Pith. sign in

REVIEW 4 major objections 6 minor 86 references

Leveraging active learning-enhanced machine-learned interatomic potential for efficient infrared spectra prediction

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read An active-learning workflow trains a neural-network potential and a dipole model on a few hundred DFT calculations per molecule and reproduces both ab initio and experimental infrared spectra.

desk verdict A well-engineered, open-source workflow for MLIP-based IR spectra with credible 100x DFT-cost savings; the main gap is untested dipole-model coverage on production trajectories. read the letter →

arxiv 2506.13486 v1 pith:MBOOX5YT submitted 2025-06-16 physics.chem-ph physics.comp-ph

classification physics.chem-phphysics.comp-ph
keywords infraredspectroscopyactivelearningmachine-learnedinteratomicpotentialsdatagenerationequivariantneuralnetworksmoleculardynamicsdipolemomentsspectraprediction
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 sets out to show that an active-learning workflow can train a machine-learned interatomic potential and a companion dipole-moment model from just a few hundred density functional theory (DFT) calculations per molecule. For 24 small organic molecules, roughly 600-800 selected structures per molecule (about 16,000 in total) were enough to push harmonic-frequency errors to roughly 4 cm$^{-1}$ and force errors to about 4 meV/Å. The spectra predicted from machine-learning molecular dynamics match DFT-based ab initio spectra and experimental gas-phase spectra in both peak positions and amplitudes, while cutting the number of DFT calculations by roughly a factor of 100. If correct, this makes high-throughput infrared screening of catalytically relevant molecules practical, including temperature-dependent spectra from 100 to 900 K.

What carries the argument

The load-bearing mechanism is the active-learning loop built around an ensemble of equivariant message-passing neural-network potentials. Disagreement among ensemble members' force predictions marks each configuration's uncertainty; the most uncertain structures from short simulations at three temperatures are labelled by DFT and added to the training set, and the ensemble is retrained. A second network with a vector output learns the molecular dipole moment. The infrared spectrum is the Fourier transform of the autocorrelation of the dipole time derivative along 50 ps trajectories, with a 1000 fs correlation depth, a smoothing window, and averaging over three independent trajectories; similarity to reference spectra is quantified by a correlation coefficient and a transport distance.

What would settle it

For a molecule outside the training distribution, such as one with the C=C motif beyond ethene, run the production trajectories, compute DFT dipole moments at snapshots where the force ensemble disagreed most, and compare them with the dipole model; if the dipole errors are large and the spectrum changes when training data are instead selected by dipole disagreement, the force-uncertainty proxy is refuted.

Watch

Extended reading notes

Core claim

The central claim is that data selection by force uncertainty, not exhaustive sampling, is what makes machine-learned spectra affordable and accurate. The workflow starts from geometries sampled along harmonic normal modes, then repeatedly runs short dynamics with an ensemble of neural-network potentials and adds the configurations where ensemble force predictions disagree most, cycling at 300, 500, and 700 K until harmonic-frequency errors plateau. A second neural network is trained on the accumulated set to predict dipole moments, and the spectrum is computed from the autocorrelation of the dipole time derivative over three independent 50 ps trajectories. Across all 24 molecules the machine-learned spectra agree with the DFT reference spectra (mean correlation 0.80) and with experimental spectra (mean correlation 0.81), comparable to or better than the DFT-versus-experiment agreement of 0.68; the authors conclude that anharmonic effects and temperature dependence are captured implicitly by the dynamics.

Load-bearing premise

The workflow assumes that the configurations picked because the atomic-force predictions disagreed are also enough to train the separate dipole-moment model, so that model stays accurate along the trajectories that actually produce the spectrum.

Editorial extensions

If this is right

  • IR spectra of small catalytically relevant molecules can be produced at DFT-level accuracy with roughly 100 times fewer single-point DFT calculations, making high-throughput spectral screening feasible.
  • Temperature-dependent spectra become practical by running the trained models at different temperatures, demonstrated here from 100 to 900 K, without new ab initio simulations.
  • Because the machine-learned dynamics scales roughly linearly with system size while DFT scales steeply, the same workflow should extend to larger molecules than ab initio dynamics can reach.
  • Transferability is bounded by chemical coverage: molecules with underrepresented bonding motifs, such as the C=C motif in 1,3-butadiene, show degraded spectra, so broader training sets or transfer learning are the natural route to wider applicability.

Reading between the lines

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

  • If the force-disagreement selection rule is really a sufficient proxy for dipole-moment coverage, then a variant that actively selects configurations by dipole-model disagreement should not change the spectra; running that comparison for out-of-distribution molecules would test the workflow's weakest link.
  • The same acquisition logic could be transferred to other spectra that depend on a property surface—Raman intensities from polarizability, or vibrational circular dichroism from rotatory strength—with the property-specific uncertainty as the acquisition signal.
  • The better-than-DFT agreement with experiment may come substantially from averaging three independent trajectories rather than from the potential itself; separating sampling noise from model error by varying the number of trajectories would clarify where the accuracy gain originates.
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

4 major / 6 minor

Summary. The manuscript introduces PALIRS, an active-learning workflow that trains a MACE ensemble-based machine-learned interatomic potential (MLIP) and a separate MACE dipole moment model, then uses machine-learning molecular dynamics (MLMD) trajectories and the dipole autocorrelation function (Eq. 1) to predict infrared spectra. The method is tested on 24 small organic molecules by comparing MLMD spectra against DFT-based AIMD spectra and NIST experimental spectra using Pearson correlation coefficient (PCC) and Wasserstein distance (WD). The authors report mean DFT-ML PCC 0.80 and Exp.-ML PCC 0.81, a speedup relative to AIMD, temperature-dependence studies for methanol and ethanol, and transferability tests on 8 additional molecules. The central claim is that PALIRS reproduces AIMD-quality IR spectra with fewer than 1,000 single-point DFT calculations per molecule, compared with roughly 100,000 DFT steps for AIMD.

Significance. If the central claim holds, PALIRS is a practically useful contribution to high-throughput IR spectroscopy for small organic molecules, and the open-source code and deposited datasets (Zenodo DOIs, GitLab repository) are valuable for reproducibility. The external comparison to NIST experimental spectra is a genuine benchmark, and the systematic analysis of trajectory length for spectral convergence is a useful practical guideline. The main limitation is that the dipole moment model is trained and evaluated on configurations selected by force-uncertainty active learning, so the accuracy of the dipole model on production trajectories and on out-of-distribution molecules is not directly demonstrated. This gap tempers the strength of the transferability claims but does not invalidate the core comparison for in-distribution molecules, which is supported by the DFT-ML spectral agreement.

major comments (4)
  1. [Section 5.6 and Figure 3] The test set used for the MLIP and dipole moment errors is generated from a 100 ps MLMD trajectory produced by the first MLIP of the final ensemble, with structures clustered in MBTR space and labeled by DFT. This makes the test configurations in-distribution for the force model and likely also for the dipole model, since the training set was itself collected from active-learning MLMD runs at 300, 500, and 700 K. The dipole moment MAE of 7.62 mDebye in Table 1 therefore does not measure accuracy on out-of-distribution regions or on the full production ensemble. The central claim would be considerably strengthened by evaluating the dipole model on an independently generated test set (for example, from DFT-based AIMD trajectories or from MLMD trajectories of the final ensemble at multiple temperatures) and by reporting dipole errors on the actual production trajectories used for IR spectra.
  2. [Sections 5.4 and 3] Active learning selects configurations based on force uncertainty only; the dipole moment model is trained once on the final force-selected dataset without any dipole-specific acquisition or retraining. The assumption that force-disagreement-selected configurations also sufficiently cover the dipole moment surface is load-bearing for the IR spectrum prediction, but it is not validated. The paper reports force uncertainties up to 10^-1 eV/A for 1,3-butadiene but never reports the dipole model's error for that molecule or for the other transferability molecules. The statement in the Discussion that the predicted IR spectra "remain consistent" despite elevated force uncertainties is based on the standard deviation across the three ensemble trajectories, which measures sampling variability, not dipole-model accuracy. The authors should add dipole-model validation on held-out production configurations, or introduce dipole-uncertainty-based acquisition, or provide an explicit argument why force-uncertainty sampling guarantees dipole accuracy.
  3. [Section 2, 'Assessment of ML model transferability'] The claim that the ML models "generalized well to larger, structurally similar molecules" is only partially supported by the main text: 1,3-butadiene shows clear deviations in both frequency and intensity (Figure 7b), and the full set of 8 molecules is only presented in the Supplementary Information. The transferability metrics (PCC and WD) for all 8 molecules should be reported in the main text or at least in a table, and the "generalizes well" wording should be qualified in light of the butadiene result. This is not a fatal issue, but it affects the strength of the generalizability claim made in both the Results and the Conclusion.
  4. [Section 2, 'Infrared spectra calculation and length of dynamical simulation'] The choice of 50 ps as the production trajectory length is based on a convergence study for a single molecule, methanol, comparing 20 ps and 50 ps runs (Figure 4). This length is then applied to all 24 training molecules and all 8 transferability molecules without per-molecule convergence checks. For molecules with slower conformational relaxation or low-frequency modes, 50 ps may not be sufficient for converged intensities. The authors should either provide per-molecule convergence evidence or discuss why the methanol result is expected to transfer to the other molecules in the dataset.
minor comments (6)
  1. [Table 2 and Section 2, 'Performance of PALIRS in predicting infrared spectra'] The text states that the ML predictions "align even more closely with the experimental data than the DFT-based AIMD results," but Table 2 shows Exp.-DFT WD = 0.054 and Exp.-ML WD = 0.057, so the Wasserstein distance slightly favors DFT, while the PCC favors ML (0.80 vs 0.73). The claim should be qualified as metric-dependent or rephrased.
  2. [Section 5.8] A maximum correlation depth of 1000 fs is used in the autocorrelation function, but no convergence study with respect to this parameter is reported. Since the correlation window length affects spectral resolution and statistical noise, a short sensitivity check or a justification of this value would improve the manuscript.
  3. [Section 5.7] DFT-based AIMD uses Berendsen equilibration followed by Nosé-Hoover thermostatting, whereas MLMD production runs use a Langevin thermostat with a friction coefficient of 0.01. The potential effect of different thermostats on the IR intensities is not discussed; a brief justification would be helpful.
  4. [Declarations] There is a typographical error: "Additionaly" should be "Additionally."
  5. [Section 5.1] The MACE model architecture is described, but training hyperparameters such as learning rate, batch size, number of epochs, and loss weighting for energy, forces, and dipole moments are not reported. Providing these details would improve reproducibility.
  6. [Figure 4] The text says peak positions converge by 20 ps, but the relative intensities differ between 20 and 50 ps and the 20 ps spectrum has an "inverse trend" compared with experiment. Clarify in the figure caption or text that convergence is primarily for peak positions, and that intensities require longer trajectories.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: IR spectra follow the standard dipole-autocorrelation formula and are benchmarked against independent DFT-AIMD and NIST experimental references.

full rationale

The central derivation is the dipole-moment autocorrelation expression in Eq. 1, which is a standard physical formula and is not constructed from the model's fitted parameters. The MLIP is trained on DFT energies and forces, and the dipole model is trained on DFT dipole moments; the resulting MLMD spectra are then compared with two external references: DFT-based AIMD spectra and NIST experimental spectra. No fitted parameter is renamed as a prediction: the dipole model's 7.62 mDebye MAE is an evaluated error on a separately labeled test set, not an input used to build the spectrum. The test set is generated with the final MLIP (Section 5.6), but it is used only for model assessment, while the spectral comparisons use separate production MLMD and AIMD runs. The choice of 50 ps trajectories is a methodological decision informed by the convergence analysis in Figure 4, not an equation-level identity. The self-citations [53]-[55] concern software and test-set clustering methodology and do not carry the load-bearing claim that PALIRS reproduces AIMD or experimental IR spectra. The concern that force-uncertainty active learning may not fully cover the dipole surface is a genuine validation limitation, acknowledged in the Discussion for out-of-distribution molecules such as 1,3-butadiene, but it is not an exhibit of a prediction reducing to its inputs by construction. Under the stated hard rules, no circular step can be quoted, so the honest finding is no significant circularity with score 0.

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

The workflow rests on standard DFT and MD methods, an established MLIP architecture (MACE), and a set of hand-chosen active learning and simulation parameters. No new physical entities or forces are introduced. The most consequential unvalidated assumption is that force-based active learning adequately covers the dipole moment surface needed for IR intensities.

free parameters (5)
  • Trajectory length for IR spectra = 50 ps
    Chosen after comparing 20 ps and 50 ps AIMD methanol spectra against NIST experiment (Figure 4); 50 ps adopted for all subsequent AIMD and MLMD runs.
  • Active learning stopping threshold = MAE 5 cm-1 on harmonic frequencies, max 40 iterations
    Stopping criterion chosen at the plateau of harmonic frequency MAE (Figure 2a); controls final dataset size of 16,067 structures.
  • Force uncertainty acquisition parameters = 15 structures/molecule/iteration; 5 per temperature; MD termination threshold 0.5 relative force error
    Hand-selected hyperparameters controlling dataset composition and exploration/exploitation balance; threshold 0.5 terminates unstable MLMD runs.
  • Correlation depth for IR spectra = 1000 fs
    Maximum autocorrelation depth used in Wiener-Khinchin transform; no convergence check with respect to this value is reported.
  • MACE hyperparameters = 128 channels, cutoff 3.0 A, 2 layers, body order 2, L=1
    Standard MACE architecture settings chosen without systematic optimization in this study.
assumptions (5)
  • domain assumption PBE with FHI-aims tier-1 basis and light grid settings provides sufficiently accurate energies, forces, and dipole moments for IR spectra of these molecules.
    All DFT references, training labels, and AIMD spectra use this level of theory (Methods 5.2); experimental agreement is therefore limited by PBE accuracy.
  • domain assumption The IR spectrum from the dipole time-derivative autocorrelation function (Eq. 1) with a 50 ps trajectory and 1000 fs correlation depth is converged.
    Convergence is asserted from methanol intensities in Figure 4; not demonstrated for all 24 molecules.
  • domain assumption Force-uncertainty-based active learning also improves the dipole moment model enough for accurate IR spectra.
    The dipole model is trained on the final 16,067 structures selected by force disagreement; no dipole-specific active learning or validation along production MLMD trajectories is reported.
  • domain assumption A committee of three MACE models with different random seeds provides a reliable uncertainty estimate for force predictions.
    Ensemble uncertainty is used to select data and to terminate unstable MLMD runs (Methods 5.4).
  • domain assumption Harmonic frequency MAE is a valid proxy for MLIP accuracy relevant to anharmonic IR spectra.
    Used as the active learning stopping metric (Figure 2a); anharmonic spectral accuracy is not directly optimized.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Leveraging active learning-enhanced machine-learned interatomic potential for efficient infrared spectra prediction." pith.science (2026). https://pith.science/paper/MBOOX5YT

@misc{pith2026250613486,
  author       = {Pith},
  title        = {Pith review of: Leveraging active learning-enhanced machine-learned interatomic potential for efficient infrared spectra prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MBOOX5YT}},
  note         = {Machine review of arXiv:2506.13486}
}
read the original abstract

Infrared (IR) spectroscopy is a pivotal analytical tool as it provides real-time molecular insight into material structures and enables the observation of reaction intermediates in situ. However, interpreting IR spectra often requires high-fidelity simulations, such as density functional theory based ab-initio molecular dynamics, which are computationally expensive and therefore limited in the tractable system size and complexity. In this work, we present a novel active learning-based framework, implemented in the open-source software package PALIRS, for efficiently predicting the IR spectra of small catalytically relevant organic molecules. PALIRS leverages active learning to train a machine-learned interatomic potential, which is then used for machine learning-assisted molecular dynamics simulations to calculate IR spectra. PALIRS reproduces IR spectra computed with ab-initio molecular dynamics accurately at a fraction of the computational cost. PALIRS further agrees well with available experimental data not only for IR peak positions but also for their amplitudes. This advancement with PALIRS enables high-throughput prediction of IR spectra, facilitating the exploration of larger and more intricate catalytic systems and aiding the identification of novel reaction pathways.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

86 extracted references · 37 canonical work pages

  1. [1]

    Zaera, F.: New advances in the use of infrared absorption spectroscopy for the characterization of heterogeneous catalytic reactions. Chem. Soc. Rev.43, 7624– 7663 (2014) https://doi.org/10.1039/C3CS60374A

  2. [2]

    (ed.) Fourier Transform Infrared Spectroscopy: Fundamentals and Application in Functional Groups and Nanomaterials Characterization, pp

    Khan, S.A., Khan, S.B., Khan, L.U., Farooq, A., Akhtar, K., Asiri, A.M.: In: Sharma, S.K. (ed.) Fourier Transform Infrared Spectroscopy: Fundamentals and Application in Functional Groups and Nanomaterials Characterization, pp. 317–

  3. [3]

    (eds.) IR Spectroscopic Techniques to Study Isolated Biomolecules, pp

    Rijs, A.M., Oomens, J.: In: Rijs, A.M., Oomens, J. (eds.) IR Spectroscopic Techniques to Study Isolated Biomolecules, pp. 1–42. Springer, Cham (2015). https://doi.org/10.1007/128_2014_621

  4. [4]

    19 Surface Science Reports70(4), 449–553 (2015) https://doi.org/10.1016/j.surfrep

    Costa, D., Pradier, C.-M., Tielens, F., Savio, L.: Adsorption and self-assembly of bio-organic molecules at model surfaces: A route towards increased complexity. 19 Surface Science Reports70(4), 449–553 (2015) https://doi.org/10.1016/j.surfrep. 2015.10.002

  5. [5]

    e-Journal of Surface Science and Nanotechnology13, 301–306 (2015) https://doi.org/10.1380/ejssnt.2015.301

    Hamamoto, M., Katsura, M., Nishiyama, N., Tononue, R., Nakashima, S.: Transmission ir micro-spectroscopy of interfacial water between colloidal silica particles. e-Journal of Surface Science and Nanotechnology13, 301–306 (2015) https://doi.org/10.1380/ejssnt.2015.301

  6. [7]

    The Journal of Physical Chemistry C117(43), 22341–22350 (2013) https://doi.org/ 10.1021/jp402737n

    Griffith, E.C., Guizado, T.R.C., Pimentel, A.S., Tyndall, G.S., Vaida, V.: Oxi- dized aromatic–aliphatic mixed films at the air–aqueous solution interface. The Journal of Physical Chemistry C117(43), 22341–22350 (2013) https://doi.org/ 10.1021/jp402737n

  7. [8]

    Nature Communications12(1), 6118 (2021) https://doi.org/10.1038/s41467-021-26416-3

    Su, H., Zhou, W., Zhou, W., Li, Y., Zheng, L., Zhang, H., Liu, M., Zhang, X., Sun, X., Xu, Y., Hu, F., Zhang, J., Hu, T., Liu, Q., Wei, S.: In-situ spectroscopic observation of dynamic-coupling oxygen on atomically dispersed iridium electro- catalyst for acidic water oxidation. Nature Communications12(1), 6118 (2021) https://doi.org/10.1038/s41467-021-26416-3

  8. [9]

    Advanced Functional Materials32(16), 2111193 (2022) https://doi.org/10.1002/adfm.202111193

    Lai, W., Ma, Z., Zhang, J., Yuan, Y., Qiao, Y., Huang, H.: Dynamic evolution of active sites in electrocatalytic co2 reduction reaction: Fundamental understand- ing and recent progress. Advanced Functional Materials32(16), 2111193 (2022) https://doi.org/10.1002/adfm.202111193

Show all 86 references
  1. [10]

    International Journal of Quantum Chemistry115(3), 107–136 (2015) https://doi.org/10.1002/qua.24811

    Deglmann, P., Schäfer, A., Lennartz, C.: Application of quantum calcula- tions in the chemical industry—an overview. International Journal of Quantum Chemistry115(3), 107–136 (2015) https://doi.org/10.1002/qua.24811

  2. [11]

    Nature Communications11(1), 1513 (2020) https://doi.org/10.1038/ s41467-020-15340-7

    Lansford, J.L., Vlachos, D.G.: Infrared spectroscopy data- and physics-driven machine learning for characterizing surface microstructure of complex mate- rials. Nature Communications11(1), 1513 (2020) https://doi.org/10.1038/ s41467-020-15340-7

  3. [12]

    new routes across old boundaries in computational spectroscopy

    Puzzarini, C., Bloino, J., Tasinato, N., Barone, V.: Accuracy and interpretability: The devil and the holy grail. new routes across old boundaries in computational spectroscopy. Chemical Reviews119(13), 8131–8191 (2019) https://doi.org/10. 1021/acs.chemrev.9b00007

  4. [13]

    https://doi.org/10.1016/C2010-0-68479-3

    Larkin, P.: Infrared and Raman Spectroscopy; Principles and Spectral Interpre- tation, (2011). https://doi.org/10.1016/C2010-0-68479-3

  5. [14]

    The Journal of Physical Chemistry B 107(38), 10344–10358 (2003) https://doi.org/10.1021/jp034788u

    Gaigeot, M.-P., Sprik, M.: Ab initio molecular dynamics computation of the 20 infrared spectrum of aqueous uracil. The Journal of Physical Chemistry B 107(38), 10344–10358 (2003) https://doi.org/10.1021/jp034788u

  6. [15]

    Chemical Physics387(1), 1–4 (2011) https://doi.org/10.1016/j.chemphys.2011.06.015

    Rodriguez-Betancourtt, V.-M., Quezada-Navarro, V.-M., Neff, M., Rauhut, G.: Anharmonic frequencies of [f,c,n,x] isomers (x=o,s) obtained from explicitly correlated coupled-cluster calculations. Chemical Physics387(1), 1–4 (2011) https://doi.org/10.1016/j.chemphys.2011.06.015

  7. [16]

    Journal of Computational Chemistry33(27), 2186–2198 (2012) https: //doi.org/10.1002/jcc.23036

    Weymuth, T., Haag, M.P., Kiewisch, K., Luber, S., Schenk, S., Jacob, C.R., Herrmann, C., Neugebauer, J., Reiher, M.: Movipac: Vibrational spectroscopy with a robust meta-program for massively parallel standard and inverse calcu- lations. Journal of Computational Chemistry33(27...

  8. [17]

    Thomas, M., Brehm, M., Fligg, R., Vöhringer, P., Kirchner, B.: Computing vibra- tional spectra from ab initio molecular dynamics. Phys. Chem. Chem. Phys.15, 6608–6622 (2013) https://doi.org/10.1039/C3CP44302G

  9. [18]

    (eds.) Theoretical Meth- ods for Vibrational Spectroscopy and Collision Induced Dissociation in the Gas Phase, pp

    Gaigeot, M.-P., Spezia, R.: In: Rijs, A.M., Oomens, J. (eds.) Theoretical Meth- ods for Vibrational Spectroscopy and Collision Induced Dissociation in the Gas Phase, pp. 99–151. Springer, Cham (2015). https://doi.org/10.1007/128_2014_ 620

  10. [19]

    npj Computational Materials10(1), 1–9 (2024) https: //doi.org/10.1038/s41524-024-01400-9

    Miotto, M., Monacelli, L.: Fast prediction of anharmonic vibrational spectra for complex organic molecules. npj Computational Materials10(1), 1–9 (2024) https: //doi.org/10.1038/s41524-024-01400-9

  11. [20]

    npj Computational Materials10(1), 1–12 (2024) https://doi.org/10.1038/ s41524-024-01236-3

    Bastonero, L., Marzari, N.: Automated all-functionals infrared and Raman spec- tra. npj Computational Materials10(1), 1–12 (2024) https://doi.org/10.1038/ s41524-024-01236-3

  12. [21]

    Behler, J., Parrinello, M.: Generalized neural-network representation of high- dimensional potential-energy surfaces. Phys. Rev. Lett.98, 146401 (2007) https: //doi.org/10.1103/PhysRevLett.98.146401

  13. [22]

    Physical Review Letters104(13) (2010) https: //doi.org/10.1103/PhysRevLett.104.136403

    Bartók, A.P.: Gaussian Approximation Potentials: The Accuracy of Quantum Mechanics, without the Electrons. Physical Review Letters104(13) (2010) https: //doi.org/10.1103/PhysRevLett.104.136403

  14. [23]

    Chemical Science 8(4), 3192–3203 (2017) https://doi.org/10.1039/C6SC05720A

    Smith, J.S., Isayev, O., Roitberg, A.E.: ANI-1: an extensible neural network potential with DFT accuracy at force field computational cost. Chemical Science 8(4), 3192–3203 (2017) https://doi.org/10.1039/C6SC05720A

  15. [24]

    Chemical Science8(10), 6924–6935 (2017) https://doi.org/10.1039/C7SC02267K 21

    Gastegger, M., Behler, J., Marquetand, P.: Machine learning molecular dynamics for the simulation of infrared spectra. Chemical Science8(10), 6924–6935 (2017) https://doi.org/10.1039/C7SC02267K 21

  16. [25]

    The Journal of Chemical Physics148(24), 241722 (2018) https://doi.org/10.1063/1.5019779

    Schütt, K.T., Sauceda, H.E., Kindermans, P.-J., Tkatchenko, A., Müller, K.-R.: SchNet – A deep learning architecture for molecules and materials. The Journal of Chemical Physics148(24), 241722 (2018) https://doi.org/10.1063/1.5019779

  17. [26]

    Journal of Chemical Theory and Computation16(8), 5410–5421 (2020) https://doi.org/10

    Zaverkin, V., Kästner, J.: Gaussian moments as physically inspired molecular descriptors for accurate and scalable machine learning potentials. Journal of Chemical Theory and Computation16(8), 5410–5421 (2020) https://doi.org/10. 1021/acs.jctc.0c00347

  18. [27]

    Journal of Chemical Theory and Computation17(10), 6658–6670 (2021) https://doi.org/10.1021/acs.jctc.1c00527

    Zaverkin, V., Holzmüller, D., Steinwart, I., Kästner, J.: Fast and sample-efficient interatomic neural network potentials for molecules and materials based on gaus- sian moments. Journal of Chemical Theory and Computation17(10), 6658–6670 (2021) https://doi.org/10.1021/acs.jct...

  19. [28]

    Chemical Reviews 121(16), 10073–10141 (2021) https://doi.org/10.1021/acs.chemrev.1c00022

    Deringer, V.L., Bartók, A.P., Bernstein, N., Wilkins, D.M., Ceriotti, M., Csányi, G.: Gaussian process regression for materials and molecules. Chemical Reviews 121(16), 10073–10141 (2021) https://doi.org/10.1021/acs.chemrev.1c00022

  20. [29]

    Nature Communications 13(1), 2453 (2022) https://doi.org/10.1038/s41467-022-29939-5

    Batzner, S., Musaelian, A., Sun, L., Geiger, M., Mailoa, J.P., Kornbluth, M., Molinari, N., Smidt, T.E., Kozinsky, B.: E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials. Nature Communications 13(1), 2453 (2022) https://doi.org/10.103...

  21. [30]

    Nature Communications14(1), 579 (2023) https://doi.org/10.1038/ s41467-023-36329-y

    Musaelian, A., Batzner, S., Johansson, A., Sun, L., Owen, C.J., Kornbluth, M., Kozinsky, B.: Learning local equivariant representations for large-scale atomistic dynamics. Nature Communications14(1), 579 (2023) https://doi.org/10.1038/ s41467-023-36329-y

  22. [31]

    Physical Review Letters 120(3), 036002 (2018) https://doi.org/10.1103/PhysRevLett.120.036002

    Grisafi, A., Wilkins, D.M., Csányi, G., Ceriotti, M.: Symmetry-Adapted Machine Learning for Tensorial Properties of Atomistic Systems. Physical Review Letters 120(3), 036002 (2018) https://doi.org/10.1103/PhysRevLett.120.036002

  23. [32]

    Journal of Chemical Theory and Computation15(6), 3678–3693 (2019) https://doi.org/10.1021/acs.jctc.9b00181

    Unke, O.T., Meuwly, M.: Physnet: A neural network for predicting energies, forces, dipole moments, and partial charges. Journal of Chemical Theory and Computation15(6), 3678–3693 (2019) https://doi.org/10.1021/acs.jctc.9b00181

  24. [33]

    Chemical Science12(34), 11473–11483 (2021) https://doi.org/10.1039/D1SC02742E

    Gastegger, M., Schütt, K.T., Müller, K.-R.: Machine learning of solvent effects on molecular spectra and reactions. Chemical Science12(34), 11473–11483 (2021) https://doi.org/10.1039/D1SC02742E

  25. [34]

    In: Meila, M., Zhang, T

    Schütt, K., Unke, O., Gastegger, M.: Equivariant message passing for the pre- diction of tensorial properties and molecular spectra. In: Meila, M., Zhang, T. (eds.) Proceedings of the 38th International Conference on Machine Learning. Proceedings of Machine Learning Research, ...

  26. [35]

    Journal of Chemical Theory and Computation18(9), 5492–5501 (2022) https://doi.org/10.1021/acs

    Beckmann, R., Brieuc, F., Schran, C., Marx, D.: Infrared Spectra at Coupled 22 Cluster Accuracy from Neural Network Representations. Journal of Chemical Theory and Computation18(9), 5492–5501 (2022) https://doi.org/10.1021/acs. jctc.2c00511

  27. [36]

    https://doi.org/10.48550/arXiv

    Batatia, I., Batzner, S., Kovács, D.P., Musaelian, A., Simm, G.N.C., Drautz, R., Ortner, C., Kozinsky, B., Csányi, G.: The Design Space of E(3)-Equivariant Atom-Centered Interatomic Potentials (2022). https://doi.org/10.48550/arXiv. 2205.06643

  28. [37]

    Advances in Neural Information Processing Systems35, 11423–11436 (2022)

    Batatia, I., Kovacs, D.P., Simm, G., Ortner, C., Csányi, G.: Mace: Higher order equivariant message passing neural networks for fast and accurate force fields. Advances in Neural Information Processing Systems35, 11423–11436 (2022)

  29. [38]

    The Journal of Chemical Physics 158(22), 224108 (2023) https://doi.org/10.1063/5.0150379

    Tang, Z., Bromley, S.T., Hammer, B.: A machine learning potential for simu- lating infrared spectra of nanosilicate clusters. The Journal of Chemical Physics 158(22), 224108 (2023) https://doi.org/10.1063/5.0150379

  30. [39]

    Nature Computational Science3(11), 957–964 (2023) https://doi.org/10.1038/ s43588-023-00550-y

    Zou, Z., Zhang, Y., Liang, L., Wei, M., Leng, J., Jiang, J., Luo, Y., Hu, W.: A deep learning model for predicting selected organic molecular spectra. Nature Computational Science3(11), 957–964 (2023) https://doi.org/10.1038/ s43588-023-00550-y

  31. [40]

    Journal of Chemical Information and Modeling 64(12), 4613–4629 (2024) https://doi.org/10.1021/acs.jcim.4c00378

    Stienstra, C.M.K., Hebert, L., Thomas, P., Haack, A., Guo, J., Hopkins, W.S.: Graphormer-IR: Graph Transformers Predict Experimental IR Spectra Using Highly Specialized Attention. Journal of Chemical Information and Modeling 64(12), 4613–4629 (2024) https://doi.org/10.1021/acs...

  32. [41]

    Chemical Papers78(5), 3149–3173 (2024) https://doi.org/10.1007/ s11696-024-03301-z

    Krzyżanowski, M., Matyszczak, G.: Machine learning prediction of organic moieties from the IR spectra, enhanced by additionally using the derivative IR data. Chemical Papers78(5), 3149–3173 (2024) https://doi.org/10.1007/ s11696-024-03301-z

  33. [42]

    https://arxiv.org/abs/2501.18876

    Yuan, M., Zou, Z., Hu, W.: QMe14S, A Comprehensive and Efficient Spectral Dataset for Small Organic Molecules (2025). https://arxiv.org/abs/2501.18876

  34. [43]

    The Journal of Chemical Physics153(3), 034702 (2020) https://doi.org/10.1063/5.0005084

    Rowe,P.,Deringer,V.L.,Gasparotto,P.,Csányi,G.,Michaelides,A.:Anaccurate and transferable machine learning potential for carbon. The Journal of Chemical Physics153(3), 034702 (2020) https://doi.org/10.1063/5.0005084

  35. [44]

    Physical Review X8(4), 041048 (2018)

    Bartók, A.P., Kermode, J., Bernstein, N., Csányi, G.: Machine learning a general- purpose interatomic potential for silicon. Physical Review X8(4), 041048 (2018)

  36. [45]

    Nature Communications11(1), 5461 (2020) https://doi.org/10.1038/s41467-020-19168-z

    Deringer, V.L., Caro, M.A., Csányi, G.: A general-purpose machine-learning force field for bulk and nanostructured phosphorus. Nature Communications11(1), 5461 (2020) https://doi.org/10.1038/s41467-020-19168-z

  37. [46]

    Computational Materials Science140, 171–180 (2017) https://doi.org/10.1016/j.commatsci.2017.08.031

    Podryabinkin, E.V., Shapeev, A.V.: Active learning of linearly parametrized 23 interatomic potentials. Computational Materials Science140, 171–180 (2017) https://doi.org/10.1016/j.commatsci.2017.08.031

  38. [47]

    Computational Materials Science156, 148–156 (2019) https://doi.org/10

    Gubaev, K., Podryabinkin, E.V., Hart, G.L.W., Shapeev, A.V.: Accelerating high-throughputsearchesfornewalloyswithactivelearningofinteratomicpoten- tials. Computational Materials Science156, 148–156 (2019) https://doi.org/10. 1016/j.commatsci.2018.09.031

  39. [48]

    npj Computational Materials6(1), 1–11 (2020) https: //doi.org/10.1038/s41524-020-0283-z

    Vandermause, J., Torrisi, S.B., Batzner, S., Xie, Y., Sun, L., Kolpak, A.M., Kozinsky, B.: On-the-fly active learning of interpretable Bayesian force fields for atomistic rare events. npj Computational Materials6(1), 1–11 (2020) https: //doi.org/10.1038/s41524-020-0283-z

  40. [49]

    npj Computational Materials9(1), 1–14 (2023) https://doi.org/10.1038/s41524-023-01104-6

    Oord, C., Sachs, M., Kovács, D.P., Ortner, C., Csányi, G.: Hyperactive learning for data-driven interatomic potentials. npj Computational Materials9(1), 1–14 (2023) https://doi.org/10.1038/s41524-023-01104-6

  41. [50]

    Nature Computational Science3(3), 230–239 (2023) https: //doi.org/10.1038/s43588-023-00406-5

    Kulichenko, M., Barros, K., Lubbers, N., Li, Y.W., Messerly, R., Tretiak, S., Smith, J.S., Nebgen, B.: Uncertainty-driven dynamics for active learning of inter- atomic potentials. Nature Computational Science3(3), 230–239 (2023) https: //doi.org/10.1038/s43588-023-00406-5

  42. [51]

    npj Computational Materials9(1), 1– 11 (2023) https://doi.org/10.1038/s41524-023-01180-8

    Tan, A.R., Urata, S., Goldman, S., Dietschreit, J.C.B., Gómez-Bombarelli, R.: Single-model uncertainty quantification in neural network potentials does not consistently outperform model ensembles. npj Computational Materials9(1), 1– 11 (2023) https://doi.org/10.1038/s41524-023-01180-8

  43. [52]

    npj Computational Materials10(1), 1–18 (2024) https://doi.org/10.1038/s41524-024-01254-1

    Zaverkin, V., Holzmüller, D., Christiansen, H., Errica, F., Alesiani, F., Takamoto, M., Niepert, M., Kästner, J.: Uncertainty-biased molecular dynamics for learning uniformly accurate interatomic potentials. npj Computational Materials10(1), 1–18 (2024) https://doi.org/10.1038...

  44. [53]

    Ghosh, K., Todorović, M., Vehtari, A., Rinke, P.: Active learning of molecular data for task-specific objectives. J. Chem. Phys.162(1), 014103 (2025)

  45. [54]

    Homm,H.,Laakso,J.,Rinke,P.:Efficientdatasetgenerationformachinelearning halide perovskite alloys. Phys. Rev. Mater.9, 053802 (2025)

  46. [55]

    GitLab (2024)

    Bhatia, N., Krejčí, O., Rinke, P.: PALIRS. GitLab (2024). https://gitlab.com/ cest-group/PALIRS

  47. [56]

    The Journal of Chemical Physics153(10), 104105 (2020) https://doi.org/10.1063/5.0016004

    Schran, C., Brezina, K., Marsalek, O.: Committee neural network potentials con- trol generalization errors and enable active learning. The Journal of Chemical Physics153(10), 104105 (2020) https://doi.org/10.1063/5.0016004

  48. [57]

    The Journal of Chemical Physics 24 148(24), 241733 (2018) https://doi.org/10.1063/1.5023802

    Smith, J.S., Nebgen, B., Lubbers, N., Isayev, O., Roitberg, A.E.: Less is more: Sampling chemical space with active learning. The Journal of Chemical Physics 24 148(24), 241733 (2018) https://doi.org/10.1063/1.5023802

  49. [58]

    Qu, C., Yu, Q., Bowman, J.M.: Permutationally Invariant Potential Energy Sur- faces.AnnualReviewofPhysicalChemistry69(Volume69,2018),151–175(2018) https://doi.org/10.1146/annurev-physchem-050317-021139

  50. [59]

    Computer Physics Communications180(11), 2175–2196 (2009) https://doi.org/ 10.1016/j.cpc.2009.06.022

    Blum, V., Gehrke, R., Hanke, F., Havu, P., Havu, V., Ren, X., Reuter, K., Schef- fler, M.: Ab initio molecular simulations with numeric atom-centered orbitals. Computer Physics Communications180(11), 2175–2196 (2009) https://doi.org/ 10.1016/j.cpc.2009.06.022

  51. [60]

    Journal of Computational Physics228(22), 8367–8379 (2009) https://doi.org/10.1016/j

    Havu, V., Blum, V., Havu, P., Scheffler, M.: EfficientO(N)integration for all- electron electronic structure calculation using numeric basis functions. Journal of Computational Physics228(22), 8367–8379 (2009) https://doi.org/10.1016/j. jcp.2009.08.008

  52. [61]

    Computer Physics Communications192, 60–69 (2015) https://doi.org/10.1016/j.cpc.2015.02.021

    Levchenko, S.V., Ren, X., Wieferink, J., Johanni, R., Rinke, P., Blum, V., Schef- fler, M.: Hybrid functionals for large periodic systems in an all-electron, numeric atom-centered basis framework. Computer Physics Communications192, 60–69 (2015) https://doi.org/10.1016/j.cpc.2...

  53. [62]

    Ren, X., Rinke, P., Blum, V., Wieferink, J., Tkatchenko, A., Andrea, S., Reuter, K., Blum, V., Scheffler, M.: Resolution-of-identity approach to Hartree-Fock, hybrid density functionals, RPA, MP2, and GW with numeric atom-centered orbital basis functions. New J. Phys.14, 053020 (2012)

  54. [63]

    Machine Learning: Science and Technology3(4), 045017 (2022) https: //doi.org/10.1088/2632-2153/aca005

    Huo, H., Rupp, M.: Unified representation of molecules and crystals for machine learning. Machine Learning: Science and Technology3(4), 045017 (2022) https: //doi.org/10.1088/2632-2153/aca005

  55. [64]

    Computer Physics Communications247, 106949 (2020) https://doi.org/10.1016/j.cpc.2019.106949

    Himanen, L., Jäger, M.O.J., Morooka, E.V., Federici Canova, F., Ranawat, Y.S., Gao, D.Z., Rinke, P., Foster, A.S.: Dscribe: Library of descriptors for machine learning in materials science. Computer Physics Communications247, 106949 (2020) https://doi.org/10.1016/j.cpc.2019.106949

  56. [65]

    Journal of Chemical Theory and Computation17(2), 985–995 (2021) https://doi.org/10.1021/acs.jctc.0c01279

    Esch, B.v.d., Peters, L.D.M., Sauerland, L., Ochsenfeld, C.: Quantitative Com- parison of Experimental and Computed IR-Spectra Extracted from Ab Initio Molecular Dynamics. Journal of Chemical Theory and Computation17(2), 985–995 (2021) https://doi.org/10.1021/acs.jctc.0c01279

  57. [66]

    Infrared Spectra

    Wallace, W.E.: NIST Chemistry WebBook, NIST Standard Reference Database Number 69. Linstrom, P. J. and Mallard, W. G.; National Institute of Standards and Technology; Gaithersburg, MD. Vol. 20899, (retrieved September 12, 2024); Chapter "Infrared Spectra" by NIST Mass Spectrom...

  58. [67]

    Chemical Physics514, 44–54 (2018) https: //doi.org/10.1016/j.chemphys.2017.12.015

    Sagiv, L., Hirshberg, B., Gerber, R.B.: Anharmonic vibrational spectroscopy cal- culations using theab initioCSP method: Applications to H2CO3, (H2CO3)2, H2CO3-H2O and isotopologues. Chemical Physics514, 44–54 (2018) https: //doi.org/10.1016/j.chemphys.2017.12.015

  59. [68]

    Perdew, J.P., Burke, K., Ernzerhof, M.: Generalized Gradient Approximation Made Simple [Phys. Rev. Lett. 77, 3865 (1996)]. Physical Review Letters78(7), 1396–1396 (1997) https://doi.org/10.1103/PhysRevLett.78.1396

  60. [69]

    The Journal of Chemical Physics99(6), 4597–4610 (1993) https: //doi.org/10.1063/1.466059

    Lenthe, E.v., Baerends, E.J., Snijders, J.G.: Relativistic regular two-component hamiltonians. The Journal of Chemical Physics99(6), 4597–4610 (1993) https: //doi.org/10.1063/1.466059

  61. [70]

    Physical ReviewLetters102(7),073005(2009)https://doi.org/10.1103/PhysRevLett.102

    Tkatchenko, A., Scheffler, M.: Accurate Molecular Van Der Waals Interactions from Ground-State Electron Density and Free-Atom Reference Data. Physical ReviewLetters102(7),073005(2009)https://doi.org/10.1103/PhysRevLett.102. 073005

  62. [71]

    Springer Series in Operations Research and Financial Engineering

    Nocedal, J., Wright, S.J.: Numerical Optimization. Springer Series in Operations Research and Financial Engineering. Springer, ??? (2006). https://doi.org/10. 1007/978-0-387-40065-5 .http://link.springer.com/10.1007/978-0-387-40065-5

  63. [72]

    Journal of Physics: Condensed Matter29(27), 273002 (2017)

    Larsen, A.H., Mortensen, J.J., Blomqvist, J., Castelli, I.E., Christensen, R., Dułak, M., Friis, J., Groves, M.N., Hammer, B., Hargus, C., Hermes, E.D., Jen- nings, P.C., Jensen, P.B., Kermode, J., Kitchin, J.R., Kolsbjerg, E.L., Kubal, J., Kaasbjerg, K., Lysgaard, S., Maronss...

  64. [73]

    Scientific Data1(1), 140022 (2014) https://doi.org/10.1038/sdata.2014.22

    Ramakrishnan, R., Dral, P.O., Rupp, M., Lilienfeld, O.A.: Quantum chemistry structures and properties of 134 kilo molecules. Scientific Data1(1), 140022 (2014) https://doi.org/10.1038/sdata.2014.22

  65. [74]

    Tech- nical Report MSR-TR-2000-65, Microsoft Research (May 2000)

    Bennett, K.P., Bradley, P.S., Demiriz, A.: Constrained k-means clustering. Tech- nical Report MSR-TR-2000-65, Microsoft Research (May 2000). https://www. microsoft.com/en-us/research/publication/constrained-k-means-clustering/

  66. [75]

    The Journal of Chemical Physics81(8), 3684–3690 (1984) https://doi.org/10.1063/1.448118

    Berendsen, H.J.C., Postma, J.P.M., Gunsteren, W.F., DiNola, A., Haak, J.R.: Molecular dynamics with coupling to an external bath. The Journal of Chemical Physics81(8), 3684–3690 (1984) https://doi.org/10.1063/1.448118

  67. [76]

    The Journal of Chemical Physics81(1), 511–519 (1984) https://doi

    Nosé, S.: A unified formulation of the constant temperature molecular dynamics methods. The Journal of Chemical Physics81(1), 511–519 (1984) https://doi. org/10.1063/1.447334 26

  68. [77]

    Phys- ical Review A31(3), 1695–1697 (1985) https://doi.org/10.1103/PhysRevA.31

    Hoover, W.G.: Canonical dynamics: Equilibrium phase-space distributions. Phys- ical Review A31(3), 1695–1697 (1985) https://doi.org/10.1103/PhysRevA.31. 1695

  69. [78]

    Physical Review Letters102(2), 020601 (2009) https://doi.org/10.1103/PhysRevLett.102.020601

    Ceriotti, M., Bussi, G., Parrinello, M.: Langevin Equation with Colored Noise for Constant-Temperature Molecular Dynamics Simulations. Physical Review Letters102(2), 020601 (2009) https://doi.org/10.1103/PhysRevLett.102.020601

  70. [79]

    Wiener,N.:Generalizedharmonicanalysis.ActaMathematica55(none),117–258 (1930) https://doi.org/10.1007/BF02546511

  71. [80]

    The Bell System Tech- nical Journal37(1), 185–282 (1958) https://doi.org/10.1002/j.1538-7305.1958

    Blackman, R.B., Tukey, J.W.: The measurement of power spectra from the point of view of communications engineering — Part I. The Bell System Tech- nical Journal37(1), 185–282 (1958) https://doi.org/10.1002/j.1538-7305.1958. tb03874.x

  72. [81]

    Journal of Chemical Theory and Computation15(1), 448–455 (2019) https://doi.org/10

    Schütt, K.T., Kessel, P., Gastegger, M., Nicoli, K.A., Tkatchenko, A., Müller, K.-R.: SchNetPack: A Deep Learning Toolbox For Atomistic Systems. Journal of Chemical Theory and Computation15(1), 448–455 (2019) https://doi.org/10. 1021/acs.jctc.8b00908

  73. [82]

    Journal of Machine Learning Research12, 2825–2830 (2011)

    Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Prettenhofer, P., Weiss, R., Dubourg, V., Vanderplas, J., Passos, A., Cournapeau, D., Brucher, M., Perrot, M., Duchesnay, E.: Scikit-learn: Machine learning in Python. Journal of Mach...

  74. [83]

    Journal of Chemical Theory and Computation16(5), 3307–3315 (2020) https://doi.org/10.1021/acs.jctc.0c00126

    Henschel, H., Andersson, A.T., Jespers, W., Mehdi Ghahremanpour, M., Spoel, D.: Theoretical Infrared Spectra: Quantitative Similarity Measures and Force Fields. Journal of Chemical Theory and Computation16(5), 3307–3315 (2020) https://doi.org/10.1021/acs.jctc.0c00126

  75. [84]

    Journal of Chemical Theory and Computation16(11), 7044–7060 (2020) https://doi.org/10.1021/acs.jctc.0c00877

    Pracht, P., Grant, D.F., Grimme, S.: Comprehensive Assessment of GFN Tight- Binding and Composite Density Functional Theory Methods for Calculating Gas- Phase Infrared Spectra. Journal of Chemical Theory and Computation16(11), 7044–7060 (2020) https://doi.org/10.1021/acs.jctc.0c00877

  76. [85]

    International Journal of Computer Vision40(2), 99–121 (2000) https://doi.org/10.1023/A:1026543900054

    Rubner, Y., Tomasi, C., Guibas, L.J.: The Earth Mover’s Distance as a Metric for Image Retrieval. International Journal of Computer Vision40(2), 99–121 (2000) https://doi.org/10.1023/A:1026543900054

  77. [86]

    Himanen, L., Geurts, A., Foster, A.S., Rinke, P.: Data-driven materials science: Status, challenges, and perspectives. Adv. Sci.6(21), 1900808 (2019) 27

  78. [344]

    https://doi.org/10.1007/978-3-319-92955-2_9

    Springer, Cham (2018). https://doi.org/10.1007/978-3-319-92955-2_9

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.