pith. sign in

arxiv: 2603.11973 · v2 · submitted 2026-03-12 · ❄️ cond-mat.soft

Metadensity functional learning for classical fluids: Regularizing with pair correlations

Stefanie M. Kampa , Florian Samm\"uller , Matthias Schmidt This is my paper

Pith reviewed 2026-05-15 12:06 UTC · model grok-4.3

classification ❄️ cond-mat.soft
keywords metadensity functional theoryclassical density functional theorypair correlationsneural functionalsinhomogeneous fluidsregularizationone-dimensional systemsOrnstein-Zernike inversion
0
0 comments X

The pith

Metadensity functional dependence on the pair potential gives direct access to fluid pair correlations without Ornstein-Zernike inversion.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper examines consequences of neural metadensity functional theory for inhomogeneous classical fluids in one dimension. It uses the built-in dependence of the metadensity on the pair potential to construct metadirect differentiation routes that recover the bulk pair correlation structure. Local neural functional learning is regularized by matching these differentiated results against test-particle insertion data obtained from an initial metadensity functional. A reader would care because the method supplies a first-principles route to pair structure that supports soft-matter design and permits the pair potential to be changed at prediction time.

Core claim

In classical density functional theory the neural metadensity functional's explicit dependence on the pair potential permits metadirect functional differentiation to obtain the pair correlation functions; these are matched to accurate test-particle data from an initial locally trained metadensity functional, thereby regularizing the learning process while circumventing Ornstein-Zernike inversion.

What carries the argument

Metadirect functional differentiation of the metadensity functional with respect to the pair potential, performed via neural functional line integration and automatic differentiation.

If this is right

  • The pair potential can be varied on the fly during prediction without retraining.
  • Local neural functional learning is regularized solely by internal consistency between differentiation routes and test-particle data.
  • Pair correlation functions are recovered from first principles for inhomogeneous fluids.
  • Efficient variational calculus is realized through automatic differentiation for short-ranged one-dimensional systems.
  • Different computational routes to the same pair structure are forced into agreement.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same metadirect route may be tested on exactly solvable one-dimensional models to quantify numerical accuracy.
  • Extension to two or three dimensions would test whether the regularization remains stable for more realistic geometries.
  • The approach could support inverse design loops in which pair potentials are adjusted to achieve target structures without separate inversion steps.
  • It offers a natural interface with simulation data for hybrid neural-functional models of soft matter.

Load-bearing premise

Matching pair structures from metadirect differentiation against test-particle data from the metadensity functional supplies effective regularization without introducing inconsistencies or requiring extra fitting.

What would settle it

A direct numerical comparison in a one-dimensional fluid with short-ranged interactions showing that the pair correlation obtained by metadirect differentiation deviates measurably from independent test-particle or simulation data.

Figures

Figures reproduced from arXiv: 2603.11973 by Florian Samm\"uller, Matthias Schmidt, Stefanie M. Kampa.

Figure 1
Figure 1. Figure 1: FIG. 1. Overview of the two-stage neural functional learning (Sec. III). From left to right: Training data consists of simulation [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Neural functional results for the bulk pair distribution function [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Illustration of the relationship of the bulk metacompressibility and the pair distribution function. a) Scaled bulk [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. Illustration of the metadensity functional depen [PITH_FULL_IMAGE:figures/full_fig_p010_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Demonstration of metadensity functional application in an inhomogeneous system. The particles interact via a [PITH_FULL_IMAGE:figures/full_fig_p011_5.png] view at source ↗
read the original abstract

We investigate and exploit consequences of the recent neural metadensity functional theory [Kampa et al., Phys. Rev. Lett. 134, 107301 (2025), 10.1103/PhysRevLett.134.107301] for describing the physics of inhomogeneous fluids. The metadensity dependence on the pair potential is relevant for soft matter design and Henderson inversion and it allows one to change the pair potential on the fly at prediction stage. Here we consider one-dimensional systems with short-ranged (truncated) interparticle forces and draw on the functional pair potential dependence to investigate 'metadirect' routes towards the bulk fluid pair correlation structure. Classical density functional theory provides the required functional relationships. Efficient variational calculus is implemented by neural functional line integration and automatic differentiation. We regularize local learning of neural functionals by comparing the pair structure from different routes. Thereby results from metadirect functional differentiation are matched against accurate test particle data from an initial locally trained metadensity functional. Accessing the pair structure via the metadensity functional dependence circumvents Ornstein-Zernike inversion and it is based on first principles.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 1 minor

Summary. The manuscript extends neural metadensity functional theory to inhomogeneous classical fluids by exploiting the explicit dependence of the learned functional on the pair potential. It derives bulk pair correlation functions via metadirect functional differentiation (using neural line integration and automatic differentiation) and regularizes the local training of the neural model by matching these structures to test-particle insertion data generated from the same initial neural functional. The approach is demonstrated on one-dimensional systems with short-ranged truncated interactions and is claimed to circumvent Ornstein-Zernike inversion while remaining first-principles based.

Significance. If the regularization procedure can be shown to enforce consistency with exact classical DFT relations (such as sum rules) independently of the shared neural approximation, the work would offer a useful route for learning functionals that remain consistent under changes to the pair potential at inference time. The technical implementation via automatic differentiation for variational calculus is a clear strength for reproducibility and efficiency. However, the dependence on matching outputs from the identical approximate model limits the strength of the first-principles claim and reduces the potential impact relative to methods that incorporate independent constraints or external data.

major comments (3)
  1. [Abstract / regularization step] Abstract and regularization description: matching the metadirect pair structure (obtained by functional differentiation of the neural metadensity) against test-particle data generated from the identical locally trained neural functional creates a circular dependence. Agreement between the two routes can occur even when the underlying neural approximation violates exact DFT identities (e.g., compressibility sum rule or virial theorem), so the procedure does not automatically guarantee consistency when the pair potential is varied at prediction stage.
  2. [Abstract] The claim that the metadirect route 'is based on first principles' (abstract) is weakened because both the metadirect differentiation and the test-particle reference derive from the same data-fitted neural representation rather than from exact functional identities independent of the approximation.
  3. [Results / validation] No explicit validation metrics (error bars, quantitative discrepancy measures, or tests on unseen pair potentials) are described, making it impossible to assess whether the regularization improves predictive accuracy or merely enforces internal consistency within the training distribution.
minor comments (1)
  1. [Abstract] The term 'metadirect' is used in the abstract without a concise definition; a parenthetical clarification linking it to functional differentiation of the metadensity would improve immediate readability.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their careful reading and constructive comments. We address each major comment point by point below. Revisions have been made to the manuscript to clarify limitations, qualify claims, and add quantitative validation where appropriate.

read point-by-point responses

  1. Referee: [Abstract / regularization step] Abstract and regularization description: matching the metadirect pair structure (obtained by functional differentiation of the neural metadensity) against test-particle data generated from the identical locally trained neural functional creates a circular dependence. Agreement between the two routes can occur even when the underlying neural approximation violates exact DFT identities (e.g., compressibility sum rule or virial theorem), so the procedure does not automatically guarantee consistency when the pair potential is varied at prediction stage.

    Authors: We agree that matching the two routes within the same approximate neural model does not automatically enforce exact DFT identities such as the compressibility sum rule. The regularization enforces internal consistency between the metadirect differentiation and test-particle routes for the learned functional, but this consistency is approximate. In the revised manuscript we have added explicit discussion of this limitation in the abstract and methods, clarifying that the procedure does not guarantee exact sum-rule compliance independent of the neural approximation. We retain the approach because it still enables first-principles access to pair correlations via the explicit pair-potential dependence of the metadensity functional. revision: partial

  2. Referee: [Abstract] The claim that the metadirect route 'is based on first principles' (abstract) is weakened because both the metadirect differentiation and the test-particle reference derive from the same data-fitted neural representation rather than from exact functional identities independent of the approximation.

    Authors: The metadirect route rests on the exact classical DFT relation that the excess free-energy functional depends explicitly on the pair potential; functional differentiation with respect to this dependence yields the pair correlation function without Ornstein-Zernike inversion. While the neural representation itself is approximate, the underlying functional identities remain first-principles. We have revised the abstract to qualify the statement, specifying that the route leverages exact DFT functional relationships within the learned metadensity framework. revision: yes

  3. Referee: [Results / validation] No explicit validation metrics (error bars, quantitative discrepancy measures, or tests on unseen pair potentials) are described, making it impossible to assess whether the regularization improves predictive accuracy or merely enforces internal consistency within the training distribution.

    Authors: We thank the referee for this observation. The revised manuscript now includes quantitative metrics: mean-squared errors between the metadirect and test-particle pair-correlation functions, standard deviations from multiple independent training runs (reported as error bars), and explicit tests on pair potentials withheld from the training set to assess generalization beyond the training distribution. revision: yes

Circularity Check

2 steps flagged

Regularization matches metadirect outputs to test-particle data generated from the same locally trained neural functional

specific steps
  1. fitted input called prediction [Abstract]
    "We regularize local learning of neural functionals by comparing the pair structure from different routes. Thereby results from metadirect functional differentiation are matched against accurate test particle data from an initial locally trained metadensity functional."

    Test-particle data is generated from the identical initial locally trained metadensity functional, so matching the differentiated pair structure to it enforces agreement between two outputs of the same fitted neural model rather than to independent external data or exact DFT identities.

  2. self citation load bearing [Abstract]
    "We investigate and exploit consequences of the recent neural metadensity functional theory [Kampa et al., Phys. Rev. Lett. 134, 107301 (2025), 10.1103/PhysRevLett.134.107301] for describing the physics of inhomogeneous fluids."

    The metadensity functional and its neural implementation are taken directly from the authors' own prior work; the regularization procedure then depends on outputs produced by that same learned representation without independent verification or external benchmarks.

full rationale

The derivation chain starts from the authors' prior neural metadensity functional (self-cited PRL 2025) and adds a regularization step that generates test-particle pair data from an initial local training of that same functional, then enforces agreement with its own metadirect differentiation. This loop is load-bearing for the claimed first-principles regularization and can hold even when the neural approximation violates exact DFT sum rules. The formal metadirect route itself is non-circular, but the practical consistency enforcement reduces to internal model consistency rather than external validation.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the functional relationships supplied by classical density functional theory and on the trainability of the neural metadensity functional from the authors' prior work. No new physical entities are introduced.

free parameters (1)
  • neural network parameters
    Weights and biases of the metadensity functional are learned from data, constituting fitted parameters whose values are not reported.
axioms (1)
  • domain assumption Classical density functional theory supplies the functional relationships needed for variational calculus and pair-structure extraction.
    Invoked explicitly in the abstract as the foundation for the metadirect routes.

pith-pipeline@v0.9.0 · 5504 in / 1333 out tokens · 62579 ms · 2026-05-15T12:06:28.611013+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

101 extracted references · 101 canonical work pages

  1. [1]

    J. P. Hansen and I. R. McDonald,Theory of Simple Liquids, 4th ed. (Academic Press, London, 2013)

    work page 2013

  2. [2]

    Evans, The nature of the liquid-vapour interface and other topics in the statistical mechanics of non-uniform, classical fluids, Adv

    R. Evans, The nature of the liquid-vapour interface and other topics in the statistical mechanics of non-uniform, classical fluids, Adv. Phys.28, 143 (1979)

    work page 1979

  3. [3]

    Evans, Density functionals in the theory of nonuni- form fluids, Chap

    R. Evans, Density functionals in the theory of nonuni- form fluids, Chap. 3 inFundamentals of Inhomogeneous Fluids, edited by D. Henderson (Dekker, New York, 1992)

    work page 1992

  4. [4]

    Evans, M

    R. Evans, M. Oettel, R. Roth, and G. Kahl, New devel- opments in classical density functional theory, J. Phys.: Condens. Matter28, 240401 (2016)

    work page 2016

  5. [5]

    Huang, G

    B. Huang, G. F. von Rudorff, O. A. von Lilienfeld, The central role of density functional theory in the AI age, Science381, 170 (2023)

    work page 2023

  6. [6]

    N. D. Mermin, Thermal properties of the inhomoge- neous electron gas, Phys. Rev.137, A1441 (1965)

    work page 1965

  7. [7]

    Baus, Broken symmetry and invariance properties of classical fluids, Mol

    M. Baus, Broken symmetry and invariance properties of classical fluids, Mol. Phys.51, 211 (1984)

    work page 1984

  8. [8]

    J. R. Henderson, Statistical mechanical sum rules, Chap. 2 inFundamentals of Inhomogeneous Fluids, edited by D. Henderson (Dekker, New York, 1992)

    work page 1992

  9. [9]

    Hermann and M

    S. Hermann and M. Schmidt, Noether’s theorem in sta- tistical mechanics, Commun. Phys.4, 176 (2021)

    work page 2021

  10. [10]

    M¨ uller, S

    J. M¨ uller, S. Hermann, F. Samm¨ uller, and M. Schmidt, Gauge invariance of equilibrium statistical mechanics, 14 Phys. Rev. Lett.133, 217101 (2024); Editors’ Sugges- tion; PRL’s Collection of the Year 2024; Featured in Physics17, 163 (2024) by B. Rotenberg

    work page 2024

  11. [11]

    M¨ uller, F

    J. M¨ uller, F. Samm¨ uller, and M. Schmidt, Why gauge invariance applies to statistical mechanics, J. Phys. A: Math. Theor.58, 125003 (2025)

    work page 2025

  12. [12]

    Samm¨ uller, S

    F. Samm¨ uller, S. Hermann, D. de las Heras, and M. Schmidt, Noether-constrained correlations in equilib- rium liquids, Phys. Rev. Lett.130, 268203 (2023)

    work page 2023

  13. [13]

    Samm¨ uller, S

    F. Samm¨ uller, S. Hermann, D. de las Heras, and M. Schmidt, Neural functional theory for inhomogeneous fluids: Fundamentals and applications, Proc. Natl. Acad. Sci.120, e2312484120 (2023)

    work page 2023

  14. [14]

    Samm¨ uller, M

    F. Samm¨ uller, M. Schmidt, and R. Evans, Neural den- sity functional theory of liquid-gas phase coexistence, Phys. Rev. X15, 011013 (2025); Featured in Physics 18, 17 (2025)

    work page 2025

  15. [15]

    S. M. Kampa, F. Samm¨ uller, M. Schmidt, and R. Evans, Metadensity functional theory for classical fluids: Ex- tracting the pair potential, Phys. Rev. Lett.134, 107301 (2025); PRL Editors’ Suggestion

    work page 2025

  16. [16]

    Samm¨ uller, S

    F. Samm¨ uller, S. Robitschko, S. Hermann, and M. Schmidt, Hyperdensity functional theory of soft mat- ter, Phys. Rev. Lett.133, 098201 (2024); PRL Editors’ Suggestion

    work page 2024

  17. [17]

    Samm¨ uller and M

    F. Samm¨ uller and M. Schmidt, Determining the chem- ical potential via universal density functional learning, Phys. Rev. Lett.136, 068202 (2026); PRL Editors’ Sug- gestion

    work page 2026

  18. [18]

    A. T. Bui and S. J. Cox, Learning classical density func- tionals for ionic fluids, Phys. Rev. Lett.134, 148001 (2025)

    work page 2025

  19. [19]

    Dijkman, M

    J. Dijkman, M. Dijkstra, R. van Roij, M. Welling, J.-W. van de Meent, and B. Ensing, Learning neural free-energy functionals with pair-correlation matching, Phys. Rev. Lett.134, 056103 (2025)

    work page 2025

  20. [20]

    Samm¨ uller and M

    F. Samm¨ uller and M. Schmidt, Neural density function- als: Local learning and pair-correlation matching, Phys. Rev. E110, L032601 (2024); (Letter, Editors’ Sugges- tion)

    work page 2024

  21. [21]

    Robitschko, F

    S. Robitschko, F. Samm¨ uller, M. Schmidt, and R. Evans, Learning the bulk and interfacial physics of liquid-liquid phase separation with neural density func- tionals, J. Chem. Phys.163, 161101 (2025). (Commu- nication)

    work page 2025

  22. [22]

    A. T. Bui and S. J. Cox, Dielectrocapillarity for exquisite control of fluids, Nat. Commun.17, 2661 (2026)

    work page 2026

  23. [23]

    P. Cats, S. Kuipers, S. de Wind, R. van Damme, G. M. Coli, M. Dijkstra, and R. van Roij, Machine-learning free-energy functionals using density profiles from sim- ulations, APL Mater.9, 031109 (2021)

    work page 2021

  24. [24]

    M. M. Kelley, J. Quinton, K. Fazel, N. Karimitari, C. Sutton, R. Sundararaman, Bridging electronic and clas- sical density-functional theory using universal machine- learned functional approximations, J. Chem. Phys.161, 144101 (2024)

    work page 2024

  25. [25]

    Yatsyshin, S

    P. Yatsyshin, S. Kalliadasis, and A. B. Duncan, Physics- constrained Bayesian inference of state functions in clas- sical density-functional theory, J. Chem. Phys.156, 074105 (2022)

    work page 2022

  26. [26]

    Malpica-Morales, P

    A. Malpica-Morales, P. Yatsyshin, M. A. Duran- Olivencia, and S. Kalliadasis, Physics-informed Bayesian inference of external potentials in classical density functional theory, J. Chem. Phys.159, 104109 (2023)

    work page 2023

  27. [27]

    Monti, P

    E. Monti, P. Yatsyshin, K. Gkagkas, A. B. Duncan, Learning density functionals to bridge particle and con- tinuum scales, arXiv:2512.23840 [physics.comp-ph] (ac- cess date: 2026-04-15)

    work page arXiv 2026

  28. [28]

    X. Fang, M. Gu and J. Wu, Reliable emulation of com- plex functionals by active learning with error control, J. Chem. Phys.157, 214109 (2022)

    work page 2022

  29. [29]

    J. Yang, R. Pan, J. Sun, J. Wup, High-dimensional op- erator learning for molecular density functional theory, J. Chem. Theo. Comput.21, 5905 (2025)

    work page 2025

  30. [30]

    R. Pan, X. Fang, K. Azizzadenesheli, M. Liu-Schiaffini, M. Gu, and J. Wu, Neural operators for forward and inverse potential–density mappings in classical density functional theory, J. Chem. Phys.163, 164120 (2025)

    work page 2025

  31. [31]

    K. Ram, J. Dijkman, R. van Roij, J.-W. van de Meent, B. Ensing, M. Welling, and D. Cremers, Learned free- energy functionals from pair-correlation matching for dynamical density functional theory, Phys. Rev. E112, 045314 (2025)

    work page 2025

  32. [32]

    Santos-Silva, P

    T. Santos-Silva, P. I. C. Teixeira, C. Anquetil-Deck, and D. J. Cleaver, Neural-network approach to modeling liq- uid crystals in complex confinement, Phys. Rev. E89, 053316 (2014)

    work page 2014

  33. [33]

    Lin and M

    S.-C. Lin and M. Oettel, A classical density functional from machine learning and a convolutional neural net- work, SciPost Phys.6, 025 (2019)

    work page 2019

  34. [34]

    S.-C. Lin, G. Martius, and M. Oettel, Analytical clas- sical density functionals from an equation learning net- work, J. Chem. Phys.152, 021102 (2020)

    work page 2020

  35. [35]

    Glitsch, J

    F. Glitsch, J. Weimar, and M. Oettel, Neural density functional theory in higher dimensions with convolu- tional layers, Phys. Rev. E111, 055305 (2025)

    work page 2025

  36. [36]

    Simon, L

    A. Simon, L. Belloni, D. Borgis, and M. Oettel, The orientational structure of a model patchy particle fluid: simulations, integral equations, density functional the- ory and machine learning, J. Chem. Phys.162, 034503 (2025)

    work page 2025

  37. [37]

    Simon, J

    A. Simon, J. Weimar, G. Martius, and M. Oettel, Ma- chine learning of a density functional for anisotropic patchy particles, J. Chem. Theory Comput.20, 1062 (2024)

    work page 2024

  38. [38]

    Samm¨ uller, S

    F. Samm¨ uller, S. Hermann, and M. Schmidt, Why neu- ral functionals suit statistical mechanics, J. Phys.: Con- dens. Matter36, 243002 (2024); (Topical Review)

    work page 2024

  39. [39]

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

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

    work page 1989

  40. [40]

    Roth, Fundamental measure theory for hard-sphere mixtures: a review, J

    R. Roth, Fundamental measure theory for hard-sphere mixtures: a review, J. Phys.: Condens. Matter22, 063102 (2010)

    work page 2010

  41. [41]

    K. L. Y. Zhou, A. T. Bui and S. J. Cox, The roles of bulk and surface thermodynamics in the selective adsorption of a confined azeotropic mixture, J. Phys. Chem. B, ASAP Article (2026)

    work page 2026

  42. [42]

    Samm¨ uller and M

    F. Samm¨ uller and M. Schmidt, Why hyperdensity func- tionals describe any equilibrium observable, J. Phys.: Condens. Matter37, 083001 (2025); (Topical Review)

    work page 2025

  43. [43]

    Samm¨ uller, Neural functional theory for inhomoge- neous fluids – Tutorial,https://github.com/sfalmo/ NeuralDFT-Tutorial(access date: 2026-04-15)

    F. Samm¨ uller, Neural functional theory for inhomoge- neous fluids – Tutorial,https://github.com/sfalmo/ NeuralDFT-Tutorial(access date: 2026-04-15)

    work page 2026

  44. [44]

    Schmidt, Power functional theory for many-body dy- namics, Rev

    M. Schmidt, Power functional theory for many-body dy- namics, Rev. Mod. Phys.94, 015007 (2022). 15

    work page 2022

  45. [45]

    de las Heras, T

    D. de las Heras, T. Zimmermann, F. Samm¨ uller, S. Her- mann, and M. Schmidt, Perspective: How to overcome dynamical density functional theory, J. Phys.: Condens. Matter35, 271501 (2023); (Invited Perspective)

    work page 2023

  46. [46]

    Zimmermann, F

    T. Zimmermann, F. Samm¨ uller, S. Hermann, M. Schmidt, and D. de las Heras, Neural force functional for non-equilibrium many-body colloidal systems, Mach. Learn.: Sci. Technol.5, 035062 (2024)

    work page 2024

  47. [47]

    Fortini, D

    A. Fortini, D. de las Heras, J. M. Brader, and M. Schmidt, Superadiabatic forces in Brownian many- body dynamics, Phys. Rev. Lett.113, 167801 (2014)

    work page 2014

  48. [48]

    Krinninger, M

    P. Krinninger, M. Schmidt, and J. M. Brader, Nonequi- librium phase behaviour from minimization of free power dissipation, Phys. Rev. Lett.117, 208003 (2016)

    work page 2016

  49. [49]

    de las Heras and M

    D. de las Heras and M. Schmidt, Velocity gradient power functional for Brownian dynamics, Phys. Rev. Lett.120, 028001 (2018)

    work page 2018

  50. [50]

    N. C. X. Stuhlm¨ uller, T. Eckert, D. de las Heras, and M. Schmidt, Structural nonequilibrium forces in driven colloidal systems, Phys. Rev. Lett.121, 098002 (2018)

    work page 2018

  51. [51]

    Hermann, D

    S. Hermann, D. de las Heras, and M. Schmidt, Non- negative interfacial tension in phase-separated active Brownian particles, Phys. Rev. Lett.123, 268002 (2019)

    work page 2019

  52. [52]

    de las Heras and M

    D. de las Heras and M. Schmidt, Flow and structure in nonequilibrium Brownian many-body systems, Phys. Rev. Lett.125, 018001 (2020)

    work page 2020

  53. [53]

    L. L. Treffenst¨ adt and M. Schmidt, Universality in driven and equilibrium hard sphere liquid dynamics, Phys. Rev. Lett.126, 058002 (2021)

    work page 2021

  54. [54]

    Renner, M

    J. Renner, M. Schmidt, and D. de las Heras, Shear and bulk acceleration viscosities in simple fluids, Phys. Rev. Lett.128, 094502 (2022)

    work page 2022

  55. [55]

    A. J. Archer and R. Evans, Binary Gaussian core model: Fluid-fluid phase separation and interfacial properties, Phys. Rev. E64, 041501 (2001)

    work page 2001

  56. [56]

    A. J. Archer, C. N. Likos and R. Evans, Soft-core bi- nary fluid exhibiting aλ-line and freezing to a highly delocalized crystal, J. Phys.: Condens. Matter16, L297 (2004)

    work page 2004

  57. [57]

    A. J. Archer, B. Chacko, and R. Evans, The standard mean-field treatment of inter-particle attraction in clas- sical DFT is better than one might expect, J. Chem. Phys.147, 034501 (2017)

    work page 2017

  58. [58]

    A. J. Archer, and R. Evans, Relationship between local molecular field theory and density functional theory for nonuniform liquids, J. Chem. Phys.138, 014502 (2013)

    work page 2013

  59. [59]

    Schmidt, Density-functional theory for soft poten- tials by dimensional crossover, Phys

    M. Schmidt, Density-functional theory for soft poten- tials by dimensional crossover, Phys. Rev. E60, R6291 (1999)

    work page 1999

  60. [60]

    Schmidt, A density functional for additive mixtures, Phys

    M. Schmidt, A density functional for additive mixtures, Phys. Rev. E62, 3799 (2000)

    work page 2000

  61. [61]

    Schmidt, Fluid structure from density functional theory, Phys

    M. Schmidt, Fluid structure from density functional theory, Phys. Rev. E62, 4976 (2000)

    work page 2000

  62. [62]

    Schmidt, Isometric and metamorphic operations on the space of local fundamental measures, Mol

    M. Schmidt, Isometric and metamorphic operations on the space of local fundamental measures, Mol. Phys. 109, 1253 (2011)

    work page 2011

  63. [63]

    R. L. Henderson, A uniqueness theorem for fluid pair correlation functions, Phys. Lett. A49, 197 (1974)

    work page 1974

  64. [64]

    G. M. Coli, E. Boattini, L. Filion, and M. Dijkstra, Inverse design of soft materials via a deep learning-based evolutionary strategy, Sci. Adv.8, eabj6731 (2022)

    work page 2022

  65. [65]

    B. A. Lindquist, R. B. Jadrich, and T. M. Truskett, Communication: Inverse design for self-assembly via on-the-fly optimization, J. Chem. Phys.145, 111101 (2016)

    work page 2016

  66. [66]

    Z. M. Sherman, M. P. Howard, B. A. Lindquist, R. B. Jadrich, and T. M. Truskett, Inverse methods for design of soft materials, J. Chem. Phys.152, 140902 (2020)

    work page 2020

  67. [67]

    J. K. Percus, Approximation methods in classical sta- tistical mechanics, Phys. Rev. Lett.8, 462 (1962)

    work page 1962

  68. [68]

    G¨ ul, R

    M. G¨ ul, R. Roth, and R. Evans, Using test particle sum rules to construct accurate functionals in classical den- sity functional theory, Phys. Rev. E110, 064115 (2024)

    work page 2024

  69. [69]

    G¨ ul, R

    M. G¨ ul, R. Roth, and R. Evans, Using test particle sum rules to improve approximations in classical density functional theory: White-Bear and White-Bear mark II versions of the Lutsko functional, Phys. Rev. E113, 034104 (2026)

    work page 2026

  70. [70]

    A. G. Baydin, B. A. Pearlmutter, A. A. Radul, and J. M. Siskind, Automatic differentiation in machine learning: A survey, J. Mach. Learn. Res.18, 1 (2018)

    work page 2018

  71. [71]

    Stierle, G

    R. Stierle, G. Bauer, N. Thiele, B. Bursik, P. Rehner, and J. Gross, Classical density functional theory in three dimensions with GPU-accelerated automatic differenti- ation: Computational performance analysis using the example of adsorption in covalent-organic frameworks, Chem. Eng. Sci.298, 120380 (2024)

    work page 2024

  72. [72]

    M. C. Stewart and R. Evans, Phase behavior and struc- ture of a fluid confined between competing (solvophobic and solvophilic) walls, Phys. Rev. E86, 031601 (2012)

    work page 2012

  73. [73]

    Evans and M

    R. Evans and M. C. Stewart, The local compressibility of liquids near non-adsorbing substrates: a useful mea- sure of solvophobicity and hydrophobicity?, J. Phys.: Condens. Matter27, 194111 (2015)

    work page 2015

  74. [74]

    Evans, M

    R. Evans, M. C. Stewart, and N. B. Wilding, A unified description of hydrophilic and superhydrophobic sur- faces in terms of the wetting and drying transitions of liquids, Proc. Natl. Acad. Sci.116, 23901 (2019)

    work page 2019

  75. [75]

    M. K. Coe, R. Evans, and N. B. Wilding, Density deple- tion and enhanced fluctuations in water near hydropho- bic solutes: identifying the underlying physics, Phys. Rev. Lett.128, 045501 (2022)

    work page 2022

  76. [76]

    M. K. Coe, R. Evans, and N. B. Wilding, Measures of fluctuations for a liquid near critical drying, Phys. Rev. E105, 044801 (2022)

    work page 2022

  77. [77]

    Eckert, N

    T. Eckert, N. C. X. Stuhlm¨ uller, F. Samm¨ uller, and M. Schmidt, Fluctuation profiles in inhomogeneous fluids, Phys. Rev. Lett.125, 268004 (2020)

    work page 2020

  78. [78]

    Eckert, N

    T. Eckert, N. C. X. Stuhlm¨ uller, F. Samm¨ uller, and M. Schmidt, Local measures of fluctuations in inhomo- geneous liquids: Statistical mechanics and illustrative applications, J. Phys.: Condens. Matter35, 425102 (2023)

    work page 2023

  79. [79]

    Tarazona and R

    P. Tarazona and R. Evans, Long ranged correlations at a solid-fluid interface a signature of the approach to complete wetting, Mol. Phys.47, 1033 (1982)

    work page 1982

  80. [80]

    M. K. Coe, R. Evans, and N. B. Wilding, Understanding the physics of hydrophobic solvation, J. Chem. Phys. 158, 034508 (2023)

    work page 2023

Showing first 80 references.