Pith. sign in

REVIEW 4 major objections 4 minor 58 references

How Machine Learning Predicts Fluid Densities under Nanoconfinement

T0 review · 4 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read A random forest model trained on molecular simulation data can predict the layered density profiles of fluids in nanoscale slit pores, replacing fresh simulations for new state points.

desk verdict Plausible ML-surrogate application for confined-fluid density profiles, but the supplied text is corrupted and the 'excellent interpolation' claim cannot be verified without state-point-level validation; worth a referee if the clean PDF has it. read the letter →

arxiv 2508.17732 v1 pith:CJOD2W4Z submitted 2025-08-25 cond-mat.soft

Yuanhao Li This is my paper
classification cond-mat.soft
keywords machinelearningrandomforestnanoconfinedfluidsdensityprofilesfluidlayeringmolecularsimulationsurrogateslitporesLennard-Jones
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

Nanoconfined fluids form layered density profiles that are notoriously hard to predict analytically, so researchers typically rely on molecular simulation for each new condition. This paper claims that a random forest regression model, trained on a large set of simulation-derived density profiles for a Lennard-Jones fluid in slit pores, can replace those simulations: given a temperature, pore width, and position, it returns the local density with accuracy that matches simulation for conditions inside the training range, and degrades gracefully just outside it. The practical upside is that once trained, the model costs almost nothing per prediction, so practitioners can scan large parameter spaces of pore geometries and temperatures in seconds rather than CPU-days. The paper positions this as a step toward generalizable, lower-cost models of nanoconfined fluid structure.

What carries the argument

The central object is a random forest regressor—an ensemble of decision trees—that learns the mapping (T, H, z) → ρ(z), where T is temperature, H is slit-pore width, and z is position across the pore. The forest is trained on density profiles extracted from molecular simulation (the data generation engine), and its trees partition the input space into piecewise-constant regions, effectively creating a nonparametric interpolation table. The model's role is to act as a surrogate: to reproduce the simulation's output for arbitrary state points without any new simulation, with the tree ensemble providing the interpolation mechanism.

What would settle it

Evaluate the trained model at a pore width and temperature that lie between training points but across a known layering transition (e.g., the width where the number of fluid layers increases by one); if the random forest prediction shows a smooth blend of the two regimes rather than one or the other, the smooth-interpolation premise is falsified.

Watch

Extended reading notes

Core claim

The paper demonstrates that a random forest regressor, taking as input the temperature, the pore width, and the position along the confined dimension, can accurately reconstruct the full inhomogeneous density profile of a Lennard-Jones fluid confined in a slit pore. Trained on density profiles precomputed from molecular simulation across a grid of state points, the model interpolates smoothly between those points, reproducing the layering peaks and troughs with high fidelity. It also retains 'modest extrapolative ability' when tested outside the training range of temperatures and widths, suggesting that what the tree ensemble has learned is not pure memorization but a representation of how l

Load-bearing premise

The training grid of temperatures and pore widths is dense enough that the true density profile changes smoothly between grid points; if a layering transition falls between two trained points, the model will interpolate across physics it has never seen.

Editorial extensions

If this is right

  • Within the trained temperature–width envelope, any arbitrary pore size and temperature can be queried instantly, enabling quick scans of confinement effects on fluid structure.
  • The same surrogate strategy should transfer to other confined-fluid observables that are smooth functions of state variables, such as solvation forces or diffusion coefficients.
  • Because the model shows some extrapolative ability, it may capture simple scaling laws of layering with pore width, which could inform physics-based closures.
  • A random forest baseline at this accuracy sets a benchmark any more complex neural-network surrogate must beat.
  • The trained model can be embedded in multiscale workflows that need fast estimates of local densities, e.g., in coarse-grained simulations or pore-network models.

Reading between the lines

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

  • The success of a piecewise-constant interpolator suggests the confined density profile is an exceptionally smooth function of (T, H, z) away from layer transitions; if so, classical density functional theory or even simple spline interpolation could achieve similar accuracy, and the paper's contribution is largely an efficient, nonparametric way to store simulation-generated profiles.
  • The main untested hazard is the possibility of abrupt layering transitions (e.g., formation of a new molecular layer at a critical width) falling between training grid points; the model would interpolate across the transition, producing a smooth mixture of two regimes that corresponds to no real fluid state. A targeted test at transition boundaries would define the real interpolation envelope.
  • The paper's demonstration uses a model fluid (Lennard-Jones in a slit); extending to chemically heterogeneous pores, where wall-fluid interactions vary and the profile is no longer a single smooth function, will likely require additional input features or much larger data sets.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The manuscript claims that a random forest trained on large molecular simulation datasets can serve as a reliable surrogate for predicting inhomogeneous density profiles of Lennard-Jones fluids in slit nanopores. The abstract asserts 'excellent interpolative capabilities over a wide range of temperatures and confining lengthscales' and 'modest extrapolative ability' relative to direct simulation. The legible fragments of the manuscript contain hyperparameter-like tables and parameter lists, but the abstract and these fragments do not report quantitative error metrics, dataset size, the state-point grid, or the train/test protocol. The central claim is thus currently asserted rather than demonstrated in the material available for review.

Significance. If the claims were fully substantiated, the paper would offer a practical, low-cost surrogate for nanoconfined fluid structure, bypassing repeated molecular simulation for practitioners. The choice of a random forest is a reasonable, interpretable baseline, and the motivation—overcoming the lack of generalizable analytical theories for fluid layering—is well founded. However, the significance is conditional on out-of-sample validation: a surrogate trained on simulation output is only useful if it predicts accurately at state points not in the training set. The current manuscript does not establish that condition, and the piecewise-constant nature of random forests makes the interpolation and extrapolation claims especially sensitive to the validation design.

major comments (4)
  1. [Abstract / Results (train-test protocol)] The central claim of 'excellent interpolative capabilities' requires predictions at state points (T,H) excluded from training. The abstract and legible results give no indication of how the train/test split was made. For a random forest, a random split of configurations drawn from the same state points is not an interpolation test: the model can memorize each profile, and the reported performance reduces to leaf averaging rather than physical prediction. Please report leave-one-state-point-out or leave-one-width-out cross-validation, including the number of held-out state points and per-state-point errors.
  2. [Abstract (performance metrics)] The terms 'excellent' and 'modest' are not quantified. The paper should report RMSE, MAE, maximum local error, and R² relative to the simulation reference, ideally as functions of T and H. It should also specify the dataset size, the number of state points, and the grid spacings ΔT and ΔH. Without these, the performance claims are not falsifiable and the surrogate cannot be compared with other approaches.
  3. [Abstract / Extrapolation claim] Random forests are piecewise-constant predictors; outside the training range they output leaf averages or boundary constants. The claimed 'modest extrapolative ability' must be accompanied by the exact extrapolation protocol: which variable was extrapolated (T or H), how far beyond the training envelope, and the comparison simulation data at those extrapolated points. If the evaluation was performed only inside the convex hull of the training state points, the term 'extrapolation' is misleading.
  4. [Interpolation across H (layer transitions)] The interpolation claim presumes the density profile is a smooth, single-valued function of T, H, and position z. Nanoconfined fluids undergo qualitative changes in layer count as the pore width crosses threshold values; this layering transition is not a smooth function of H. If the training grid in H is not fine enough to resolve these transitions, the random forest will silently interpolate across distinct physical regimes. The authors should demonstrate that the H-grid resolves layer insertion/removal and that no phase-transition-like discontinuity lies between trained state points.
minor comments (4)
  1. [Full text / General] The supplied manuscript text is heavily corrupted or encoded; equations and Methods sections could not be verified. The authors should ensure the posted version is a clean, readable PDF.
  2. [Tables / Figures] Table and figure captions are missing or illegible in the supplied text. Captions should identify which entries are hyperparameters, which are simulation parameters, and which are performance metrics. Figures need axis labels, units, and error bars reflecting the statistical uncertainty of the MD reference.
  3. [Text artifact] A companion identifier 'arXiv:2508.17733 [astro-ph.GA]' appears within the body text. This appears to be an artifact and should be removed unless the authors intend a cross-reference.
  4. [Title / Abstract consistency] The title refers to 'fluid densities' while the abstract focuses on inhomogeneous density profiles. Please clarify whether the target is the full profile ρ(z) or an integrated/average density, and keep the terminology consistent throughout.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the ML surrogate claim is an empirical interpolation claim validated against independent simulation data, not a derivation that reduces to its inputs.

full rationale

The paper is an empirical surrogate-modeling study. Its central claim is that a random forest trained on large molecular simulation data sets can reproduce nanoconfined fluid density profiles at nearby state points and show modest extrapolative ability. This is not a first-principles derivation, so the circularity patterns enumerated (self-definition, fitted-input-called-prediction, self-citation chains, uniqueness imported from authors, ansatz smuggled via citation, renaming a known result) do not apply. The target quantity—the density profile—is generated by independent molecular simulations, and the random forest is an approximant fitted to those data. In-range agreement is a supervised-learning interpolation claim; it would reduce to a fit only if the validation set were not held out, but the garbled text provides no quotable evidence that the paper conflates training accuracy with prediction. The abstract explicitly separates the modest extrapolation claim from the interpolation claim, which is the honest structure for a surrogate model. No load-bearing self-citation or definitional identity was found, so the appropriate finding is no significant circularity.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The ledger is nearly empty of invented physics, which is good: the paper postulates no new entities or forces, and its free parameters are ML configuration choices rather than physical constants. The central burden sits in two places: the surrogate inherits the accuracy of the Lennard-Jones simulation data treated as ground truth, and the interpolation claim rests on an unstated smoothness assumption about how layering varies with temperature and pore width. No numbers could be extracted from the corrupted full text to populate the ledger further.

free parameters (2)
  • Random forest hyperparameters = not disclosed in abstract
    Number of trees, depth, minimum leaf size, and feature set determine the surrogate's accuracy; the abstract does not report them, and the corrupted full text hides the methods section.
  • Training state-point grid (T*, H*) = not disclosed in abstract
    The claimed interpolation range is defined by which temperatures and pore widths were simulated; the abstract does not state the grid density, which is the premise for excellent interpolation.
assumptions (3)
  • domain assumption Molecular simulation (Lennard-Jones slit-pore) data is an accurate ground truth for confined-fluid density profiles.
    The surrogate is only as good as its training data; any force-field or finite-size error in the simulations is inherited by the predictions. Invoked by the abstract's framing of the training set as large molecular simulation data sets.
  • domain assumption Local density at position z is a smooth, deterministic function of temperature and pore width, learnable by a tree ensemble from discrete state points.
    Excellent interpolation across discrete training state points presupposes the profile varies smoothly with T* and H*; if layering changes abruptly with pore width, a coarse grid misleads the surrogate.
  • domain assumption Held-out evaluation sets are statistically independent of training state points.
    The interpolation/extrapolation distinction in the abstract only means something if the reported performance was measured on state points excluded from training; no protocol is given in the abstract.

how reviews work

0 comments
Cite this review

Pith. "Pith review of How Machine Learning Predicts Fluid Densities under Nanoconfinement." pith.science (2026). https://pith.science/paper/CJOD2W4Z

@misc{pith2026250817732,
  author       = {Pith},
  title        = {Pith review of: How Machine Learning Predicts Fluid Densities under Nanoconfinement},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CJOD2W4Z}},
  note         = {Machine review of arXiv:2508.17732}
}
read the original abstract

Fluids under nanoscale confinement differ -- and often dramatically -- from their bulk counterparts. A notorious feature of nanoconfined fluids is their inhomogeneous density profile along the confining dimension, which plays a key role in many fluid structural and transport phenomena in nanopores. Nearly five decades of theoretical efforts on predicting this phenomenon (fluid layering) have shown that its complexity resists purely analytical treatments; as a consequence, nearly all current approaches make extensive use of molecular simulations, and tend not to have generalizable predictive capabilities. In this work, we demonstrate that machine-learning-based models (in particular, a random forest model), trained upon large molecular simulation data sets, can serve as reliable surrogates in lieu of further molecular simulation. We show that this random forest model has excellent interpolative capabilities over a wide range of temperatures and confining lengthscales, and even has modest extrapolative ability. These results provide a promising pathway forward for developing models of nanoconfined fluid properties that are generalizable, lower cost than ``pure" molecular simulation, and sufficiently predictive for fluids-in-nanopores practitioners.}

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

58 extracted references · 56 canonical work pages

  1. [1]

    title Nanofluidics coming of age

    author Bocquet, L. title Nanofluidics coming of age . journal Nature Materials volume 19 , pages 254--256 ( year 2020 )

    work page 2020

  2. [2]

    , author Zhou, R

    author Sun, C. , author Zhou, R. , author Zhao, Z. & author Bai, B. title Nanoconfined Fluids: What Can We Expect from Them? journal The Journal of Physical Chemistry Letters volume 11 , pages 4678--4692 ( year 2020 )

    work page 2020

  3. [3]

    author Krekelberg, W. P. , author Siderius, D. W. , author Shen, V. K. , author Truskett, T. M. & author Errington, J. R. title Connection between thermodynamics and dynamics of simple fluids in highly attractive pores . journal Langmuir volume 29 , pages 14527--14535 ( year 2013 ). arXiv: http://dx.doi.org/10.1021/la4037327

    work page doi:10.1021/la4037327 2013

  4. [4]

    , author Truskett, T

    author Mittal, J. , author Truskett, T. M. , author Errington, J. R. & author Hummer, G. title Layering and position-dependent diffusive dynamics of confined fluids . journal Phys. Rev. Lett. volume 100 , pages 145901 ( year 2008 )

    work page 2008

  5. [5]

    & author Aluru, N

    author Barati Farimani, A. & author Aluru, N. R. title Spatial diffusion of water in carbon nanotubes: From fickian to ballistic motion . journal J. Phys. Chem. B volume 115 , pages 12145--12149 ( year 2011 ). arXiv: http://dx.doi.org/10.1021/jp205877b

    work page doi:10.1021/jp205877b 2011

  6. [6]

    author Wang, G. J. & author Hadjiconstantinou, N. G. title Layered fluid structure and anomalous diffusion under nanoconfinement . journal Langmuir volume 34 , pages 6976--6982 ( year 2018 ). note PMID: 29775320 , arXiv: https://doi.org/10.1021/acs.langmuir.8b01540

    work page doi:10.1021/acs.langmuir.8b01540 2018

  7. [7]

    & author Wang, G

    author Li, Y. & author Wang, G. J. title How to produce confidence intervals instead of confidence tricks: Representative sampling for molecular simulations of fluid self-diffusion under nanoscale confinement . journal The Journal of Chemical Physics volume 156 , pages 114113 ( year 2022 )

    work page 2022

  8. [8]

    , author Peeters, F

    author Neek-Amal, M. , author Peeters, F. M. , author Grigorieva, I. V. & author Geim, A. K. title Commensurability Effects in Viscosity of Nanoconfined Water . journal ACS Nano volume 10 , pages 3685--3692 ( year 2016 )

    work page 2016

Show all 58 references
  1. [9]

    , author Chen, J

    author Alexeev, D. , author Chen, J. , author Walther, J. H. , author Giapis, K. P. , author Angelikopoulos, P. & author Koumoutsakos, P. title Kapitza resistance between few-layer graphene and water: Liquid layering effects . journal Nano Lett. volume 15 , pages 0.5mm 5744--5...

    2015

  2. [10]

    , author Goicochea, J

    author Hu, M. , author Goicochea, J. V. , author Michel, B. & author Poulikakos, D. title Water Nanoconfinement Induced Thermal Enhancement at Hydrophilic Quartz Interfaces . journal Nano Letters volume 10 , pages 279--285 ( year 2010 )

    2010

  3. [11]

    , author Kanke, Y

    author Nagoe, A. , author Kanke, Y. , author Oguni, M. & author Namba, S. title Findings of Cp Maximum at 233 K for the Water within Silica Nanopores and Very Weak Dependence of the Tmax on the Pore Size . journal The Journal of Physical Chemistry B volume 114 , pages 13940--1...

    2010

  4. [12]

    author Kim, B. H. , author Beskok, A. & author Cagin, T. title Thermal interactions in nanoscale fluid flow: molecular dynamics simulations with solid--liquid interfaces . journal Microfluidics and Nanofluidics volume 5 , pages 551--559 ( year 2008 )

    2008

  5. [13]

    , author Shen, Y

    author Shi, K. , author Shen, Y. , author Santiso, E. E. & author Gubbins, K. E. title Microscopic pressure tensor in cylindrical geometry: Pressure of water in a carbon nanotube . journal Journal of Chemical Theory and Computation volume 16 , pages 5548--5561 ( year 2020 ). n...

    2020 doi

  6. [14]

    & author Maroo, S

    author Zou, A. & author Maroo, S. C. title Nano-confinement effects on liquid pressure ( year 2020 ). arXiv: 2012.02265

    2020 arXiv

  7. [15]

    , author Tanaka, H

    author Koga, K. , author Tanaka, H. & author Zeng, X. C. title First-order transition in confined water between high-density liquid and low-density amorphous phases . journal Nature volume 408 , pages 564--567 ( year 2000 )

    2000

  8. [16]

    , author Zeng, X

    author Koga, K. , author Zeng, X. C. & author Tanaka, H. title Freezing of confined water: A bilayer ice phase in hydrophobic nanopores . journal Phys. Rev. Lett. volume 79 , pages 5262--5265 ( year 1997 )

    1997

  9. [17]

    , author Hatano, I

    author Takaiwa, D. , author Hatano, I. , author Koga, K. & author Tanaka, H. title Phase diagram of water in carbon nanotubes . journal Proceedings of the National Academy of Sciences volume 105 , pages 39--43 ( year 2008 ). arXiv: https://www.pnas.org/content/105/1/39.full.pdf

    2008

  10. [18]

    & author Aluru, N

    author Barati Farimani , A. & author Aluru, N. R. title Existence of Multiple Phases of Water at Nanotube Interfaces . journal The Journal of Physical Chemistry C volume 120 , pages 23763--23771 ( year 2016 )

    2016

  11. [19]

    author Agrawal, K. V. , author Shimizu, S. , author Drahushuk, L. W. , author Kilcoyne, D. & author Strano, M. S. title Observation of extreme phase transition temperatures of water confined inside isolated carbon nanotubes . journal Nature Nanotechnology volume 12 , pages 267...

    2017

  12. [20]

    , author Lichter, S

    author Han, M. , author Lichter, S. , author Lin, C. & author Perng, Y. title Micromechanics of the moving contact line . journal NASA Conference Publication pages 621--626 ( year 1996 ). note Proceedings of the 1996 3rd Microgravity Fluid Physics Conference ; Conference date:...

    1996

  13. [21]

    , author Horinek, D

    author Sendner, C. , author Horinek, D. , author Bocquet, L. & author Netz, R. R. title Interfacial water at hydrophobic and hydrophilic surfaces: Slip, viscosity, and diffusion . journal Langmuir volume 25 , pages 10768--10781 ( year 2009 )

    2009

  14. [22]

    & author Barrat, J.-L

    author Bocquet, L. & author Barrat, J.-L. title Hydrodynamic boundary conditions, correlation functions, and kubo relations for confined fluids . journal Phys. Rev. E volume 49 , pages 3079--3092 ( year 1994 )

    1994

  15. [23]

    & author Bocquet, L

    author Barrat, J.-L. & author Bocquet, L. title Large slip effect at a nonwetting fluid-solid interface . journal Phys. Rev. Lett. volume 82 , pages 4671--4674 ( year 1999 )

    1999

  16. [24]

    & author Barrat, J.-L

    author Bocquet, L. & author Barrat, J.-L. title Flow boundary conditions from nano- to micro-scales . journal Soft Matter volume 3 , pages 685--693 ( year 2007 )

    2007

  17. [25]

    , author Roxin, A

    author Lichter, S. , author Roxin, A. & author Mandre, S. title Mechanisms for liquid slip at solid surfaces . journal Phys. Rev. Lett. volume 93 , pages 086001 ( year 2004 )

    2004

  18. [26]

    , author Roxin, A

    author Martini, A. , author Roxin, A. , author Snurr, R. Q. , author Wang, Q. & author Lichter, S. title Molecular mechanisms of liquid slip . journal Journal of Fluid Mechanics volume 600 , pages 257--269 ( year 2008 )

    2008

  19. [27]

    , author Martini, A

    author Lichter, S. , author Martini, A. , author Snurr, R. Q. & author Wang, Q. title Liquid slip in nanoscale channels as a rate process . journal Phys. Rev. Lett. volume 98 , pages 226001 ( year 2007 )

    2007

  20. [28]

    , author Nicholson, D

    author Rowley, L. , author Nicholson, D. & author Parsonage, N. title Grand ensemble monte-carlo studies of physical adsorption . journal Molecular Physics volume 31 , pages 389--407 ( year 1976 ). arXiv: http://dx.doi.org/10.1080/00268977600100291

    1976 doi

  21. [29]

    author Abraham, F. F. title The interfacial density profile of a lennard-jones fluid in contact with a (100) lennard-jones wall and its relationship to idealized fluid/wall systems: A monte carlo simulation . journal The Journal of Chemical Physics volume 68 , pages 3713--3716...

    1978 doi

  22. [30]

    & author Praestgaard, E

    author Toxvaerd, S. & author Praestgaard, E. title Molecular dynamics calculation of the liquid structure up to a solid surface . journal The Journal of Chemical Physics volume 67 , pages 5291--5295 ( year 1977 ). arXiv: https://doi.org/10.1063/1.434707

    1977 doi

  23. [31]

    title The structure and thermodynamics of a solid-fluid interface

    author Toxvaerd, S. title The structure and thermodynamics of a solid-fluid interface . journal The Journal of Chemical Physics volume 74 , pages 1998--2005 ( year 1981 ). arXiv: https://doi.org/10.1063/1.441243

    1998 doi

  24. [32]

    , author Rice, S

    author He, Y. , author Rice, S. A. & author Xu, X. title Analytic solution of the ornstein-zernike relation for inhomogeneous liquids . journal The Journal of Chemical Physics volume 145 , pages 234508 ( year 2016 ). arXiv: https://doi.org/10.1063/1.4972020

    2016 doi

  25. [33]

    title Free-energy model for the inhomogeneous hard-sphere fluid mixture and density-functional theory of freezing

    author Rosenfeld, Y. title Free-energy model for the inhomogeneous hard-sphere fluid mixture and density-functional theory of freezing . journal Phys. Rev. Lett. volume 63 , pages 980--983 ( year 1989 )

    1989

  26. [34]

    author Stillinger, F. H. & author Buff, F. P. title Equilibrium statistical mechanics of inhomogeneous fluids . journal The Journal of Chemical Physics volume 37 , pages 1--12 ( year 1962 ). arXiv: https://doi.org/10.1063/1.1732927

    1962 doi

  27. [35]

    author Yang, J. H. , author Schultz, A. J. , author Errington, J. R. & author Kofke, D. A. title Calculation of inhomogeneous-fluid cluster expansions with application to the hard-sphere/hard-wall system . journal The Journal of Chemical Physics volume 138 , pages 134706 ( yea...

    2013 doi

  28. [36]

    author Sisan, T. B. & author Lichter, S. title The end of nanochannels . journal Microfluidics and Nanofluidics volume 11 , pages 787--791 ( year 2011 )

    2011

  29. [37]

    author Raghunathan, A. V. , author Park, J. H. & author Aluru, N. R. title Interatomic potential-based semiclassical theory for lennard-jones fluids . journal The Journal of Chemical Physics volume 127 , pages 174701 ( year 2007 ). arXiv: https://doi.org/10.1063/1.2793070

    2007 doi

  30. [38]

    author Wang, G. J. & author Hadjiconstantinou, N. G. title Why are fluid densities so low in carbon nanotubes? journal Phys. Fluids volume 27 ( year 2015 )

    2015

  31. [39]

    author Wang, G. J. & author Hadjiconstantinou, N. G. title Molecular mechanics and structure of the fluid-solid interface in simple fluids . journal Phys. Rev. Fluids volume 2 , pages 094201 ( year 2017 )

    2017

  32. [40]

    & author Zewail, A

    author Yang, D.-S. & author Zewail, A. H. title Ordered water structure at hydrophobic graphite interfaces observed by 4d, ultrafast electron crystallography . journal Proceedings of the National Academy of Sciences volume 106 , pages 4122--4126 ( year 2009 ). arXiv: https://w...

    2009

  33. [41]

    author Uhlig, M. R. , author Martin-Jimenez, D. & author Garcia, R. title Atomic-scale mapping of hydrophobic layers on graphene and few-layer MoS2 and WSe2 in water . journal Nature Communications volume 10 , pages 2606 ( year 2019 )

    2019

  34. [42]

    & author R., M

    author Hansen, J.-P. & author R., M. I. title Theory of Simple Liquids ( publisher Elsevier , year 2006 )

    2006

  35. [43]

    author Myers, T. G. title Why are slip lengths so large in carbon nanotubes? journal Microfluidics and Nanofluidics volume 10 , pages 1141--1145 ( year 2011 )

    2011

  36. [44]

    author Thomas, J. A. & author McGaughey, A. J. H. title Reassessing Fast Water Transport Through Carbon Nanotubes . journal Nano Letters volume 8 , pages 2788--2793 ( year 2008 )

    2008

  37. [45]

    & author Aluru, N

    author Heiranian, M. & author Aluru, N. R. title Nanofluidic transport theory with enhancement factors approaching one . journal ACS Nano volume 14 , pages 272--281 ( year 2020 )

    2020

  38. [46]

    title Fast parallel algorithms for short-range molecular dynamics

    author Plimpton, S. title Fast parallel algorithms for short-range molecular dynamics . journal Journal of computational physics volume 117 , pages 1--19 ( year 1995 )

    1995

  39. [47]

    author Berendsen, H. J. C. , author Grigera, J. R. & author Straatsma, T. P. title The missing term in effective pair potentials . journal J. Phys. Chem. volume 91 , pages 6269--6271 ( year 1987 ). arXiv: http://pubs.acs.org/doi/pdf/10.1021/j100308a038

    1987 doi

  40. [48]

    title A unified formulation of the constant temperature molecular dynamics methods

    author Nos \'e , S. title A unified formulation of the constant temperature molecular dynamics methods . journal The Journal of chemical physics volume 81 , pages 511--519 ( year 1984 )

    1984

  41. [49]

    author Hoover, W. G. title Canonical dynamics: Equilibrium phase-space distributions . journal Physical review A volume 31 , pages 1695 ( year 1985 )

    1985

  42. [50]

    author Pedregosa, F. et al. title Scikit-learn: Machine learning in P ython . journal Journal of Machine Learning Research volume 12 , pages 2825--2830 ( year 2011 )

    2011

  43. [51]

    & author Guestrin, C

    author Chen, T. & author Guestrin, C. title Xgboost: A scalable tree boosting system . In booktitle Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining , pages 785--794 ( year 2016 )

    2016

  44. [52]

    author Paszke, A. et al. title Pytorch: An imperative style, high-performance deep learning library . In booktitle Advances in neural information processing systems , pages 8026--8037 ( year 2019 )

    2019

  45. [53]

    , author Hinton, G

    author Srivastava, N. , author Hinton, G. , author Krizhevsky, A. , author Sutskever, I. & author Salakhutdinov, R. title Dropout: a simple way to prevent neural networks from overfitting . journal The journal of machine learning research volume 15 , pages 1929--1958 ( year 2014 )

    1929

  46. [54]

    author Kingma, D. P. & author Ba, J. title Adam: A method for stochastic optimization . journal arXiv preprint arXiv:1412.6980 ( year 2014 )

    2014 arXiv

  47. [55]

    , author Giovambattista, N

    author Gao, Z. , author Giovambattista, N. & author Sahin, O. title Phase Diagram of Water Confined by Graphene . journal Scientific Reports volume 8 , pages 6228 ( year 2018 )

    2018

  48. [56]

    & author Verlet, L

    author Hansen, J.-P. & author Verlet, L. title Phase transitions of the lennard-jones system . journal Phys. Rev. volume 184 , pages 151--161 ( year 1969 )

    1969

  49. [57]

    , " * write output.state after.block = add.period write newline

    ENTRY address archive archiveprefix author booktitle chapter edition editor eprint howpublished institution journal key month note number organization pages publisher school series title type url volume year label INTEGERS output.state before.all mid.sentence after.sentence af...

  50. [58]

    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 gl...

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.