Pith. sign in

REVIEW 3 major objections 4 minor 59 references

Active learning and explicit electrostatics enable accurate modeling of electrolytes

T0 review · 3 major / 4 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read Active learning can automatically build machine-learned potentials that simulate carbonate electrolytes, with ionic conductivities within 11% of experiment (6% when explicit electrostatics is added).

desk verdict Solid active-learning pipeline for electrolyte MTPs; the 6% explicit-electrostatics conductivity claim is softer than the abstract suggests. read the letter →

arxiv 2510.03479 v3 pith:KS744ZOV submitted 2025-10-03 physics.chem-ph

Olga Chalykh , Mikhail Polovinkin , Dmitry Korogod , Nikita Rybin , Alexander Shapeev This is my paper
classification physics.chem-ph
keywords activelearningmomenttensorpotentialsmachine-learnedinteratomicionicconductivityelectrolyteschargeredistributionGreen-KuboEC/EMCsolvent
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 makes two connected claims. First, D-optimality-based active learning—adding configurations only when a model's extrapolation grade crosses a threshold—can replace the manual, physically motivated dataset augmentation that earlier EC/EMC machine-learned potential work required, and can yield moment tensor potentials with stable liquid densities for ethylene carbonate, ethyl methyl carbonate, mixtures, and LiPF6 solutions. Second, explicitly adding electrostatics through a charge-redistribution (QRd) term delivers equal or better accuracy with fewer parameters in solvent mixtures, and predicts ionic conductivity with a 6% mean deviation from experiment. A reader should care because composition screening for battery electrolytes is bottlenecked by expensive training-set construction and by local potentials' neglect of long-range electrostatics. The salt-solution test is more equivocal: the fixed-charge QRd model produced only 2.8 ns of total trajectory, and the paper itself warns this undermines the reliability of its 6% conductivity figure.

What carries the argument

Two components carry the argument. (1) D-optimality active learning: from the matrix of energy derivatives with respect to MTP parameters, the MaxVol algorithm selects the most linearly independent rows; the extrapolation grade γ of any new configuration is the largest component of its derivative vector expressed in that basis, with thresholds γ_save≈2 and γ_break≈10 controlling when configurations are added and when AL-MD is stopped. This automates dataset growth. (2) The QRd charge-redistribution term: point charges qi = b_zi + s_zi (Q_total − Σb)/(Σs), fixed per atomic type, are added to the short-range MTP energy and fitted alongside the MTP parameters. Its analytic simplicity is what le

What would settle it

Run MTP20-QRd (or an environment-dependent-charge variant) for a 25 ns trajectory at 300 K in 3EC:7EMC with a 500 ps correlation time, matching the MTP20 protocol used here, and check whether the mean absolute deviation from experimental conductivity stays below 6%; if the Green-Kubo plateau shifts or the simulation blows up, that number is an artifact of short trajectories.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that the D-optimality active-learning criterion, using the MaxVol algorithm and an extrapolation-grade threshold, can automatically assemble compact training sets (hundreds to about 6300 configurations) that keep MTP-based molecular dynamics in the liquid phase without manual density inflation or isolated-molecule augmentation. The resulting potentials transfer across EC/EMC ratios and 280–320 K, matching measured densities to about 6% and ionic conductivities to within 11% (MTP20). The paper further argues that augmenting a level-16 MTP with the QRd fixed-charge Coulomb term reaches the accuracy of a level-20 plain MTP with 389 rather than 651 machine

Load-bearing premise

The 6% conductivity result for MTP20-QRd rests on assuming that 2.8 ns of accumulated trajectory (14 completed runs at 300 K, 100 ps equilibration, 100 ps correlation time) is enough for a converged Green-Kubo integral—an assumption the paper itself says undermines the reliability of that number.

Editorial extensions

If this is right

  • The AL pipeline removes the need for manual training-set augmentation (density inflation/deflation, isolated molecules) that earlier EC/EMC MTP work required; training sets of roughly 500–1400 configurations sufficed for pure solvents and mixtures.
  • A short-range MTP without explicit electrostatics is enough for roughly 11% ionic-conductivity accuracy at multiple EC:EMC ratios and three temperatures, supporting the view that most electrostatics is captured within a 5 Å cutoff.
  • Explicit electrostatics buys parameter efficiency in the stable regime: a level-16 QRd model with 389 parameters matches a level-20 plain MTP with 651 parameters on energies and forces of solvent mixtures.
  • The QRd model's transport accuracy is coupled to a stability penalty in salt solutions: it reached 6% conductivity only at the trained composition and on trajectory statistics the paper itself calls insufficient.
  • Force errors concentrate on carbonyl carbons (and, in the salt solution, phosphorus), yet the Li–O radial distribution function still matched AIMD, so substantial relative force errors on Li+ (~40%) were tolerable for structure prediction by MTP20.

Reading between the lines

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

  • If environment-dependent charges (charge equilibration or equivariant charge prediction) stabilize the long-range term, the parameter-count and training-set advantages seen here could extend to salt solutions; the paper's own discussion points the same way.
  • The 6% conductivity figure should be read as provisional until a fixed-charge or environment-dependent QRd model sustains multi-nanosecond trajectories with a correlation time of at least 500 ps; a stable rerun is a direct test.
  • The near-identical conductivities despite clearly different ion-pair populations (MTP20: 100% solvent-separated pairs; MTP20-QRd: about 95% SSIPs plus some contact pairs and aggregates) suggest ionic conductivity at this salt concentration is insensitive to pairing in short trajectories—useful for prediction but weak as a structural discriminator.
  • The submitted metadata abstract says environment-dependent charges 'further improve accuracy and stability,' but the body reports MD instability for the implemented fixed-charge QRd scheme and lists environment-dependent charges as future work; readers should weigh the body's evidence over the abstract's wording.
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 / 4 minor

Summary. The manuscript presents a D-optimality/MaxVol active-learning workflow for training system-specific moment tensor potentials (MTPs) for EC/EMC binary solvents and 1M LiPF6 in EC/EMC, and compares the resulting short-range MTPs with MTP-QRd, an MTP augmented by a fixed-charge redistribution (QRd) term. The MTPs are reported to give stable densities (within ~6%) for pure solvents and mixtures, accurate Li–O RDFs, and ionic conductivities within 11% of experiment across temperatures and solvent compositions. The MTP-QRd model is reported to reach comparable or better accuracy than MTP at lower MTP level for salt-free mixtures and to reproduce the 3:7 EC:EMC LiPF6 conductivity within 6% of experiment, but its MD is acknowledged to be unstable, with only 2.8 ns of usable trajectory at 300 K and a much shorter equilibration/correlation protocol.

Significance. If upheld, the paper would establish a practical, hands-off AL pipeline for building system-specific MLIPs for liquid electrolytes, with strong validation of the short-range MTP: 25 ns of stable MD, ~6% density error, 11% conductivity error, and Li–O RDFs that improve on ReaxFF relative to AIMD. The D-optimality/MaxVol procedure is well-defined, the MLIP-2 and MLIP-4 codes are public, and the training sets are promised for release; these are concrete strengths. The MTP-QRd parameter-efficiency claim for EC/EMC mixtures is plausible but incompletely demonstrated because some QRd models could not cover all compositions. The headline 6% explicit-electrostatics conductivity result is the main liability: the underlying trajectories are short, unstable, and the paper itself states that this undermines the result's reliability.

major comments (3)
  1. [2.3.3, 4.6, Table 1] The central abstract claim of 6% ionic conductivity for MTP20-QRd rests on 17/14/6 completed runs out of 50 at 280/300/320 K, total sampling of 3.4/2.8/1.2 ns, 100 ps equilibration (versus 5 ns for MTP), a 100 ps correlation time, and averaging over the last 20 ps of the Green–Kubo integral. The manuscript itself states that this 'undermines the reliability of the ionic conductivity values obtained with MTP20-QRd' (Sec. 2.3.3). The subsequent assertion that the value is 'unlikely to be severely affected' is not supported by any demonstrated decay of the current autocorrelation function to zero within 100 ps, nor by convergence checks with respect to equilibration length or starting configuration. The 6% figure and the related statement in Sec. 3 that MTP-QRd 'outperformed MTP by 5%' should be removed from the abstract/discussion or replaced by a clearly labeled preliminary estimate accom
  2. [2.2, 2.3.3, 3] The parameter-efficiency advantage claimed for MTP-QRd is broader than the evidence. For EC/EMC mixtures the paper notes that 'some MTP-QRd models could not model all mixture compositions' (Sec. 2.2), and for LiPF6 solution MTP20-QRd was stable only in the single 3EC:7EMC composition corresponding to its training set (Sec. 2.3.3). The abstract's wording that the extended MTP 'achieves accuracy comparable to MTP ... with fewer parameters' therefore needs an explicit scope restriction to the compositions where QRd MD is actually stable. As written, the claim suggests a general advantage that the paper's own stability data contradict.
  3. [Abstract (front matter) vs 4.2, 2.3.2] The front-matter abstract states that charge redistribution was assessed 'using either fixed or environment-dependent charges' and that 'environment-dependent charges further improve accuracy and the stability of simulations.' The body implements only the fixed-charge QRd scheme (Sec. 4.2, Eq. 8) and reports that QRd MD is unstable for LiPF6 in pure EC (Sec. 2.3.2); no environment-dependent charge model is trained or tested. This is an internal inconsistency that attributes to the paper a result it does not contain. The abstract must be corrected to describe only the fixed-charge QRd model that was actually used.
minor comments (4)
  1. [Figure 5 caption] The caption contains two panels labelled (d): 'd) Coordination number (CN) distributions in ionic pairs present in training set' and then 'd) Force error magnitudes...' The second should be relettered (e) or (f), and subsequent panel labels adjusted.
  2. [Supplementary Information, S1] The text 'where xx% of the EC and xx% of the EMC configurations are inherited from the initial training set' contains unprocessed placeholders. Please replace with actual percentages.
  3. [Throughout] Several typographical errors need correction: 'unpysical' (Sec. 2.3.2), 'Maxvell' (Sec. 4.6), 'primitivity' (Sec. 2.3.3), 'respressed' (Sec. 4.2), 'neigboring' (Sec. 4.1), 'compering' (Sec. 9), and 'on the over hand' (Sec. 2.3.3).
  4. [Figure 3] The caption says RMSEs are 'compared with PBE-D3 calculations,' which is clear, but the text could state explicitly that reference energies/forces are DFT values computed on MTP20-generated validation configurations, so that readers do not mistake the validation set for AIMD-sampled configurations.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation: conductivity and density are forward MD predictions benchmarked against external experiment; self-citations are present but not load-bearing.

full rationale

The central claims are not circular by construction. Ionic conductivity is obtained by Green-Kubo integration (Eq. 13) of charge-current autocorrelation from equilibrium MD after training; the training loss (Eq. 10) contains only DFT energies and forces, so neither conductivity nor density is a fitted target. Densities are compared to experimental interpolation [42,14], and conductivities are compared to experimental interpolation [46]. Active-learning selection uses the extrapolation grade (Eqs. 11-12), but the training labels come from DFT and the final property evaluations are external benchmarks. MTP-QRd parameters are fit to the same DFT data, and its 6% conductivity deviation is a forward simulation result, not a recycled input. Some citations are to the authors' own MLIP/AL software and the unpublished QRd paper [38], but the QRd equations are given in Sec. 4.2, and the model is benchmarked here against independent DFT and experiment, so these citations are not load-bearing and do not force the conclusion. The paper itself flags a serious sampling limitation for MTP20-QRd — "This fact undermines the reliability of the ionic conductivity values obtained with MTP20-QRd" (Sec. 2.3.3) — which is a convergence/robustness concern rather than a circularity. Given only minor, non-load-bearing self-citations, the circularity score is low.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The paper contributes a workflow and benchmarks, not a closed-form derivation; the ledger therefore captures the fitting choices (MTP/QRd parameters, AL thresholds, cutoff, loss weights) and the modeling assumptions (DFT reference quality, γ validity, Green-Kubo convergence, ion-only charge current, cutoff sufficiency). No new physical entities are introduced.

free parameters (6)
  • MTP linear and radial parameters = not enumerated (383 at level 16, 651 at level 20)
    Fitted by BFGS to DFT energies/forces (Sec. 4.3); these are the machine-learning parameters whose count is central to the "fewer parameters" claim.
  • QRd charge parameters b, s per element = not enumerated
    Fitted to DFT data in the MTP-QRd model (Sec. 4.2); fixed per atomic type, causing the carbonyl/ether oxygen degeneracy the authors acknowledge.
  • Active learning thresholds γ_save≈2, γ_break≈10 = ≈2, ≈10
    Hand-chosen in Sec. 4.4; control which configurations enter the training set and when to terminate AL-MD trajectories.
  • MTP cutoff Rcut = 5 Å
    Chosen following Ref. [14] (Sec. 2.1); excludes long-range electrostatics beyond 5 Å, motivating the QRd comparison.
  • Loss function weights w_e and w_f = 1 and 0.01
    Set in Sec. 4.3; balance energy and force fitting, a modeling choice affecting all trained potentials.
  • MTP level (12/16/20) = levels 12, 16, 20 compared
    Model capacity chosen by the authors; the central parameter-efficiency claim compares level-16 QRd against level-20 MTP.
assumptions (5)
  • domain assumption PBE-D3 DFT with PAW pseudopotentials is an accurate reference for EC/EMC/LiPF6 energies and forces.
    Invoked in Sec. 4.5 with justification from Ref. [14]; all MLIP training and validation inherits this functional's systematic errors.
  • domain assumption The extrapolation grade γ is a valid uncertainty measure for selecting diverse and physically relevant configurations.
    Assumed from refs [24,25] and used throughout Sec. 4.4; the entire AL pipeline rests on this.
  • domain assumption Green-Kubo conductivity converges at τc=500 ps for MTP and τc=100 ps for MTP-QRd.
    Convergence examined in Sec. 4.6 and SI; for MTP-QRd the choice is acknowledged as a minimal value from MTP-based reference calculations despite much shorter trajectories.
  • domain assumption Charge current in LiPF6/EC/EMC is carried only by Li+ and PF6−; the P-atom velocity represents the PF6− center of mass.
    Stated in Sec. 4.6 as an approximation to reduce noise; solvent-mediated charge transport is neglected.
  • domain assumption A 5 Å cutoff captures the essential short-range interactions for these molecular liquids.
    Adopted in Sec. 2.1 following Ref. [14]; the QRd comparison is motivated precisely by this locality limitation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Active learning and explicit electrostatics enable accurate modeling of electrolytes." pith.science (2026). https://pith.science/paper/KS744ZOV

@misc{pith2026251003479,
  author       = {Pith},
  title        = {Pith review of: Active learning and explicit electrostatics enable accurate modeling of electrolytes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KS744ZOV}},
  note         = {Machine review of arXiv:2510.03479}
}
abstract

Machine learning interatomic potentials (MLIPs) offer near-\textit{ab initio} accuracy with the efficiency of classical force fields, making them attractive for modeling electrolytes. Collecting a diverse training set is essential for their accuracy and reliability, and explicit treatment of strong electrostatic interactions may be necessary. In this work, we demonstrated that D-optimality-based active learning can automatically generate diverse training sets for moment tensor potentials (MTPs), enabling reliable molecular dynamics simulations of pure ethylene carbonate (EC), ethyl methyl carbonate (EMC), their mixtures, and LiPF$_6$ solutions. The resulting MTPs exhibit excellent transferability across various EC/EMC compositions, producing ionic conductivities within 11\% mean deviations from experiments. In addition, we assessed the impact of explicitly incorporating electrostatics by augmenting MTP with charge redistribution schemes using either fixed or environment-dependent charges. Our results show that the augmented MTP achieves the same or higher accuracy than standard model with fewer parameters, while environment-dependent charges further improve accuracy and the stability of simulations.

Figures

Figures reproduced from arXiv: 2510.03479 by the authors.

Figure 1
Figure 1. a) Mean densities predicted by an ensemble of 3 MTPs, together with 1- [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Mean densities predicted by an ensemble of three MLIPs with 1-σ confidence intervals, com￾pared to literature values [14] obtained by interpo￾lating experimental data [42]. reaching 11% mean percent deviation in the latter case. Furthermore, MTP16 predictions exhibit high uncertainty associated with the random initialization of MTP parameters. This causes a large overlap in the 1-σ intervals of the predicted densiti… view at source ↗
Figure 3
Figure 3. Total, intra- and intermolecular energies per atom and forces, obtained with ensembles of three [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Intermolecular force error norms averaged over configurations taken from the 3EC:7EMC valida [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: a) Ionic pair types; b) Ionic pair composition of training set; c) number of ligands in Li [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Temperature dependency of ionic conductivity predicted by MTP [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: a) Interconnected SSIPs (Network) found in MD produced by MTP20, with only EC molecules being shown; b) distribution of number of Li+ ions involved into Networks, encountered in MD pro￾duced by MTP20. monly bridged separate SSIPs into the network (Fig￾ure 7 (a)), coord…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

59 extracted references · 4 linked inside Pith

  1. [1]

    Boosting rechargeable batteries R&D by multiscale modeling: myth or reality?

    Alejandro A Franco et al. “Boosting rechargeable batteries R&D by multiscale modeling: myth or reality?” In: Chemical reviews 119.7 (2019), pp. 4569–4627

    2019

  2. [2]

    CALiSol-23: Experi- mental electrolyte conductivity data for var- ious Li-salts and solvent combinations

    Paolo de Blasio et al. “CALiSol-23: Experi- mental electrolyte conductivity data for var- ious Li-salts and solvent combinations”. In: Scientific Data 11.1 (2024), p. 750

    2024

  3. [3]

    Ab initio simulations of liquid electrolytes for energy conversion and storage

    Tuan Anh Pham. “Ab initio simulations of liquid electrolytes for energy conversion and storage”. In: International Journal of Quan- tum Chemistry 119.1 (2019), e25795

    2019

  4. [4]

    Lithium ion sol- vation and diffusion in bulk organic elec- trolytes from first-principles and classical re- active molecular dynamics

    Mitchell T Ong et al. “Lithium ion sol- vation and diffusion in bulk organic elec- trolytes from first-principles and classical re- active molecular dynamics”. In: The Jour- nal of Physical Chemistry B 119.4 (2015), pp. 1535–1545

    2015

  5. [5]

    The solvation struc- ture, transport properties and reduction be- havior of carbonate-based electrolytes of lithium-ion batteries

    Tingzheng Hou et al. “The solvation struc- ture, transport properties and reduction be- havior of carbonate-based electrolytes of lithium-ion batteries”. In: Chem. Sci. 12 (44 2021), pp. 14740–14751

    2021

  6. [6]

    Develop- ment of many- body polarizable force fields for Li-battery components: 1. Ether, Alkane, and carbonate-based solvents

    Oleg Borodin and Grant D Smith. “Develop- ment of many- body polarizable force fields for Li-battery components: 1. Ether, Alkane, and carbonate-based solvents”. In: The Jour- nal of Physical Chemistry B 110.12 (2006), pp. 6279–6292

    2006

  7. [7]

    Quan- tum chemistry and molecular dynamics sim- ulation study of dimethyl carbonate: ethylene carbonate electrolytes doped with LiPF6

    Oleg Borodin and Grant D Smith. “Quan- tum chemistry and molecular dynamics sim- ulation study of dimethyl carbonate: ethylene carbonate electrolytes doped with LiPF6”. In: The Journal of Physical Chemistry B 113.6 (2009), pp. 1763–1776

    2009

  8. [8]

    A foundation model for atomistic materials chemistry

    Ilyes Batatia et al. “A foundation model for atomistic materials chemistry”. In: arXiv preprint arXiv:2401.00096 (2023)

    arXiv 2023

Show all 59 references
  1. [9]

    Application of pretrained universal machine-learning interatomic po- tential for physicochemical simulation of liq- uid electrolytes in Li-ion battery

    Suyeon Ju et al. “Application of pretrained universal machine-learning interatomic po- tential for physicochemical simulation of liq- uid electrolytes in Li-ion battery”. In: arXiv preprint arXiv:2501.05211 (2025)

    2025 arXiv

  2. [10]

    High-dimensional neural network potential for liquid electrolyte simulations

    Steven Dajnowicz et al. “High-dimensional neural network potential for liquid electrolyte simulations”. In: The Journal of Physical Chemistry B 126.33 (2022), pp. 6271–6280. 16

    2022

  3. [11]

    Systematic softening in universal machine learning interatomic po- tentials

    Bowen Deng et al. “Systematic softening in universal machine learning interatomic po- tentials”. In: npj Computational Materials 11.1 (2025), p. 9

    2025

  4. [12]

    An efficient forgetting-aware fine-tuning framework for pretrained univer- sal machine-learning interatomic potentials

    Jisu Kim et al. “An efficient forgetting-aware fine-tuning framework for pretrained univer- sal machine-learning interatomic potentials”. In: arXiv preprint arXiv:2506.15223 (2025)

    2025 arXiv

  5. [13]

    Universal Machine Learning Potentials under Pressure

    Antoine Loew et al. “Universal Machine Learning Potentials under Pressure”. In: arXiv preprint arXiv:2508.17792 (2025)

    2025 arXiv

  6. [14]

    Machine learn- ing force fields for molecular liquids: Ethy- lene Carbonate/Ethyl Methyl Carbonate bi- nary solvent

    Ioan-Bogdan Magd ˘au et al. “Machine learn- ing force fields for molecular liquids: Ethy- lene Carbonate/Ethyl Methyl Carbonate bi- nary solvent”. In: npj Computational Mate- rials 9.1 (2023), p. 146

    2023

  7. [15]

    Transferability of Data Sets between Machine-Learned Inter- atomic Potential Algorithms

    Samuel P Niblett et al. “Transferability of Data Sets between Machine-Learned Inter- atomic Potential Algorithms”. In: Journal of Chemical Theory and Computation (2025)

    2025

  8. [16]

    A fourth-generation high- dimensional neural network potential with accurate electrostatics including non-local charge transfer

    Tsz Wai Ko et al. “A fourth-generation high- dimensional neural network potential with accurate electrostatics including non-local charge transfer”. In: Nature communications 12.1 (2021), p. 398

    2021

  9. [17]

    A deep potential model with long-range electrostatic interactions

    Linfeng Zhang et al. “A deep potential model with long-range electrostatic interactions”. In: The Journal of Chemical Physics 156.12 (2022)

    2022

  10. [18]

    Phys- Net: A neural network for predicting en- ergies, forces, dipole moments, and partial charges

    Oliver T Unke and Markus Meuwly. “Phys- Net: A neural network for predicting en- ergies, forces, dipole moments, and partial charges”. In: Journal of chemical theory and computation 15.6 (2019), pp. 3678–3693

    2019

  11. [19]

    The TensorMol-0.1 model chemistry: a neural network augmented with long-range physics

    Kun Yao et al. “The TensorMol-0.1 model chemistry: a neural network augmented with long-range physics”. In: Chemical science 9.8 (2018), pp. 2261–2269

    2018

  12. [20]

    Accurate fourth- generation machine learning potentials by electrostatic embedding

    Tsz Wai Ko et al. “Accurate fourth- generation machine learning potentials by electrostatic embedding”. In: Journal of Chemical Theory and Computation 19.12 (2023), pp. 3567–3579

    2023

  13. [21]

    Less is more: Sampling chemical space with active learning

    Justin S Smith et al. “Less is more: Sampling chemical space with active learning”. In: The Journal of chemical physics 148.24 (2018)

    2018

  14. [22]

    On-the-fly ac- tive learning of interpretable Bayesian force fields for atomistic rare events

    Jonathan Vandermause et al. “On-the-fly ac- tive learning of interpretable Bayesian force fields for atomistic rare events”. In:npj Com- putational Materials 6.1 (2020), p. 20

    2020

  15. [23]

    Training data selection for accuracy and transferability of interatomic potentials

    David Montes de Oca Zapiain et al. “Training data selection for accuracy and transferability of interatomic potentials”. In: npj Computa- tional Materials 8.1 (2022), p. 189

    2022

  16. [24]

    Active learning of linearly parametrized interatomic potentials

    Evgeny V Podryabinkin and Alexander V Shapeev. “Active learning of linearly parametrized interatomic potentials”. In: Computational Materials Science 140 (2017), pp. 171–180

    2017

  17. [25]

    Accelerating high- throughput searches for new alloys with active learning of interatomic potentials

    Konstantin Gubaev et al. “Accelerating high- throughput searches for new alloys with active learning of interatomic potentials”. In: Computational Materials Science 156 (2019), pp. 148–156

    2019

  18. [26]

    How to find a good submatrix

    Sergei A Goreinov et al. “How to find a good submatrix”. In: Matrix Methods: Theory, Al- gorithms And Applications: Dedicated to the Memory of Gene Golub . World Scientific, 2010, pp. 247–256

    2010

  19. [27]

    The MLIP pack- age: moment tensor potentials with MPI and active learning

    Ivan S Novikov et al. “The MLIP pack- age: moment tensor potentials with MPI and active learning”. In: Machine Learning: Science and Technology 2.2 (Dec. 2020), p. 025002

    2020

  20. [28]

    Thermophysical proper- ties of molten FLiNaK: A moment tensor po- tential approach

    Nikita Rybin et al. “Thermophysical proper- ties of molten FLiNaK: A moment tensor po- tential approach”. In: Journal of Molecular Liquids 410 (2024), p. 125402

    2024

  21. [29]

    Ring polymer molecu- lar dynamics and active learning of moment tensor potential for gas-phase barrierless re- actions: Application to S+ H2

    Ivan S Novikov, Alexander V Shapeev, and Yury V Suleimanov. “Ring polymer molecu- lar dynamics and active learning of moment tensor potential for gas-phase barrierless re- actions: Application to S+ H2”. In: The Jour- nal of chemical physics 151.22 (2019)

    2019

  22. [30]

    Accelerating structure pre- diction of molecular crystals using ac- tively trained moment tensor potential

    Nikita Rybin, Ivan S Novikov, and Alexan- der Shapeev. “Accelerating structure pre- diction of molecular crystals using ac- tively trained moment tensor potential”. In: Physical Chemistry Chemical Physics 27.10 (2025), pp. 5141–5148

    2025

  23. [31]

    Actively trained magnetic moment tensor potentials for me- chanical, dynamical, and thermal properties of paramagnetic CrN

    Alexey S Kotykhov et al. “Actively trained magnetic moment tensor potentials for me- chanical, dynamical, and thermal properties of paramagnetic CrN”. In: Physical Review B 111.9 (2025), p. 094438. 17

    2025

  24. [32]

    Bayesian infer- ence of composition-dependent phase dia- grams

    Timofei Miryashkin et al. “Bayesian infer- ence of composition-dependent phase dia- grams”. In: Physical Review B108.17 (2023), p. 174103

    2023

  25. [33]

    Moment Tensor Poten- tial and Equivariant Tensor Network Poten- tial with explicit dispersion interactions

    Olga Chalykh et al. “Moment Tensor Poten- tial and Equivariant Tensor Network Poten- tial with explicit dispersion interactions”. In: arXiv preprint arXiv:2504.15760 (2025)

    2025

  26. [34]

    Accelerat- ing crystal structure prediction by machine- learning interatomic potentials with active learning

    Evgeny V Podryabinkin et al. “Accelerat- ing crystal structure prediction by machine- learning interatomic potentials with active learning”. In: Physical Review B 99.6 (2019), p. 064114

    2019

  27. [35]

    Accelerating the global search of adsorbate molecule positions using machine-learning interatomic potentials with active learning

    Olga Klimanova, Nikita Rybin, and Alexan- der Shapeev. “Accelerating the global search of adsorbate molecule positions using machine-learning interatomic potentials with active learning”. In: Physical Chemistry Chemical Physics 27.17 (2025), pp. 9201– 9210

    2025

  28. [36]

    Moment Tensor Potentials: A Class of Systematically Im- provable Interatomic Potentials

    Alexander V . Shapeev. “Moment Tensor Potentials: A Class of Systematically Im- provable Interatomic Potentials”. In: Multi- scale Modeling & Simulation 14.3 (2016), pp. 1153–1173

    2016

  29. [37]

    The MLIP package: moment tensor potentials with MPI and ac- tive learning

    Ivan S Novikov et al. “The MLIP package: moment tensor potentials with MPI and ac- tive learning”. In: Machine Learning: Sci- ence and Technology 2.2 (2020), p. 025002

    2020

  30. [38]

    Incorporating Coulomb interactions with fixed charges in Moment Tensor Potentials and Equivariant Tensor Network Potentials

    Dmitry Korogod et al. Incorporating Coulomb interactions with fixed charges in Moment Tensor Potentials and Equivariant Tensor Network Potentials. 2025

    2025

  31. [39]

    Alexander Shapeev. MLIP-4. 2024

    2024

  32. [40]

    Performance and cost as- sessment of machine learning interatomic po- tentials

    Yunxing Zuo et al. “Performance and cost as- sessment of machine learning interatomic po- tentials”. In: The Journal of Physical Chem- istry A 124.4 (2020), pp. 731–745

    2020

  33. [41]

    Towards reliable cal- culations of thermal rate constants: Ring polymer molecular dynamics for the OH+ HBr→ Br+ H2O reaction

    Ivan S Novikov et al. “Towards reliable cal- culations of thermal rate constants: Ring polymer molecular dynamics for the OH+ HBr→ Br+ H2O reaction”. In: Chemical Physics Letters 856 (2024), p. 141620

    2024

  34. [42]

    The properties of ethylene carbonate and its use in electrochemical ap- plications a literature review

    Paul H Johnson. “The properties of ethylene carbonate and its use in electrochemical ap- plications a literature review.” In: (1985)

    1985

  35. [43]

    MLIP-3: Active learning on atomic environments with mo- ment tensor potentials

    Evgeny Podryabinkin et al. “MLIP-3: Active learning on atomic environments with mo- ment tensor potentials”. In: The Journal of Chemical Physics 159.8 (2023)

    2023

  36. [44]

    SolvationAnaly- sis: A Python toolkit for understanding liq- uid solvation structure in classical molecular dynamics simulations

    Orion Archer Cohen et al. “SolvationAnaly- sis: A Python toolkit for understanding liq- uid solvation structure in classical molecular dynamics simulations”. In: Journal of Open Source Software 8.84 (2023), p. 5183

    2023

  37. [45]

    MDAnalysis: a Python package for the rapid analysis of molecular dynamics simulations

    Richard J Gowers et al. MDAnalysis: a Python package for the rapid analysis of molecular dynamics simulations . Tech. rep. Los Alamos National Laboratory (LANL), Los Alamos, NM (United States), 2019

    2019

  38. [46]

    Change of conductivity with salt content, solvent composition, and tem- perature for electrolytes of LiPF6 in ethylene carbonate-ethyl methyl carbonate

    MS Ding et al. “Change of conductivity with salt content, solvent composition, and tem- perature for electrolytes of LiPF6 in ethylene carbonate-ethyl methyl carbonate”. In: Jour- nal of the Electrochemical Society 148.10 (2001), A1196

    2001

  39. [47]

    Effect of salt concentration on properties of mixed carbonate-based electrolyte for Li-ion batter- ies: a molecular dynamics simulation study

    Hasty Haghkhah, Behnam Ghalami Choo- bar, and Sepideh Amjad-Iranagh. “Effect of salt concentration on properties of mixed carbonate-based electrolyte for Li-ion batter- ies: a molecular dynamics simulation study”. In: Journal of Molecular Modeling 26.8 (2020), p. 220

    2020

  40. [48]

    Structure of the Li+ ion close environment in various solvents

    PR Smirnov. “Structure of the Li+ ion close environment in various solvents”. In:Russian Journal of General Chemistry 89.12 (2019), pp. 2443–2452

    2019

  41. [49]

    Communica- tion—microscopic view of the ethylene carbonate based lithium-ion battery elec- trolyte by x-ray scattering

    Zhange Feng et al. “Communica- tion—microscopic view of the ethylene carbonate based lithium-ion battery elec- trolyte by x-ray scattering”. In: Journal of the Electrochemical Society 166.2 (2019), A47

    2019

  42. [50]

    Transport phe- nomena in low temperature lithium-ion bat- tery electrolytes

    Alexandra J Ringsby et al. “Transport phe- nomena in low temperature lithium-ion bat- tery electrolytes”. In: Journal of The Electro- chemical Society 168.8 (2021), p. 080501

    2021

  43. [51]

    Enhanced Ion Solvation and Conductivity in Lithium-Ion Electrolytes via Tailored EMC-TMS Solvent Mixtures: A Molecular Dynamics Study

    Jitti Kasemchainan, Siriporn Teerabu- ranapong, and Manaswee Suttipong. “Enhanced Ion Solvation and Conductivity in Lithium-Ion Electrolytes via Tailored EMC-TMS Solvent Mixtures: A Molecular Dynamics Study”. In: ACS omega 10.2 (2025), pp. 2141–2149. 18

    2025

  44. [52]

    Software V ASP, vienna (1999)

    Georg Kresse and J Furthmüller. “Software V ASP, vienna (1999)”. In:Phys. Rev. B54.11 (1996), p. 169

    1999

  45. [53]

    Generalized gradient approxima- tion made simple

    John P Perdew, Kieron Burke, and Matthias Ernzerhof. “Generalized gradient approxima- tion made simple”. In:Physical review letters 77.18 (1996), p. 3865

    1996

  46. [54]

    A consistent and accu- rate ab initio parametrization of density func- tional dispersion correction (DFT-D) for the 94 elements H-Pu

    Stefan Grimme et al. “A consistent and accu- rate ab initio parametrization of density func- tional dispersion correction (DFT-D) for the 94 elements H-Pu”. In: The Journal of chem- ical physics 132.15 (2010)

    2010

  47. [55]

    Markoff Random Pro- cesses and the Statistical Mechanics of Time- Dependent Phenomena. II. Irreversible Pro- cesses in Fluids

    Melville S. Green. “Markoff Random Pro- cesses and the Statistical Mechanics of Time- Dependent Phenomena. II. Irreversible Pro- cesses in Fluids”. In: The Journal of Chemi- cal Physics 22.3 (1954), pp. 398–413

    1954

  48. [56]

    Statistical-Mechanical The- ory of Irreversible Processes. I. General The- ory and Simple Applications to Magnetic and Conduction Problems

    Ryogo Kubo. “Statistical-Mechanical The- ory of Irreversible Processes. I. General The- ory and Simple Applications to Magnetic and Conduction Problems”. In: Journal of the Physical Society of Japan 12.6 (1957), pp. 570–586

    1957

  49. [57]

    Esti- mates of Electrical Conductivity from Molec- ular Dynamics Simulations: How to Invest the Computational Effort

    Piotr Kubisiak and Andrzej Eilmes. “Esti- mates of Electrical Conductivity from Molec- ular Dynamics Simulations: How to Invest the Computational Effort”. In: The Jour- nal of Physical Chemistry B 124.43 (2020), pp. 9680–9689

    2020

  50. [58]

    Estimating ionic con- ductivity of ionic liquids: Nernst–Einstein and Einstein formalisms

    Ashutosh Kumar Verma, Amey S. Thorat, and Jindal K. Shah. “Estimating ionic con- ductivity of ionic liquids: Nernst–Einstein and Einstein formalisms”. In: Journal of Ionic Liquids 4.1 (2024), p. 100089. ISSN : 2772-4220

    2024

  51. [59]

    Best practices for computing transport properties 1. Self- diffusivity and viscosity from equilibrium molecular dynamics [article v1. 0]

    Edward J Maginn et al. “Best practices for computing transport properties 1. Self- diffusivity and viscosity from equilibrium molecular dynamics [article v1. 0]”. In: Liv- ing Journal of Computational Molecular Sci- ence 1.1 (2019), pp. 6324–6324. 19 Supplementary Information ...

    2019

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.