pith. sign in

arxiv: 2507.12960 · v1 · submitted 2025-07-17 · ⚛️ physics.chem-ph · physics.comp-ph· physics.plasm-ph

Accelerated free energy estimation in ab initio path integral Monte Carlo simulations

Pontus Svensson , Fotios Kalkavouras , Uwe Hernandez Acosta , Zhandos A. Moldabekov , Panagiotis Tolias , Jan Vorberger , Tobias Dornheim This is my paper

Pith reviewed 2026-05-19 04:44 UTC · model grok-4.3

classification ⚛️ physics.chem-ph physics.comp-phphysics.plasm-ph
keywords path integral Monte Carlofree energy estimationuniform electron gasfermion sign problemextrapolation techniquesab initio simulationschemical accuracyEwald interaction
0
0 comments X

The pith

Combining intermediate-reference acceleration with ξ-extrapolation enables free energy evaluation for 1000 electrons with finite-size and statistical errors below chemical accuracy.

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

The paper introduces a methodology to speed up free energy calculations in path integral Monte Carlo simulations of the uniform electron gas. By using a spherically averaged Ewald interaction as an inexpensive intermediate reference, the costly interaction part of the free energy can be computed much more rapidly. A ξ-extrapolation method is also employed to mitigate the fermion sign problem at large particle numbers. When these techniques are combined, the free energy for systems as large as 1000 electrons can be obtained with both finite-size and statistical errors below the threshold of chemical accuracy. This opens the door to more accurate modeling of systems important for planetary science and inertial confinement fusion.

Core claim

The central discovery is that an intermediate artificial reference system based on the spherically averaged Ewald interaction, paired with a ξ-extrapolation technique, permits the interaction contribution to the free energy to be evaluated up to 18 times faster while resolving the sign problem sufficiently to achieve chemical accuracy for a 1000-electron system.

What carries the argument

The intermediate artificial reference system employing the spherically averaged Ewald interaction, which reduces the numerical cost of interaction evaluations and allows the correction to the true interaction to be computed with manageable variance.

If this is right

  • The interaction contribution to the free energy is evaluated up to 18 times faster than using only the Ewald method.
  • Free energy calculations become feasible for systems of 1000 electrons.
  • Both finite-size and statistical errors can be kept below chemical accuracy.
  • The approach extends to systems with low to moderate quantum degeneracy relevant for fusion modeling.

Where Pith is reading between the lines

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

  • This method could be adapted to other reference systems or particle types to further increase efficiency.
  • It may allow simulations of warm dense matter at scales previously computationally prohibitive.
  • Connections to other many-body techniques might yield hybrid acceleration strategies.

Load-bearing premise

The spherically averaged Ewald reference stays close enough to the actual interactions that the correction term maintains low enough variance at large particle counts.

What would settle it

Performing the free energy calculation with both the accelerated method and a full Ewald computation on a system size where the latter is still practical, and finding that the results differ by more than chemical accuracy.

Figures

Figures reproduced from arXiv: 2507.12960 by Fotios Kalkavouras, Jan Vorberger, Panagiotis Tolias, Pontus Svensson, Tobias Dornheim, Uwe Hernandez Acosta, Zhandos A. Moldabekov.

Figure 1
Figure 1. Figure 1: Schematic showing the different systems used to compute the free energies [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The average sign SEw for the UEG at rs = 3.23 and θ = 1.0, for different system sizes N and permutation weight ξ. Note that ξ < 0 and all values of SEw are less than unity. The error bars, correspond to 95%-confidence intervals estimated from simulations with vary￾ing initial conditions. The extrapolation of the confidence interval assuming aS is independent of ξ is shown in the highlighted areas. The poin… view at source ↗
Figure 3
Figure 3. Figure 3: Finite size and finite P corrected ex￾change correlation free energies for the UEG at rs = 3.23 and θ = 1.0 shown as the difference from the GDSMFB parametrisation88 with a value of f Ref. xc = −0.15529 Ha (Ref.). The re￾maining N dependence is a fraction of a per￾cent. The computations have been performed in a variety of ways, using the physical interac￾tion (Ewald-only) or accelerated method (Ac￾celerate… view at source ↗
Figure 4
Figure 4. Figure 4: The size (a) and scaling (b) of the N dependent free energy for the UEG at rs = 3.23 and θ = 1.0. The correction term ∆f YR-Ew a and FSC vanish as N → ∞ while the contribution from the η-ensemble and the sign converges to a finite value. The subtraction of the infinite system size contribution N = ∞ in (b) was fa￾cilitated by a fit ∆f = d0 +d1N −d2 for N ≤ 66. The interaction components are seen to con￾ver… view at source ↗
read the original abstract

We present a methodology for accelerating the estimation of the free energy from path integral Monte Carlo simulations by considering an intermediate artificial reference system where interactions are inexpensive to evaluate numerically. Using the spherically averaged Ewald interaction as this intermediate reference system for the uniform electron gas, the interaction contribution for the free energy was evaluated up to 18 times faster than the Ewald-only method. Furthermore, a $\xi$-extrapolation technique was tested and applied to alleviate the fermion sign problem and to resolve the sign for large particle numbers. Combining these two techniques enabled the evaluation of the free energy for a system of 1000 electrons, where both finite-size and statistical errors are below chemical accuracy. The general procedure can be applied to systems relevant for planetary and inertial confinement fusion modeling with low to moderate levels of quantum degeneracy.

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

2 major / 1 minor

Summary. The manuscript presents a methodology to accelerate free energy estimation in ab initio path integral Monte Carlo simulations of the uniform electron gas. An intermediate reference system based on the spherically averaged Ewald interaction is used to evaluate the interaction contribution up to 18 times faster than direct Ewald sampling. A ξ-extrapolation technique is introduced and applied to mitigate the fermion sign problem, enabling free-energy calculations for N=1000 electrons in which both finite-size and statistical errors fall below chemical accuracy. The procedure is framed as applicable to low-to-moderate degeneracy plasmas relevant to planetary and inertial-confinement-fusion modeling.

Significance. If the reference-system correction and ξ-extrapolation are shown to introduce no uncontrolled bias, the reported speed-up and the ability to reach N=1000 with chemical accuracy would constitute a practical advance for large-scale PIMC free-energy calculations. The direct sampling against an external Ewald benchmark avoids circularity and supplies a concrete basis for the variance-reduction claim.

major comments (2)
  1. [ξ-extrapolation section] The section describing the ξ-extrapolation procedure: the headline claim that free energies at N=1000 are unbiased and below chemical accuracy rests on the assumption that a low-order fit in ξ recovers the ξ=1 limit without residual bias larger than the target accuracy. No direct comparisons to exact diagonalization, small-N PIMC, or high-temperature expansions are provided to bound extrapolation uncertainty, and the reported error bars are stated to be purely statistical.
  2. [Abstract and §3] Abstract and the results section reporting the 18-fold speed-up: the claim that the spherically averaged Ewald reference permits an 18-times faster interaction contribution relies on the correction term having acceptably low variance at large particle numbers. No quantitative demonstration is given that the variance of this correction remains controlled when N reaches 1000.
minor comments (1)
  1. Notation for the extrapolation variable ξ and the fitting procedure should be defined more explicitly so that the polynomial order and fitting range can be reproduced from the text alone.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their thoughtful review and positive assessment of the work's potential significance. We address the two major comments below, agreeing where the manuscript can be strengthened and proposing concrete revisions.

read point-by-point responses

  1. Referee: [ξ-extrapolation section] The section describing the ξ-extrapolation procedure: the headline claim that free energies at N=1000 are unbiased and below chemical accuracy rests on the assumption that a low-order fit in ξ recovers the ξ=1 limit without residual bias larger than the target accuracy. No direct comparisons to exact diagonalization, small-N PIMC, or high-temperature expansions are provided to bound extrapolation uncertainty, and the reported error bars are stated to be purely statistical.

    Authors: We agree that an explicit bound on extrapolation bias would strengthen the headline claim. The current manuscript reports only statistical uncertainties and relies on observed stability of the low-order polynomial fit across the simulated ξ range. In the revised version we will add a dedicated subsection with direct validation on smaller systems (N ≤ 38) where we compare the ξ-extrapolated free energy against both direct PIMC at ξ=1 and available high-temperature expansion results. This will quantify any residual systematic error and allow us to state an upper bound relative to chemical accuracy. The reported error bars will remain statistical, but we will explicitly discuss the additional systematic contribution from the fit. revision: yes

  2. Referee: [Abstract and §3] Abstract and the results section reporting the 18-fold speed-up: the claim that the spherically averaged Ewald reference permits an 18-times faster interaction contribution relies on the correction term having acceptably low variance at large particle numbers. No quantitative demonstration is given that the variance of this correction remains controlled when N reaches 1000.

    Authors: We acknowledge that the manuscript does not present an explicit scaling plot of the correction-term variance up to N=1000. The 18-fold speedup figure was obtained from production runs that included the N=1000 case, but the supporting variance data were not shown. In the revision we will add a figure in §3 (or a new supplementary figure) that reports the variance of the difference between the full Ewald and spherically averaged Ewald estimators as a function of N, explicitly including the N=1000 data point. This will demonstrate that the correction variance remains controlled and thereby substantiate the reported computational gain. revision: yes

Circularity Check

0 steps flagged

Central results from direct PIMC sampling against independent Ewald reference; no load-bearing self-definition or fitted prediction

full rationale

The paper defines the intermediate reference as the spherically averaged Ewald interaction (independent of the target free energy) and applies ξ-extrapolation as a numerical technique to mitigate the sign problem. Free energies are obtained from explicit PIMC sampling runs, with reported errors being statistical and finite-size rather than extrapolation fits that reproduce the input by construction. No equation reduces the claimed free energy to a parameter fitted inside the same dataset, and no uniqueness theorem or ansatz is imported solely via self-citation to force the result. This matches the reader's assessment of low circularity; the derivation chain remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The method relies on the existence of a well-behaved intermediate reference whose energy difference to the true interaction has low variance; this is an ad-hoc modeling choice rather than a derived result.

axioms (1)
  • domain assumption The spherically averaged Ewald interaction constitutes a valid and sufficiently close reference for the uniform electron gas.
    Invoked when the authors choose this reference to accelerate the interaction contribution (abstract).

pith-pipeline@v0.9.0 · 5700 in / 1303 out tokens · 25981 ms · 2026-05-19T04:44:41.396040+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

24 extracted references · 24 canonical work pages

  1. [1]

    P.; Stringari, S

    (1) Giorgini, S.; Pitaevskii, L. P.; Stringari, S. Theory of ultracold atomic Fermi gases. Reviews of Modern Physics 2008, 80,

    work page 2008

  2. [2]

    Physical Re- view Letters 1992, 69,

    work page 1992

  3. [3]

    M.; Manninen, M

    (3) Reimann, S. M.; Manninen, M. Elec- tronic structure of quantum dots. Re- views of Modern Physics 2002, 74,

    work page 2002

  4. [4]

    Abnormal quan- tum moment of inertia and structural properties of electrons in 2D and 3D quantum dots: An ab initio path-integral Monte Carlo study

    (4) Dornheim, T.; Yan, Y. Abnormal quan- tum moment of inertia and structural properties of electrons in 2D and 3D quantum dots: An ab initio path-integral Monte Carlo study. New Journal of Physics 2022, 24, 113024. (5) Lorenzen, W.; Becker, A.; Redmer, R. In Frontiers and Challenges in Warm Dense Matter ; Graziani, F., Desjar- lais, M. P., Redmer, R., Tr...

    work page 2022

  5. [5]

    Material prop- erties of Saturn’s interior from ab ini- tio simulations

    (10) Preising, M.; French, M.; Mankovich, C.; Soubiran, F.; Redmer, R. Material prop- erties of Saturn’s interior from ab ini- tio simulations. The Astrophysical Jour- nal Supplement Series 2023, 269,

    work page 2023

  6. [6]

    (12) Fortov, V. E. Extreme states of matter on Earth and in space. Physics-Uspekhi 2009, 52,

    work page 2009

  7. [7]

    (13) Chandrasekhar, S. XLVIII. The density of white dwarf stars. The London, Ed- inburgh, and Dublin Philosophical Mag- azine and Journal of Science 1931, 11, 592–596. (14) Chabrier, G.; Brassard, P.; Fontaine, G.; Saumon, D. Cooling sequences and color-magnitude diagrams for cool white dwarfs with hydrogen atmospheres. The Astrophysical Journal 2000, 543,

    work page 1931

  8. [8]

    H.; Pethick, C

    (15) Gudmundsson, E. H.; Pethick, C. J.; Ep- stein, R. I. Structure of neutron star envelopes. Astrophysical Journal 1983, 272, 286–300. (16) Haensel, P.; Potekhin, A. Y.; Yakovlev, D. G. Neutron Stars 1: Equa- tion of State and Structure ; Springer: New York, NY,

    work page 1983

  9. [9]

    Laser compression of matter to super-high densities: Ther- monuclear (CTR) applications

    (17) Nuckolls, J.; Wood, L.; Thiessen, A.; Zimmerman, G. Laser compression of matter to super-high densities: Ther- monuclear (CTR) applications. Nature 1972, 239, 139–142. (18) Betti, R.; Hurricane, O. A. Inertial- confinement fusion with lasers. Nature Physics 2016, 12, 435–448. (19) Hurricane, O. A.; Patel, P. K.; Betti, R.; Froula, D. H.; Regan, S. P....

    work page 1972

  10. [10]

    P.; Zunger, A

    (30) Perdew, J. P.; Zunger, A. Self-interaction correction to density-functional approx- imations for many-electron systems. Physical Review B 1981, 23,

    work page 1981

  11. [11]

    V.; Sjostrom, T.; Dufty, J.; Trickey, S

    (31) Karasiev, V. V.; Sjostrom, T.; Dufty, J.; Trickey, S. Accurate homogeneous elec- tron gas exchange-correlation free energy for local spin-density calculations. Phys- ical Review Letters 2014, 112, 076403. (32) Dornheim, T.; Groth, S.; Bonitz, M. The uniform electron gas at warm dense mat- ter conditions. Physics Reports 2018, 744, 1–86. (33) Ceperley...

    work page 2014

  12. [12]

    A.; Vor- berger, J.; Dornheim, T

    (34) B¨ ohme, M.; Moldabekov, Z. A.; Vor- berger, J.; Dornheim, T. Static electronic density response of warm dense hydro- gen: Ab initio path integral Monte Carlo Simulations. Phys. Rev. Lett. 2022, 129, 066402. (35) Dornheim, T. Fermion sign problem in path integral Monte Carlo simulations: Quantum dots, ultracold atoms, and warm dense matter. Physical ...

    work page 2022

  13. [13]

    Thermodynamics of fermions at any temperature based on parametrized partition function

    (41) Xiong, Y.; Xiong, H. Thermodynamics of fermions at any temperature based on parametrized partition function. Physical Review E 2023, 107, 055308. (42) Morresi, T.; Garberoglio, G. Normal liq- uid 3He studied by path-integral Monte Carlo with a parametrized partition func- tion. Physical Review B 2025, 111, 014521. (43) Dornheim, T.; Schwalbe, S.; Mol...

    work page 2023

  14. [14]

    P.; Moldabekov, Z

    (44) Dornheim, T.; Schwalbe, S.; B¨ ohme, M. P.; Moldabekov, Z. A.; Vorberger, J.; Tolias, P. Ab initio path integral Monte Carlo simulations of warm dense two-component systems without fixed nodes: Structural proper- ties. The Journal of Chemical Physics 2024, 160, 164111. (45) Dornheim, T.; D¨ oppner, T.; To- lias, P.; B¨ ohme, M. P.; Fletcher, L. B.; G...

    work page 2024

  15. [15]

    P.; Moldabekov, Z

    (46) Dornheim, T.; Schwalbe, S.; Tolias, P.; B¨ ohme, M. P.; Moldabekov, Z. A.; Vor- berger, J. Ab initio density response and local field factor of warm dense hydrogen. Matter and Radiation at Extremes 2024, 9, 057401. (47) Dornheim, T.; Moldabekov, Z.; Schwalbe, S.; Tolias, P.; Vorberger, J. Fermionic free energies from ab initio path integral Monte Car...

    work page arXiv 2024

  16. [16]

    High-pressure phase diagram of beryllium from ab initio free-energy cal- culations

    (55) Wu, J.; Gonz´ alez-Cataldo, F.; Mil- itzer, B. High-pressure phase diagram of beryllium from ab initio free-energy cal- culations. Physical Review B 2021, 104, 014103. (56) Militzer, B.; Gonz´ alez-Cataldo, F.; Zhang, S.; Driver, K. P.; Soubi- ran, F. First-principles equation of state database for warm dense matter com- putation. Physical Review E 2...

    work page 2021

  17. [17]

    (59) Pople, J. A. Nobel lecture: Quantum chemical models. Reviews of Modern Physics 1999, 71,

    work page 1999

  18. [18]

    (60) Zastrau, U.; Appel, K.; Baehtz, C.; Baehr, O.; Batchelor, L.; Bergh¨ auser, A.; Banjafar, M.; Brambrink, E.; Ceran- tola, V.; Cowan, T. E. et al. The High Energy Density Scientific Instrument at the European XFEL. Journal of Syn- chrotron Radiation 2021, 28, 1393–1416. (61) Fletcher, L. B.; Vorberger, J.; Schu- maker, W.; Ruyer, C.; Goede, S.; Galtie...

    work page 2021

  19. [19]

    Prediction of a roton-type feature in warm dense hydrogen

    (62) Hamann, P.; Kordts, L.; Filinov, A.; Bonitz, M.; Dornheim, T.; Vorberger, J. Prediction of a roton-type feature in warm dense hydrogen. Phys. Rev. Res. 2023, 5, 033039. (63) Frenkel, D.; Smit, B. Understanding Molecular Simulation: From Algorithms to Applications, 2nd ed.; Academic Press: San Diego, USA, 2002; Chapter Free En- ergy Calculations, pp 1...

    work page arXiv 2023

  20. [20]

    (93) Loos, P.-F.; Gill, P. M. W. The uniform electron gas. Wiley Interdisciplinary Re- views: Computational Molecular Science 2016, 6, 410–429. (94) Giuliani, G.; Vignale, G. Quantum The- ory of the Electron Liquid ; Cambridge University Press: Cambridge,

    work page 2016

  21. [21]

    A.; Vorberger, J

    (95) Dornheim, T.; Tolias, P.; Mold- abekov, Z. A.; Vorberger, J. Energy response and spatial alignment of the perturbed electron gas. The Journal of Chemical Physics 2023, 158, 164108. (96) Kasim, M. F.; Vinko, S. M. Learn- ing the Exchange-Correlation Functional from Nature with Fully Differentiable Density Functional Theory. Phys. Rev. Lett. 2021, 127,...

    work page arXiv 2023

  22. [22]

    (99) Hastings, W. K. Monte Carlo sampling methods using Markov chains and their applications. Biometrika 1970, 57, 97–

    work page 1970

  23. [23]

    D.; Foulkes, W

    (100) Dornheim, T.; Groth, S.; Sjostrom, T.; Malone, F. D.; Foulkes, W. M. C.; Bonitz, M. Ab initio quantum Monte Carlo simulation of the warm dense elec- tron gas in the thermodynamic limit. Physical Review Letters 2016, 117, 156403. (101) See https://github.com/fdmalone/ uegpy. (102) Singwi, K.; Tosi, M.; Land, R.; Sj¨ olander, A. Electron correlations ...

    work page 2016

  24. [24]

    An accurate equation of state for the one-component plasma in the low coupling regime

    21 (103) Caillol, J.-M.; Gilles, D. An accurate equation of state for the one-component plasma in the low coupling regime. Jour- nal of Physics A: Mathematical and The- oretical 2010, 43, 105501. (104) Sakkos, K.; Casulleras, J.; Boronat, J. High order Chin actions in path integral Monte Carlo. The Journal of Chemical Physics 2009, 130, 204109. (105) Dorn...

    work page 2010