceiling set by the reference method itself. The paper identifies three failure modes: DFT's limited accuracy (band gaps, fractional charges, strong correlation, adsorption-energy errors on the order of 0.5 eV), reproducibility gaps between DFT codes where the same functional yields different energies and forces, and data coverage bias, as in the MPTrj dataset missing 89 elements and concentrating on binary and ternary ideal bulk crystals. Evaluation also misleads: models with similar static energy and force errors diverge in MD tasks such as structural relaxation, and current MLIP inference is slower than classical potentials, making billion-atom device simulations impractical. The authors propose that universal MLIPs should instead target higher-accuracy wavefunction-based labels, simulation-informed metrology and interpretability, and performant inference workflows.

Load-bearing premise

The paper assumes that DFT label noise, meaning its known errors and code-to-code inconsistencies, is the main thing holding universal MLIPs back, rather than model capacity, data coverage, or inference cost, but it does not run a controlled test comparing MLIPs trained on DFT labels versus higher-accuracy labels on the same materials.

Editorial extensions

If this is right

  • If the paper is correct, the current generation of universal MLIPs trained solely on DFT trajectories has a practical accuracy ceiling around DFT, so further scaling of models and datasets on the same label source will not deliver device-scale reliability.
  • Evaluation would shift from static energy and force error metrics toward simulation-based metrology: structural relaxation, MD stability, and predicted experimentally measured properties.
  • New training-data efforts would target underrepresented chemical spaces and realistic conditions such as defects, interfaces, temperature, pressure, and phase changes, rather than only ideal bulk crystals.
  • Inference efficiency becomes a first-class research objective, because separating model accuracy from simulation throughput is necessary to reach device-scale systems such as transistors.
  • Hybrid ML plus quantum mechanics methods that solve higher-accuracy equations with neural ansatze, rather than only regressing energy and force labels, offer a route to properties beyond the current training targets.

Reading between the lines

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

  • Beyond the paper: a controlled head-to-head comparison — the same architecture and structures, with labels from PBE-DFT versus CCSD(T) — would directly measure how much of the MLIP error ceiling is set by the reference method.
  • Beyond the paper: if DFT label noise is the bottleneck, data distillation from large DFT trajectory sets into targeted high-accuracy labels could capture most of the benefit at a fraction of the CCSD(T) cost.
  • Beyond the paper: the metrology argument implies that benchmark leaderboards dominated by energy and force MAE may overstate progress, and simulation-property tests would likely reshuffle model rankings.
  • Beyond the paper: the same pattern, where surrogate models inherit the error of their training reference, is testable as a general principle in other physics surrogate modeling settings.
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

3 major / 6 minor

Summary. This paper is a position statement arguing that universal machine learning interatomic potentials (MLIPs) currently fall short of device-scale materials simulation because they are trained by regressing energies and forces on DFT trajectories. The authors propose three research priorities: (1) generating higher-accuracy training data with methods such as CCSD(T), (2) developing MLIP metrology tools for large-scale benchmarking, visualization, and interpretability, and (3) building computationally efficient inference workflows. The paper surveys limitations of DFT, data coverage gaps in current datasets, inadequacies of common evaluation metrics, and computational bottlenecks, and it formulates a set of concrete community challenges.

Significance. If the central claim holds, the paper usefully reframes MLIP research away from architecture-centric improvements toward data quality and evaluation methodology. Its strengths are a clear structure, a broad synthesis of the DFT and MLIP literatures, an original data-coverage analysis in Figure 3, and falsifiable recommendations (e.g., training the same architecture on CCSD(T)-level labels for matched systems should reduce end-to-end simulation error). The paper contains no new measurements, derivations, or fitted parameters, so it must be judged as a perspective rather than a research contribution. Its main weakness is that the load-bearing premise, namely that DFT label noise propagates to dominate MLIP prediction error, is asserted rather than demonstrated.

major comments (3)
  1. [Sec. 3.1, 'Limitations of DFT for Training Data Generation'] The sentence 'These errors can then further propagate to MLIP-based property prediction models' (Sec. 3.1) is the pivotal link between known DFT inaccuracies and the paper's first recommendation, but no evidence is provided that this propagation is the dominant error source in end-to-end MLIP simulations. The cited studies (Araujo et al. 2022; Lejaeghere et al. 2016; Bosoni et al. 2024) establish DFT's systematic inaccuracies and code-to-code variance for DFT itself, not the effect of those errors after ML training. The paper also acknowledges that data diversity, training objectives, and inference stability are critical (Secs. 3.2, 4, 5); without a controlled comparison of the same architecture trained on DFT versus CCSD(T)-level labels for matched materials, the title's claim that DFT trajectory regression is 'not enough' is underdetermined. The authors should either add such a comparison, even on a small set, or explicitly reframe the claim as a testable hypothesis.
  2. [Sec. 6, 'Alternative Viewpoints & Approaches' (AV1)] The dismissal of the 'DFT is good enough' viewpoint rests on the speculative sentence 'eventually the limitations of DFT as the source of the data will become the limitation for the underlying MLIP' (Sec. 6, AV1). The paper acknowledges that DFT-based MLIPs have shown 'reasonable generalization' and practical success in screening, but it provides no quantitative failure threshold and no comparison with the cost of generating CCSD(T) data. To make the central claim convincing, the authors should specify which observable property or benchmark would discriminate between the two hypotheses (e.g., a matched comparison on interfacial or defect systems where DFT is known to fail), rather than relying on an eventual-limitation argument.
  3. [Sec. 5, 'Computational & Modeling Considerations' (Figure 4)] The estimate that reaching one millisecond of simulation time on a device-scale system would require ~10^5 years is presented without derivation. The text cites 'empirical strong scaling results on 10^8 atoms by Kozinsky et al. (2023) with 1024 nodes of Nvidia A100', but it does not state the assumed timestep, integration scheme, atoms-per-node efficiency, or how the extrapolation to the device-scale regime (e.g., '~30,000 atoms per transistor' for an Intel 8086 versus 'hundreds of billions' of atoms for modern devices) is computed. As this number is used to motivate the third research priority, the estimate should be made reproducible or replaced by a range with explicit assumptions.
minor comments (6)
  1. [Abstract] The abstract contains two typos: 'Interactomic' should be 'Interatomic', and 'aargue' should be 'argue'.
  2. [Sec. 3.1] There are several typographical errors: 'viz-a-viz' should be 'vis-à-vis', 'repoducibility' should be 'reproducibility', and 'previsouly' should be 'previously'.
  3. [Fig. 3 caption] The caption spells the dataset as 'MPtraj'; elsewhere in the text it is 'MPtrj'. Please standardize.
  4. [Appendix B] The appendix has typos: 'Bechmarking' should be 'Benchmarking', 'parameteres' should be 'parameters', and 'occuring' should be 'occurring'. The dataset name is also inconsistently written as 'Mptrj'.
  5. [Fig. 2] The complexity labels O(N5), O(N7), O(N9), O(N!) are ambiguous; use superscripts (O(N^5), O(N^7), etc.) for readability.
  6. [Sec. 5] The word 'langauges' should be 'languages'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: this position paper's recommendations rest on external benchmarks and standard quantum-chemistry facts, not on self-referential derivations.

full rationale

This is a position/opinion paper rather than a derivation: it contains no fitted parameters, no equations that reduce to their inputs, and no 'prediction' that is defined in terms of the target claim. The central thesis (DFT-trajectory regression is insufficient for universal MLIPs, so higher-accuracy data such as CCSD(T) should be pursued) is supported by external evidence: Araujo et al. (2022) and Lejaeghere et al. (2016) document DFT accuracy and reproducibility limitations, and the dataset coverage analysis in Section 3.2 is computed from MPtrj and AMCSD. The paper does cite prior work co-authored by its own authors (Bihani et al. 2024a; Gonzales et al. 2024; Lee et al. 2024b) for MLIP failure and interpretability narratives, but those are concrete empirical benchmarks that are externally falsifiable and not simply restatements of the paper's conclusion. The weakest point is evidentiary, not circular: Section 3.1 asserts 'These errors can then further propagate to MLIP-based property prediction models' without a controlled DFT-versus-CCSD(T) training comparison, so the priority claim is underdetermined; this is a correctness/evidence concern, not a self-definitional reduction. No step in the paper's argument is equivalent by construction to its own inputs.

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

As a perspective, the paper introduces no free parameters or invented entities. Its argument rests on domain assumptions about DFT limitations, error propagation, and benchmark validity, most of which are plausible and literature-supported but not proven here.

assumptions (4)
  • domain assumption The accuracy of MLIPs is ultimately gated by the accuracy of the simulation method used to generate training data.
    Stated in Section 3.1: 'MLIP performance is ultimately gated by the availability of high quality data, which necessitates more accurate simulation methods'. This is load-bearing for the recommendation but not empirically demonstrated.
  • domain assumption DFT errors of ~0.5 eV in adsorption energies materially degrade MLIP-based predictions.
    Invoked in Section 3.1 via citing Araujo et al.; the propagation to MLIPs is asserted, not derived.
  • domain assumption Static energy/force MAE benchmarks do not reflect MLIP performance in MD simulations.
    Stated in Appendix B, citing Bihani et al. and Gonzales et al. This underpins the metrology recommendation but is based on limited comparative studies.
  • domain assumption CCSD(T) is a suitable reference method for generating training data for diverse solid-state materials at scale.
    Recommended in Section 3.1 and Challenge 3.1.1; feasibility for periodic and diverse systems is only supported by a few recent works (e.g., Herzog et al. 2024).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Energy & Force Regression on DFT Trajectories is Not Enough for Universal Machine Learning Interatomic Potentials." pith.science (2026). https://pith.science/paper/Q3AFQ2M6

@misc{pith2026250203660,
  author       = {Pith},
  title        = {Pith review of: Energy & Force Regression on DFT Trajectories is Not Enough for Universal Machine Learning Interatomic Potentials},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Q3AFQ2M6}},
  note         = {Machine review of arXiv:2502.03660}
}
read the original abstract

Universal Machine Learning Interactomic Potentials (MLIPs) enable accelerated simulations for materials discovery. However, current research efforts fail to impactfully utilize MLIPs due to: 1. Overreliance on Density Functional Theory (DFT) for MLIP training data creation; 2. MLIPs' inability to reliably and accurately perform large-scale molecular dynamics (MD) simulations for diverse materials; 3. Limited understanding of MLIPs' underlying capabilities. To address these shortcomings, we aargue that MLIP research efforts should prioritize: 1. Employing more accurate simulation methods for large-scale MLIP training data creation (e.g. Coupled Cluster Theory) that cover a wide range of materials design spaces; 2. Creating MLIP metrology tools that leverage large-scale benchmarking, visualization, and interpretability analyses to provide a deeper understanding of MLIPs' inner workings; 3. Developing computationally efficient MLIPs to execute MD simulations that accurately model a broad set of materials properties. Together, these interdisciplinary research directions can help further the real-world application of MLIPs to accurately model complex materials at device scale.

Figures

Figures reproduced from arXiv: 2502.03660 by the authors.

Figure 1
Figure 1. Overview of Machine Learning Interatomic Potentials (MLIP) requirements for device scale modeling. Current research focuses mainly on bulk structures in ideal conditions with regression-based training and error metric evaluation. To enable materials foundation models, we require higher quality training datasets that use more accurate simulation methods like Coupled Cluster Theory. MLIPs, in turn, should be evaluated… view at source ↗
Figure 2
Figure 2. An artistic interpretation of Jacob’s ladder (Perdew et al., 2005), extended to include wavefunction methods. Going up the ladders improves both accuracy and precision, but is generally pro￾portional to increased time complexity. We stress that, while there are many nuances associated with method choice for atomistic systems, particularly with multireference methods like CASSCF (Roos et al., 1980), these accuracy tr… view at source ↗
Figure 3
Figure 3. Frequency of elements in MPtrj dataset. The color bar in the figure represents a logarithmic scale ranging from low to high values, with the corresponding numbers indicating the frequency of each element’s presence in the MPtraj dataset. In addition to the aforementioned limitations on DFT accu￾racy, the limited set of systems studied imposes additional constraints on MLIP models generalizing to new designs. This is… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Schematic of length and time scales relevant to materials modeling. Annotations indicate approximate computational re￾quirements; each block region corresponds to the amount of effort required for the corresponding scale of compute within a “timely” fashion. Routine an…
Figure 5
Figure 5. Figure 5: Training and inference times for different MLIP architectures on the LiPS dataset based on an analysis from Bihani et al. (2024a). The MLIPs architectures include: MACE (Batatia et al., 2023), BotNet (Batatia et al., 2025), Allegro (Musaelian et al., 2023), Equiformer …
Figure 6
Figure 6. Figure 6: Normalized element frequency of Mptrj (Deng et al., 2023) containing DFT simulated and naturally occuring minerals in AMCSD (Downs & Hall-Wallace, 2003) that are experimentally measured compounds. The linear scale shows the high proportion of Mptrj related to binary an…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Full-cycle device-scale simulations of memory materials with a tailored atomic-cluster-expansion potential

    cond-mat.mtrl-sci 2025-02 conditional novelty 6.0 of 10

    A new ACE machine-learned potential enables full-cycle (RESET-SET-RESET) atomistic simulations of Ge-Sb-Te phase-change memory cells at device scale on CPU-only high-performance computers.

Reference graph

Works this paper leans on

137 extracted references · 40 canonical work pages · cited by 1 Pith paper

  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., Bambrick, J., et al

    Abramson, J., Adler, J., Dunger, J., Evans, R., Green, T., Pritzel, A., Ronneberger, O., Willmore, L., Ballard, A. J., Bambrick, J., et al. Accurate structure prediction of biomolecular interactions with alphafold 3. Nature, pp.\ 1--3, 2024

    2024

  3. [3]

    Alampara, N., Miret, S., and Jablonka, K. M. Mattext: Do language models need more than text & scale for materials modeling? In AI for Accelerated Materials Design-Vienna 2024, 2024

    2024

  4. [4]

    Towards fast, specialized machine learning force fields: Distilling foundation models via energy hessians

    Amin, I., Raja, S., and Krishnapriyan, A. Towards fast, specialized machine learning force fields: Distilling foundation models via energy hessians. arXiv preprint arXiv:2501.09009, 2025

    arXiv 2025

  5. [5]

    J., De Jong, W

    Apra, E., Bylaska, E. J., De Jong, W. A., Govind, N., Kowalski, K., Straatsma, T. P., Valiev, M., van Dam, H. J., Alexeev, Y., Anchell, J., et al. Nwchem: Past, present, and future. The Journal of chemical physics, 152 0 (18), 2020

    2020

  6. [6]

    B., Rodrigues, G

    Araujo, R. B., Rodrigues, G. L. S., dos Santos , E. C., and Pettersson, L. G. M. Adsorption energies on transition metal surfaces: Towards an accurate and balanced description. Nature Communications, 13 0 (1): 0 6853, November 2022. ISSN 2041-1723. doi:10.1038/s41467-022-34507-y

    doi:10.1038/s41467-022-34507-y 2022

  7. [7]

    M., Dzamba, M., Gao, M., Rizvi, A., Zitnick, C

    Barroso-Luque, L., Shuaibi, M., Fu, X., Wood, B. M., Dzamba, M., Gao, M., Rizvi, A., Zitnick, C. L., and Ulissi, Z. W. Open materials 2024 (omat24) inorganic materials dataset and models. arXiv preprint arXiv:2410.12771, 2024

    arXiv 2024

  8. [8]

    M., Kov \'a cs, D

    Batatia, I., Benner, P., Chiang, Y., Elena, A. M., Kov \'a cs, D. P., Riebesell, J., Advincula, X. R., Asta, M., Avaylon, M., Baldwin, W. J., et al. A foundation model for atomistic materials chemistry. arXiv preprint arXiv:2401.00096, 2023

    arXiv 2023

Show all 137 references
  1. [9]

    P., Musaelian, A., Simm, G

    Batatia, I., Batzner, S., Kov \'a cs, D. P., Musaelian, A., Simm, G. N., Drautz, R., Ortner, C., Kozinsky, B., and Cs \'a nyi, G. The design space of e (3)-equivariant atom-centred interatomic potentials. Nature Machine Intelligence, pp.\ 1--12, 2025

    2025

  2. [10]

    P., Kornbluth, M., Molinari, N., Smidt, T

    Batzner, S., Musaelian, A., Sun, L., Geiger, M., Mailoa, J. P., Kornbluth, M., Molinari, N., Smidt, T. E., and Kozinsky, B. E (3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials. Nature communications, 13 0 (1): 0 2453, 2022

    2022

  3. [11]

    Becke, A. D. Density-functional exchange-energy approximation with correct asymptotic behavior. Physical Review A, 38 0 (6): 0 3098--3100, September 1988. doi:10.1103/PhysRevA.38.3098

    1988 doi

  4. [12]

    M., Ranu, S., and Krishnan, N

    Bihani, V., Mannan, S., Pratiush, U., Du, T., Chen, Z., Miret, S., Micoulaut, M., Smedskjaer, M. M., Ranu, S., and Krishnan, N. A. Egraffbench: evaluation of equivariant graph neural network force fields for atomistic simulations. Digital Discovery, 3 0 (4): 0 759--768, 2024 a

    2024

  5. [13]

    Bihani, V., Sastry, S., Ranu, S., and Krishnan, N. M. A. Low-dimensional projections for visualizing energy landscapes of atomic systems. In AI for Accelerated Materials Design - Vienna 2024, 2024 b . URL https://openreview.net/forum?id=x3ryxZgHgu

    2024

  6. [14]

    Graph atomic cluster expansion for semilocal interactions beyond equivariant message passing

    Bochkarev, A., Lysogorskiy, Y., and Drautz, R. Graph atomic cluster expansion for semilocal interactions beyond equivariant message passing. Physical Review X, 14 0 (2): 0 021036, 2024

    2024

  7. [15]

    Bootsma, A. N. and Wheeler, S. Popular Integration Grids Can Result in Large Errors in DFT-Computed Free Energies , July 2019

    2019

  8. [16]

    u gel, S., Br \

    Bosoni, E., Beal, L., Bercx, M., Blaha, P., Bl \"u gel, S., Br \"o der, J., Callsen, M., Cottenier, S., Degomme, A., Dikan, V., Eimre, K., Flage-Larsen , E., Fornari, M., Garcia, A., Genovese, L., Giantomassi, M., Huber, S. P., Janssen, H., Kastlunger, G., Krack, M., Kresse, G...

    2024

  9. [17]

    J., Leary, C., Maclaurin, D., Necula, G., Paszke, A., Vander P las, J., Wanderman- M ilne, S., and Zhang, Q

    Bradbury, J., Frostig, R., Hawkins, P., Johnson, M. J., Leary, C., Maclaurin, D., Necula, G., Paszke, A., Vander P las, J., Wanderman- M ilne, S., and Zhang, Q. JAX : composable transformations of P ython+ N um P y programs, 2018. URL http://github.com/jax-ml/jax

    2018

  10. [18]

    E., and Welling, M

    Brandstetter, J., Worrall, D. E., and Welling, M. Message passing neural PDE solvers. In International Conference on Learning Representations, 2022. URL https://openreview.net/forum?id=vSix3HPYKSU

    2022

  11. [19]

    Does equivariance matter at scale? arXiv preprint arXiv:2410.23179, 2024

    Brehmer, J., Behrends, S., de Haan, P., and Cohen, T. Does equivariance matter at scale? arXiv preprint arXiv:2410.23179, 2024

    2024 arXiv

  12. [20]

    Brunton, S. L. and Kutz, J. N. Promising directions of machine learning for partial differential equations. Nature Computational Science, 4 0 (7): 0 483--494, 2024

    2024

  13. [21]

    B., et al

    Bunne, C., Roohani, Y., Rosen, Y., Gupta, A., Zhang, X., Roed, M., Alexandrov, T., AlQuraishi, M., Brennan, P., Burkhardt, D. B., et al. How to build the virtual cell with artificial intelligence: Priorities and opportunities. Cell, 187 0 (25): 0 7045--7063, 2024

    2024

  14. [22]

    G., Maley, S., Gibaldi, M., Simrod, S., Ogden, V., et al

    Burner, J., Luo, J., White, A., Mirmiran, A., Kwon, O., Boyd, P. G., Maley, S., Gibaldi, M., Simrod, S., Ogden, V., et al. Arc--mof: a diverse database of metal-organic frameworks with dft-derived partial atomic charges and descriptors for machine learning. Chemistry of Materi...

    2023

  15. [23]

    Addressing the Band Gap Problem with a Machine-Learned Exchange Functional

    Bystrom, K., Falletta, S., and Kozinsky, B. Addressing the Band Gap Problem with a Machine-Learned Exchange Functional . Journal of Chemical Theory and Computation, 20 0 (17): 0 7516--7532, September 2024. ISSN 1549-9618, 1549-9626. doi:10.1021/acs.jctc.4c00999

    2024 doi

  16. [24]

    and Parrinello, M

    Car, R. and Parrinello, M. Unified approach for molecular dynamics and density-functional theory. Physical review letters, 55 0 (22): 0 2471, 1985

    1985

  17. [25]

    Open catalyst 2020 (oc20) dataset and community challenges

    Chanussot, L., Das, A., Goyal, S., Lavril, T., Shuaibi, M., Riviere, M., Tran, K., Heras-Domingo, J., Ho, C., Hu, W., et al. Open catalyst 2020 (oc20) dataset and community challenges. Acs Catalysis, 11 0 (10): 0 6059--6072, 2021

    2020

  18. [26]

    Cheetham, A. K. and Seshadri, R. Artificial intelligence driving materials discovery? perspective on the article: Scaling deep learning for materials discovery. Chemistry of Materials, 36 0 (8): 0 3490--3495, 2024

    2024

  19. [27]

    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

    2022

  20. [28]

    H., Lo, A., Miret, S., Pate, B

    Cheng, A. H., Lo, A., Miret, S., Pate, B. H., and Aspuru-Guzik, A. Determining 3d structure from molecular formula and isotopologue rotational spectra in natural abundance with reflection-equivariant diffusion. The Journal of Chemical Physics, 160 0 (12), 2024

    2024

  21. [29]

    F., Reid, A

    Choudhary, K., Garrity, K. F., Reid, A. C., DeCost, B., Biacchi, A. J., Hight Walker, A. R., Trautt, Z., Hattrick-Simpers, J., Kusne, A. G., Centrone, A., et al. The joint automated repository for various integrated simulations (jarvis) for data-driven materials design. npj co...

    2020

  22. [30]

    On the Correlation Problem in Atomic and Molecular Systems

    C \'i z ek, J. On the Correlation Problem in Atomic and Molecular Systems . Calculation of Wavefunction Components in Ursell - Type Expansion Using Quantum - Field Theoretical Methods . The Journal of Chemical Physics, 45 0 (11): 0 4256--4266, December 1966. ISSN 0021-9606. do...

    1966 doi

  23. [31]

    J., Mori-S \'a nchez , P., and Yang, W

    Cohen, A. J., Mori-S \'a nchez , P., and Yang, W. Fractional spins and static correlation error in density functional theory. The Journal of Chemical Physics, 129 0 (12): 0 121104, September 2008. ISSN 0021-9606. doi:10.1063/1.2987202

    2008 doi

  24. [32]

    J., Mori-S \'a nchez , P., and Yang, W

    Cohen, A. J., Mori-S \'a nchez , P., and Yang, W. Challenges for Density Functional Theory . Chemical Reviews, 112 0 (1): 0 289--320, January 2012. ISSN 0009-2665. doi:10.1021/cr200107z

    2012 doi

  25. [33]

    J., and Ceder, G

    Deng, B., Zhong, P., Jun, K., Riebesell, J., Han, K., Bartel, C. J., and Ceder, G. Chgnet as a pretrained universal neural network potential for charge-informed atomistic modelling. Nature Machine Intelligence, 5 0 (9): 0 1031--1041, 2023

    2023

  26. [34]

    Matexpert: Decomposing materials discovery by mimicking human experts

    Ding, Q., Miret, S., and Liu, B. Matexpert: Decomposing materials discovery by mimicking human experts. In AI for Accelerated Materials Design - NeurIPS 2024, 2024. URL https://openreview.net/forum?id=srVcm79DvD

    2024

  27. [35]

    Htmd: high-throughput molecular dynamics for molecular discovery

    Doerr, S., Harvey, M., No \'e , F., and De Fabritiis, G. Htmd: high-throughput molecular dynamics for molecular discovery. Journal of chemical theory and computation, 12 0 (4): 0 1845--1852, 2016

    2016

  28. [36]

    Torchmd: A deep learning framework for molecular simulations

    Doerr, S., Majewski, M., P \'e rez, A., Kramer, A., Clementi, C., Noe, F., Giorgino, T., and De Fabritiis, G. Torchmd: A deep learning framework for molecular simulations. Journal of chemical theory and computation, 17 0 (4): 0 2355--2363, 2021

    2021

  29. [37]

    Downs, R. T. and Hall-Wallace, M. The american mineralogist crystal structure database. American Mineralogist, 88 0 (1): 0 247--250, 2003

    2003

  30. [38]

    V., Joshi, C

    Duval, A., Mathis, S. V., Joshi, C. K., Schmidt, V., Miret, S., Malliaros, F. D., Cohen, T., Lio, P., Bengio, Y., and Bronstein, M. A hitchhiker's guide to geometric gnns for 3d atomic systems. arXiv preprint arXiv:2312.07511, 2023

    2023 arXiv

  31. [39]

    Phast: Physics-aware, scalable, and task-specific gnns for accelerated catalyst design

    Duval, A., Schmidt, V., Miret, S., Bengio, Y., Hern \'a ndez-Garc \'i a, A., and Rolnick, D. Phast: Physics-aware, scalable, and task-specific gnns for accelerated catalyst design. Journal of Machine Learning Research, 25 0 (106): 0 1--26, 2024. URL http://jmlr.org/papers/v25/...

    2024

  32. [40]

    Analyzing atomic interactions in molecules as learned by neural networks

    Esders, M., Schnake, T., Lederer, J., Kabylda, A., Montavon, G., Tkatchenko, A., and M \"u ller, K.-R. Analyzing atomic interactions in molecules as learned by neural networks. Journal of Chemical Theory and Computation, 2024

    2024

  33. [41]

    Analyzing Atomic Interactions in Molecules as Learned by Neural Networks

    Esders, M., Schnake, T., Lederer, J., Kabylda, A., Montavon, G., Tkatchenko, A., and M \"u ller, K.-R. Analyzing Atomic Interactions in Molecules as Learned by Neural Networks . Journal of Chemical Theory and Computation, 21 0 (2): 0 714--729, January 2025. ISSN 1549-9618. doi...

    2025 doi

  34. [42]

    and Lenssen, J

    Fey, M. and Lenssen, J. E. Fast graph representation learning with PyTorch Geometric . In ICLR Workshop on Representation Learning on Graphs and Manifolds, 2019

    2019

  35. [43]

    C., Soklaski, R., Axelrod, S., Samsi, S., Gomez-Bombarelli, R., Coley, C

    Frey, N. C., Soklaski, R., Axelrod, S., Samsi, S., Gomez-Bombarelli, R., Coley, C. W., and Gadepally, V. Neural scaling of deep chemical models. Nature Machine Intelligence, 5 0 (11): 0 1297--1305, 2023

    2023

  36. [44]

    Fu, X., Wu, Z., Wang, W., Xie, T., Keten, S., Gomez-Bombarelli, R., and Jaakkola, T. S. Forces are not enough: Benchmark and critical evaluation for machine learning force fields with molecular simulations. Transactions on Machine Learning Research, 2023

    2023

  37. [45]

    A., Tadmor, E

    Fuemmeler, E., Wolfe, G., Gupta, A., Vita, J. A., Tadmor, E. B., and Martiniani, S. Advancing the colabfit exchange towards a web-scale data source for machine learning interatomic potentials. In AI for Accelerated Materials Design-NeurIPS 2024, 2024

    2024

  38. [46]

    S., Schoenholz, S

    Gangan, A. S., Schoenholz, S. S., Cubuk, E. D., Bauchy, M., and Krishnan, N. Force field optimization by end-to-end differentiable atomistic simulation. arXiv preprint arXiv:2409.13844, 2024

    2024 arXiv

  39. [47]

    Searching for high-value molecules using reinforcement learning and transformers

    Ghugare, R., Miret, S., Hugessen, A., Phielipp, M., and Berseth, G. Searching for high-value molecules using reinforcement learning and transformers. In The Twelfth International Conference on Learning Representations, 2024. URL https://openreview.net/forum?id=nqlymMx42E

    2024

  40. [48]

    L., Cococcioni, M., Dabo, I., et al

    Giannozzi, P., Baroni, S., Bonini, N., Calandra, M., Car, R., Cavazzoni, C., Ceresoli, D., Chiarotti, G. L., Cococcioni, M., Dabo, I., et al. Quantum espresso: a modular and open-source software project for quantum simulations of materials. Journal of physics: Condensed matter...

    2009

  41. [49]

    The non-linear nature of the cost of comprehensibility

    Goethals, S., Martens, D., and Evgeniou, T. The non-linear nature of the cost of comprehensibility. Journal of Big Data, 9 0 (1): 0 30, March 2022. ISSN 2196-1115. doi:10.1186/s40537-022-00579-2

    2022 doi

  42. [50]

    Gonzales, C., Lee, K. L. K., Mu, B., Galkin, M., and Miret, S. Data efficient training for materials property prediction using active learning querying. In AI for Accelerated Materials Design - NeurIPS 2023 Workshop, 2023. URL https://openreview.net/forum?id=waQPqifiuC

    2023

  43. [51]

    B., Martiniani, S., and Miret, S

    Gonzales, C., Fuemmeler, E., Tadmor, E. B., Martiniani, S., and Miret, S. Benchmarking of universal machine learning interatomic potentials for structural relaxation. In AI for Accelerated Materials Design - NeurIPS 2024, 2024. URL https://openreview.net/forum?id=fNyXCCZ0g6

    2024

  44. [52]

    Crystal design amidst noisy dft signals: A reinforcement learning approach

    Govindarajan, P., Reymond, M., Miret, S., Phielipp, M., and Chandar, S. Crystal design amidst noisy dft signals: A reinforcement learning approach. In AI for Accelerated Materials Design-NeurIPS 2024, 2024

    2024

  45. [53]

    G., Zitnick, C

    Gruver, N., Sriram, A., Madotto, A., Wilson, A. G., Zitnick, C. L., and Ulissi, Z. W. Fine-tuned language models generate stable inorganic materials as text. In The Twelfth International Conference on Learning Representations, 2024. URL https://openreview.net/forum?id=vN9fpfqoP1

    2024

  46. [54]

    B., and Martiniani, S

    Gupta, A., Tadmor, E. B., and Martiniani, S. Kusp: Python server for deploying ml interatomic potentials. In AI for Accelerated Materials Design-Vienna 2024, 2024

    2024

  47. [55]

    and Tibshirani, R

    Hastie, T. and Tibshirani, R. Generalized Additive Models . Statistical Science, 1 0 (3): 0 297--310, August 1986. ISSN 0883-4237, 2168-8745. doi:10.1214/ss/1177013604

    1986

  48. [56]

    and Fitzgibbon, A

    Helal, H. and Fitzgibbon, A. Mess: Modern electronic structure simulations. arXiv preprint arXiv:2406.03121, 2024

    2024 arXiv

  49. [57]

    Coupled cluster finite temperature simulations of periodic materials via machine learning

    Herzog, B., Gallo, A., Hummel, F., Badawi, M., Bu c ko, T., Leb \`e gue, S., Gr \"u neis, A., and Rocca, D. Coupled cluster finite temperature simulations of periodic materials via machine learning. npj Computational Materials, 10 0 (1): 0 1--7, April 2024. ISSN 2057-3960. doi...

    2024 doi

  50. [58]

    E., Christensen, R., Dułak, M., Friis, J., Groves, M

    Hjorth Larsen, A., Jørgen Mortensen, J., Blomqvist, J., Castelli, I. E., Christensen, R., Dułak, M., Friis, J., Groves, M. N., Hammer, B., Hargus, C., Hermes, E. D., Jennings, P. C., Bjerre Jensen, P., Kermode, J., Kitchin, J. R., Leonhard Kolsbjerg, E., Kubal, J., Kaasbjerg, ...

    2017

  51. [59]

    K., Montoya, J

    Horton, M. K., Montoya, J. H., Liu, M., and Persson, K. A. High-throughput prediction of the ground-state collinear magnetic order of inorganic materials using density functional theory. npj Computational Materials, 5 0 (1): 0 64, 2019

    2019

  52. [60]

    Difftaichi: Differentiable programming for physical simulation

    Hu, Y., Anderson, L., Li, T.-M., Sun, Q., Carr, N., Ragan-Kelley, J., and Durand, F. Difftaichi: Differentiable programming for physical simulation. In International Conference on Learning Representations, 2020. URL https://openreview.net/forum?id=B1eB5xSFvr

    2020

  53. [61]

    M., Yang, L., Linker, T., Olguin, M., Hattori, S., Luo, Y., Kalia, R

    Ibayashi, H., Razakh, T. M., Yang, L., Linker, T., Olguin, M., Hattori, S., Luo, Y., Kalia, R. K., Nakano, A., Nomura, K.-i., et al. Allegro-legato: Scalable, fast, and robust neural-network quantum molecular dynamics via sharpness-aware minimization. In International Conferen...

    2023

  54. [62]

    J., Ong, S

    Jain, A., Hautier, G., Moore, C. J., Ong, S. P., Fischer, C. C., Mueller, T., Persson, K. A., and Ceder, G. A high-throughput infrastructure for density functional theory calculations. Computational Materials Science, 50 0 (8): 0 2295--2310, 2011

    2011

  55. [63]

    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

  56. [64]

    Space group constrained crystal generation

    Jiao, R., Huang, W., Liu, Y., Zhao, D., and Liu, Y. Space group constrained crystal generation. In The Twelfth International Conference on Learning Representations, 2024. URL https://openreview.net/forum?id=jkvZ7v4OmP

    2024

  57. [65]

    Jones, R. O. Density functional theory: Its origins, rise to prominence, and future. Reviews of Modern Physics, 87 0 (3): 0 897--923, August 2015. doi:10.1103/RevModPhys.87.897

    2015 doi

  58. [66]

    Highly accurate protein structure prediction with alphafold

    Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., Tunyasuvunakool, K., Bates, R., Z \' dek, A., Potapenko, A., et al. Highly accurate protein structure prediction with alphafold. nature, 596 0 (7873): 0 583--589, 2021

    2021

  59. [67]

    Timewarp: Transferable acceleration of molecular dynamics by learning time-coarsened dynamics

    Klein, L., Foong, A., Fjelde, T., Mlodozeniec, B., Brockschmidt, M., Nowozin, S., No \'e , F., and Tomioka, R. Timewarp: Transferable acceleration of molecular dynamics by learning time-coarsened dynamics. Advances in Neural Information Processing Systems, 36, 2024

    2024

  60. [68]

    and Sham, L

    Kohn, W. and Sham, L. J. Self-consistent equations including exchange and correlation effects. Physical review, 140 0 (4A): 0 A1133, 1965

    1965

  61. [69]

    Neural operator: Learning maps between function spaces with applications to pdes

    Kovachki, N., Li, Z., Liu, B., Azizzadenesheli, K., Bhattacharya, K., Stuart, A., and Anandkumar, A. Neural operator: Learning maps between function spaces with applications to pdes. Journal of Machine Learning Research, 24 0 (89): 0 1--97, 2023

    2023

  62. [70]

    Scaling the leading accuracy of deep equivariant models to biomolecular simulations of realistic size

    Kozinsky, B., Musaelian, A., Johansson, A., and Batzner, S. Scaling the leading accuracy of deep equivariant models to biomolecular simulations of realistic size. In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, S...

    2023

  63. [71]

    and Hafner, J

    Kresse, G. and Hafner, J. Ab initio molecular-dynamics simulation of the liquid-metal--amorphous-semiconductor transition in germanium. Physical Review B, 49 0 (20): 0 14251, 1994

    1994

  64. [72]

    u hne, T. D., Iannuzzi, M., Del Ben, M., Rybkin, V. V., Seewald, P., Stein, F., Laino, T., Khaliullin, R. Z., Sch \

    K \"u hne, T. D., Iannuzzi, M., Del Ben, M., Rybkin, V. V., Seewald, P., Stein, F., Laino, T., Khaliullin, R. Z., Sch \"u tt, O., Schiffmann, F., et al. Cp2k: An electronic structure and molecular dynamics software package-quickstep: Efficient and accurate electronic structure...

    2020

  65. [73]

    Lee, C., Yang, W., and Parr, R. G. Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density. Physical Review B, 37 0 (2): 0 785--789, January 1988. doi:10.1103/PhysRevB.37.785

    1988 doi

  66. [74]

    Lee, K. L. K., Gonzales, C., Nassar, M., Spellings, M., Galkin, M., and Miret, S. Matsciml: A broad, multi-task benchmark for solid-state materials modeling. arXiv preprint arXiv:2309.05934, 2023 a

    2023 arXiv

  67. [75]

    Lee, K. L. K., Gonzales, C., Spellings, M., Galkin, M., Miret, S., and Kumar, N. Towards foundation models for materials science: The open matsci ml toolkit. In Proceedings of the SC'23 Workshops of The International Conference on High Performance Computing, Network, Storage, ...

    2023

  68. [76]

    Lee, K. L. K., Galkin, M., and Miret, S. Scaling computational performance of spherical harmonics kernels with triton. In AI for Accelerated Materials Design-Vienna 2024, 2024 a

    2024

  69. [77]

    Lee, K. L. K., Galkin, M., and Miret, S. Deconstructing equivariant representations in molecular systems. In AI for Accelerated Materials Design - NeurIPS 2024 , November 2024 b

    2024

  70. [78]

    o rkman, T., Blaha, P., Bl \

    Lejaeghere, K., Bihlmayer, G., Bj \"o rkman, T., Blaha, P., Bl \"u gel, S., Blum, V., Caliste, D., Castelli, I. E., Clark, S. J., Dal Corso, A., et al. Reproducibility in density functional theory calculations of solids. Science, 351 0 (6280): 0 aad3000, 2016

    2016

  71. [79]

    S., Kaba, S.-O., Zhu, Q., Galkin, M., Miret, S., and Ravanbakhsh, S

    Levy, D., Panigrahi, S. S., Kaba, S.-O., Zhu, Q., Galkin, M., Miret, S., and Ravanbakhsh, S. Symm CD : Symmetry-preserving crystal generation with diffusion models. In AI for Accelerated Materials Design - NeurIPS 2024, 2024. URL https://openreview.net/forum?id=V7x2KZQn2v

    2024

  72. [80]

    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. URL https://openreview.net/forum?id=KwmPfARgOTD

    2023

  73. [81]

    Generalizing denoising to non-equilibrium structures improves equivariant force fields

    Liao, Y.-L., Smidt, T., Shuaibi*, M., and Das*, A. Generalizing denoising to non-equilibrium structures improves equivariant force fields. arXiv preprint arXiv:2403.09549, 2024

    2024 arXiv

  74. [82]

    Intelligible models for classification and regression

    Lou, Y., Caruana, R., and Gehrke, J. Intelligible models for classification and regression. In Proceedings of the 18th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , pp.\ 150--158, Beijing China, August 2012. ACM. ISBN 978-1-4503-1462-6. doi:10.11...

    2012

  75. [83]

    Conceptual Problem with Calculating Electron Densities in Finite Basis Density Functional Theory

    Mayer, I., P \'a pai, I., Bak \'o , I., and Nagy, \'A . Conceptual Problem with Calculating Electron Densities in Finite Basis Density Functional Theory . Journal of Chemical Theory and Computation, 13 0 (9): 0 3961--3963, September 2017. ISSN 1549-9618. doi:10.1021/acs.jctc.7b00562

    2017 doi

  76. [84]

    G., Bushmarinov, I

    Medvedev, M. G., Bushmarinov, I. S., Sun, J., Perdew, J. P., and Lyssenko, K. A. Density functional theory is straying from the path toward the exact functional. Science, 355 0 (6320): 0 49--52, January 2017. doi:10.1126/science.aah5975

    2017 doi

  77. [85]

    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

  78. [86]

    D., Schoenholz, S

    Metz, L., Freeman, C. D., Schoenholz, S. S., and Kachman, T. Gradients are not all you need. arXiv preprint arXiv:2111.05803, 2021

    2021 arXiv

  79. [87]

    K., Chen, R

    Miller, B. K., Chen, R. T. Q., Sriram, A., and Wood, B. M. Flow MM : Generating materials with riemannian flow matching. In Forty-first International Conference on Machine Learning, 2024. URL https://openreview.net/forum?id=W4pB7VbzZI

    2024

  80. [88]

    and Krishnan, N

    Miret, S. and Krishnan, N. Are llms ready for real-world materials discovery? arXiv preprint arXiv:2402.05200, 2024

    2024 arXiv

  81. [89]

    Miret, S., Lee, K. L. K., Gonzales, C., Nassar, M., and Spellings, M. The open matsci ML toolkit: A flexible framework for machine learning in materials science. Transactions on Machine Learning Research, 2023. ISSN 2835-8856. URL https://openreview.net/forum?id=QBMyDZsPMd

    2023

  82. [90]

    A., Sanchez-Lengeling, B., Skreta, M., Venugopal, V., and Wei, J

    Miret, S., Krishnan, N. A., Sanchez-Lengeling, B., Skreta, M., Venugopal, V., and Wei, J. N. Perspective on ai for accelerated materials design at the ai4mat-2023 workshop at neurips 2023. Digital Discovery, 2024

    2023

  83. [91]

    R., van Dijk , D., Wang, Z., Gigante, S., Burkhardt, D

    Moon, K. R., van Dijk , D., Wang, Z., Gigante, S., Burkhardt, D. B., Chen, W. S., Yim, K., van den Elzen, A., Hirn, M. J., Coifman, R. R., Ivanova, N. B., Wolf, G., and Krishnaswamy, S. Visualizing structure and transitions in high-dimensional biological data. Nature Biotechno...

    2019 doi

  84. [92]

    J., Kornbluth, M., and Kozinsky, B

    Musaelian, A., Batzner, S., Johansson, A., Sun, L., Owen, C. J., Kornbluth, M., and Kozinsky, B. Learning local equivariant representations for large-scale atomistic dynamics. Nature Communications, 14 0 (1): 0 579, 2023

    2023

  85. [93]

    Orb: A fast, scalable neural network potential

    Neumann, M., Gin, J., Rhodes, B., Bennett, S., Li, Z., Choubisa, H., Hussey, A., and Godwin, J. Orb: A fast, scalable neural network potential. arXiv preprint arXiv:2410.22570, 2024

    2024 arXiv

  86. [94]

    M., Kuo, T.-S., Liu, Y., Dror, R., Brajovic, D., Yao, X., Bartolo, M., Rojas, W

    Oala, L., Maskey, M., Bat-Leah, L., Parrish, A., G \"u rel, N. M., Kuo, T.-S., Liu, Y., Dror, R., Brajovic, D., Yao, X., Bartolo, M., Rojas, W. A. G., Hileman, R., Aliment, R., Mahoney, M. W., Risdal, M., Lease, M., Samek, W., Dutta, D., Northcutt, C. G., Coleman, C., Hancock,...

    2024

  87. [95]

    P., Richards, W

    Ong, S. P., Richards, W. D., Jain, A., Hautier, G., Kocher, M., Cholia, S., Gunter, D., Chevrier, V. L., Persson, K. A., and Ceder, G. Python materials genomics (pymatgen): A robust, open-source python library for materials analysis. Computational Materials Science, 68: 0 314-...

    2013

  88. [96]

    Perdew, J. P. Density functional theory and the band gap problem. International Journal of Quantum Chemistry, 28 0 (S19): 0 497--523, 1985. ISSN 1097-461X. doi:10.1002/qua.560280846

    1985 doi

  89. [97]

    P., Burke, K., and Ernzerhof, M

    Perdew, J. P., Burke, K., and Ernzerhof, M. Generalized gradient approximation made simple. Physical review letters, 77 0 (18): 0 3865, 1996

    1996

  90. [98]

    P., Ruzsinszky, A., Tao, J., Staroverov, V

    Perdew, J. P., Ruzsinszky, A., Tao, J., Staroverov, V. N., Scuseria, G. E., and Csonka, G. I. Prescription for the design and selection of density functional approximations: More constraint satisfaction with fewer fits. The Journal of chemical physics, 123 0 (6), 2005

    2005

  91. [99]

    Matthews , A., and Foulkes, W

    Pfau, D., Spencer, J., de G. Matthews , A., and Foulkes, W. Ab-initio solution of the many-electron schr \"o dinger equation with deep neural networks. Physical Review Research, 2 0 (3): 0 033429, 2020. doi:10.1103/PhysRevResearch.2.033429

    2020 doi

  92. [100]

    Pfau, D., Axelrod, S., Sutterud, H., von Glehn , I., and Spencer, J. S. Accurate computation of quantum excited states with neural networks. Science, 385 0 (6711): 0 eadn0137, August 2024. doi:10.1126/science.adn0137

    2024 doi

  93. [101]

    Purvis, III, G. D. and Bartlett, R. J. A full coupled-cluster singles and doubles model: The inclusion of disconnected triples. The Journal of Chemical Physics, 76 0 (4): 0 1910--1918, February 1982. ISSN 0021-9606. doi:10.1063/1.443164

    1910 doi

  94. [102]

    and Krishnapriyan, A

    Qu, E. and Krishnapriyan, A. S. The importance of being scalable: Improving the speed and accuracy of neural network interatomic potentials across chemical domains. arXiv preprint arXiv:2410.24169, 2024

    2024 arXiv

  95. [103]

    Raja, S., Amin, I., Pedregosa, F., and Krishnapriyan, A. S. Stability-aware training of neural network interatomic potentials with differentiable boltzmann estimators. arXiv preprint arXiv:2402.13984, 2024

    2024 arXiv

  96. [104]

    Reddy, M. N. and Panda, D. K. A comprehensive review on finfet in terms of its device structure and performance matrices. Silicon, 14 0 (18): 0 12015--12030, 2022

    2022

  97. [105]

    E., Jain, A., Benner, P., Persson, K

    Riebesell, J., Goodall, R. E., Jain, A., Benner, P., Persson, K. A., and Lee, A. A. Matbench discovery--an evaluation framework for machine learning crystal stability prediction. arXiv preprint arXiv:2308.14920, 2023

    2023 arXiv

  98. [106]

    O., Taylor, P

    Roos, B. O., Taylor, P. R., and Sigbahn, P. E. A complete active space scf method (casscf) using a density matrix formulated super-ci approach. Chemical Physics, 48 0 (2): 0 157--173, 1980

    1980

  99. [107]

    E., Kirklin, S., Aykol, M., Meredig, B., and Wolverton, C

    Saal, J. E., Kirklin, S., Aykol, M., Meredig, B., and Wolverton, C. Materials design and discovery with high-throughput density functional theory: the open quantum materials database (oqmd). Jom, 65: 0 1501--1509, 2013

    2013

  100. [108]

    F., Romero, A

    Schmidt, J., Cerqueira, T. F., Romero, A. H., Loew, A., J \"a ger, F., Wang, H.-C., Botti, S., and Marques, M. A. Improving machine-learning models in materials science through large datasets. Materials Today Physics, 48: 0 101560, 2024

    2024

  101. [109]

    and Cubuk, E

    Schoenholz, S. and Cubuk, E. D. Jax md: a framework for differentiable physics. Advances in Neural Information Processing Systems, 33: 0 11428--11441, 2020

    2020

  102. [110]

    and Verstraete, F

    Schuch, N. and Verstraete, F. Computational complexity of interacting electrons and fundamental limitations of density functional theory. Nature physics, 5 0 (10): 0 732--735, 2009

    2009

  103. [111]

    E., Adams, P

    Shaw, D. E., Adams, P. J., Azaria, A., Bank, J. A., Batson, B., Bell, A., Bergdorf, M., Bhatt, J., Butts, J. A., Correia, T., Dirks, R. M., Dror, R. O., Eastwood, M. P., Edwards, B., Even, A., Feldmann, P., Fenn, M., Fenton, C. H., Forte, A., Gagliardo, J., Gill, G., Gorlatova...

    2021

  104. [112]

    G., Strieth-Kalthoff, F., Hao, H., Hickman, R

    Sim, M., Vakili, M. G., Strieth-Kalthoff, F., Hao, H., Hickman, R. J., Miret, S., Pablo-Garc \' a, S., and Aspuru-Guzik, A. Chemos 2.0: An orchestration architecture for chemical self-driving laboratories. Matter, 7 0 (9): 0 2959--2977, 2024

    2024

  105. [113]

    Q., Li, C., and Yang, W

    Su, N. Q., Li, C., and Yang, W. Describing strong correlation with fractional-spin correction in density functional theory. Proceedings of the National Academy of Sciences, 115 0 (39): 0 9678--9683, September 2018. doi:10.1073/pnas.1807095115

    2018 doi

  106. [114]

    Pdebench: An extensive benchmark for scientific machine learning

    Takamoto, M., Praditia, T., Leiteritz, R., MacKinlay, D., Alesiani, F., Pfl \"u ger, D., and Niepert, M. Pdebench: An extensive benchmark for scientific machine learning. Advances in Neural Information Processing Systems, 35: 0 1596--1611, 2022

    2022

  107. [115]

    R., Wang, Y., Liu, F., Xu, H., and Li, J

    Tang, H., Xiao, B., He, W., Subasic, P., Harutyunyan, A. R., Wang, Y., Liu, F., Xu, H., and Li, J. Approaching coupled-cluster accuracy for molecular electronic structures with multi-task learning. Nature Computational Science, 2024

    2024

  108. [116]

    C., Liebschner, D., Croll, T

    Terwilliger, T. C., Liebschner, D., Croll, T. I., Williams, C. J., McCoy, A. J., Poon, B. K., Afonine, P. V., Oeffner, R. D., Richardson, J. S., Read, R. J., et al. Alphafold predictions are valuable hypotheses and accelerate but do not replace experimental structure determina...

    2024

  109. [117]

    P., Aktulga, H

    Thompson, A. P., Aktulga, H. M., Berger, R., Bolintineanu, D. S., Brown, W. M., Crozier, P. S., In't Veld, P. J., Kohlmeyer, A., Moore, S. G., Nguyen, T. D., et al. Lammps-a flexible simulation tool for particle-based materials modeling at the atomic, meso, and continuum scale...

    2022

  110. [118]

    Triton: an intermediate language and compiler for tiled neural network computations

    Tillet, P., Kung, H.-T., and Cox, D. Triton: an intermediate language and compiler for tiled neural network computations. In Proceedings of the 3rd ACM SIGPLAN International Workshop on Machine Learning and Programming Languages, pp.\ 10--19, 2019

    2019

  111. [119]

    M., Goyal, S., Das, A., Heras-Domingo, J., Kolluru, A., Rizvi, A., Shoghi, N., et al

    Tran, R., Lan, J., Shuaibi, M., Wood, B. M., Goyal, S., Das, A., Heras-Domingo, J., Kolluru, A., Rizvi, A., Shoghi, N., et al. The open catalyst 2022 (oc22) dataset and challenges for oxide electrocatalysts. ACS Catalysis, 13 0 (5): 0 3066--3084, 2023

    2022

  112. [120]

    R., Lebrun-Grandié, D., Arndt, D., Ciesko, J., Dang, V., Ellingwood, N., Gayatri, R., Harvey, E., Hollman, D

    Trott, C. R., Lebrun-Grandié, D., Arndt, D., Ciesko, J., Dang, V., Ellingwood, N., Gayatri, R., Harvey, E., Hollman, D. S., Ibanez, D., Liber, N., Madsen, J., Miles, J., Poliakoff, D., Powell, A., Rajamanickam, S., Simberg, M., Sunderland, D., Turcksin, B., and Wilke, J. Kokko...

    2022

  113. [121]

    J., and Van Santen, R

    Van Beest, B., Kramer, G. J., and Van Santen, R. Force fields for silicas and aluminophosphates based on ab initio calculations. Physical review letters, 64 0 (16): 0 1955, 1990

    1955

  114. [122]

    A., Fuemmeler, E

    Vita, J. A., Fuemmeler, E. G., Gupta, A., Wolfe, G. P., Tao, A. Q., Elliott, R. S., Martiniani, S., and Tadmor, E. B. Colabfit exchange: Open-access datasets for data-driven interatomic potentials. The Journal of Chemical Physics, 159 0 (15), 2023

    2023

  115. [123]

    S., and Pfau, D

    von Glehn, I., Spencer, J. S., and Pfau, D. A Self-Attention Ansatz for Ab-initio Quantum Chemistry , April 2023

    2023

  116. [124]

    Deep graph library: A graph-centric, highly-performant package for graph neural networks

    Wang, M., Zheng, D., Ye, Z., Gan, Q., Li, M., Song, X., Zhou, J., Ma, C., Yu, L., Gai, Y., Xiao, T., He, T., Karypis, G., Li, J., and Zhang, Z. Deep graph library: A graph-centric, highly-performant package for graph neural networks. arXiv preprint arXiv:1909.01315, 2019

    1909 arXiv

  117. [125]

    Perovs-dopants: Machine learning potentials for doped bulk structures

    Wang, X., Park, S., and Miret, S. Perovs-dopants: Machine learning potentials for doped bulk structures. In AI for Accelerated Materials Design-NeurIPS 2024, 2024 a

    2024

  118. [126]

    X2- GNN : A Physical Message Passing Neural Network with Natural Generalization Ability to Large and Complex Molecules

    Wang, Z., Zhang, W., Jiang, M., Chen, Y., Zhu, Z., Yan, W., Wu, J., and Xu, X. X2- GNN : A Physical Message Passing Neural Network with Natural Generalization Ability to Large and Complex Molecules . The Journal of Physical Chemistry Letters, 15 0 (51): 0 12501--12512, Decembe...

    2024 doi

  119. [127]

    and Marquetand, P

    Westermayr, J. and Marquetand, P. Machine learning for electronically excited states of molecules. Chemical Reviews, 121 0 (16): 0 9873--9926, 2020

    2020

  120. [128]

    Wood, S. N. Generalized Additive Models . Annual Reviews of Statistics and its Application, October 2024. doi:10.1146/annurev-statistics-112723-034249

    2024 doi

  121. [129]

    Protein data bank: the single global archive for 3d macromolecular structure data

    wwPDB consortium. Protein data bank: the single global archive for 3d macromolecular structure data. Nucleic acids research, 47 0 (D1): 0 D520--D528, 2019

    2019

  122. [130]

    Small data machine learning in materials science

    Xu, P., Ji, X., Li, M., and Lu, W. Small data machine learning in materials science. npj Computational Materials, 9 0 (1): 0 42, 2023

    2023

  123. [131]

    Mattersim: A deep learning atomistic model across elements, temperatures and pressures

    Yang, H., Hu, C., Zhou, Y., Liu, X., Shi, Y., Li, J., Li, G., Chen, Z., Chen, S., Zeni, C., et al. Mattersim: A deep learning atomistic model across elements, temperatures and pressures. arXiv preprint arXiv:2405.04967, 2024

    2024 arXiv

  124. [132]

    A., Smedskjaer, M

    Yang, K., Xu, X., Yang, B., Cook, B., Ramos, H., Krishnan, N. A., Smedskjaer, M. M., Hoover, C., and Bauchy, M. Predicting the young’s modulus of silicate glasses using high-throughput molecular dynamics simulations and machine learning. Scientific reports, 9 0 (1): 0 8739, 2019

    2019

  125. [133]

    W., Sanchez-Gonzalez, A., Battaglia, P., Pascanu, R., and Godwin, J

    Zaidi, S., Schaarschmidt, M., Martens, J., Kim, H., Teh, Y. W., Sanchez-Gonzalez, A., Battaglia, P., Pascanu, R., and Godwin, J. Pre-training via denoising for molecular property prediction. In The Eleventh International Conference on Learning Representations, 2023. URL https:...

    2023

  126. [134]

    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

  127. [135]

    R., Chenthamarakshan, V., Lozano, A., Das, P., and Tang, J

    Zhang, Z., Xu, M., Jamasb, A. R., Chenthamarakshan, V., Lozano, A., Das, P., and Tang, J. Protein representation learning by geometric structure pretraining. In The Eleventh International Conference on Learning Representations, 2023. URL https://openreview.net/forum?id=to3qCB3tOh9

    2023

  128. [136]

    Predicting equilibrium distributions for molecular systems with deep learning

    Zheng, S., He, J., Liu, C., Shi, Y., Lu, Z., Feng, W., Ju, F., Wang, J., Zhu, J., Min, Y., et al. Predicting equilibrium distributions for molecular systems with deep learning. Nature Machine Intelligence, pp.\ 1--10, 2024

    2024

  129. [137]

    L., Chanussot, L., Das, A., Goyal, S., Heras-Domingo, J., Ho, C., Hu, W., Lavril, T., Palizhati, A., Riviere, M., et al

    Zitnick, C. L., Chanussot, L., Das, A., Goyal, S., Heras-Domingo, J., Ho, C., Hu, W., Lavril, T., Palizhati, A., Riviere, M., et al. An introduction to electrocatalyst design using machine learning for renewable energy storage. arXiv preprint arXiv:2010.09435, 2020

    2010 arXiv

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.