pith. sign in

arxiv: 2002.06680 · v5 · submitted 2020-02-16 · ⚛️ physics.bio-ph · cond-mat.soft· cond-mat.stat-mech· q-bio.CB· q-bio.QM

Inferring the dynamics of underdamped stochastic systems

Pith reviewed 2026-05-24 14:33 UTC · model grok-4.3

classification ⚛️ physics.bio-ph cond-mat.softcond-mat.stat-mechq-bio.CBq-bio.QM
keywords underdamped Langevin equationstochastic inferencecell migrationflockingparameter estimationmeasurement errorsdiscrete sampling
0
0 comments X

The pith

A framework called Underdamped Langevin Inference recovers the parameters of underdamped stochastic dynamics from discrete noisy trajectories.

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

The paper develops a method to extract the force, friction, and noise terms in underdamped Langevin equations when data arrives only at discrete times and carries measurement errors. This matters for systems such as migrating cells and animal groups whose motion is thought to follow these equations but cannot be observed continuously or without noise. The resulting operational procedure, Underdamped Langevin Inference, is tested on real cell trajectories and on simulated flocks that include alignment rules. It returns both the inferred parameters and a self-consistent estimate of how reliable those parameters are.

Core claim

We derive a principled framework to infer the dynamics of underdamped stochastic systems from realistic experimental trajectories, sampled at discrete times and subject to measurement errors. This framework yields an operational method, Underdamped Langevin Inference (ULI), which performs well on experimental trajectories of single migrating cells and in complex high-dimensional systems, including flocks with Viscek-like alignment interactions. Our method is robust to experimental measurement errors, and includes a self-consistent estimate of the inference error.

What carries the argument

Underdamped Langevin Inference (ULI), which recovers the deterministic and stochastic terms of the underdamped Langevin equation while correcting for discrete sampling intervals and additive measurement noise.

If this is right

  • ULI recovers accurate parameters from experimental trajectories of single migrating cells.
  • The method extends without modification to high-dimensional systems that include alignment interactions among many agents.
  • Inference remains stable even when measurement errors are comparable to the true displacements.
  • Each run supplies its own estimate of how large the inference uncertainty is.

Where Pith is reading between the lines

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

  • The same correction strategy for discrete sampling and noise could be adapted to overdamped Langevin equations or to other stochastic differential equations.
  • Running ULI on cell trajectories recorded under different conditions might reveal how effective friction or noise strength changes with environment.
  • In collective systems the inferred alignment strength could be compared directly with the Viscek-like rule used in the simulation to test which interaction form best matches data.

Load-bearing premise

The observed trajectories are produced by an underdamped Langevin process whose parameters can still be recovered when the data are recorded at discrete times and contain measurement errors.

What would settle it

Generate trajectories from a known underdamped Langevin equation, add realistic discrete sampling and measurement noise, run ULI on the resulting data, and check whether the recovered parameters agree with the known inputs inside the reported error bars.

Figures

Figures reproduced from arXiv: 2002.06680 by Chase P. Broedersz, David B. Br\"uckner, Pierre Ronceray.

Figure 1
Figure 1. Figure 1: FIG. 1 [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3 [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
read the original abstract

Many complex systems, ranging from migrating cells to animal groups, exhibit stochastic dynamics described by the underdamped Langevin equation. Inferring such an equation of motion from experimental data can provide profound insight into the physical laws governing the system. Here, we derive a principled framework to infer the dynamics of underdamped stochastic systems from realistic experimental trajectories, sampled at discrete times and subject to measurement errors. This framework yields an operational method, Underdamped Langevin Inference (ULI), which performs well on experimental trajectories of single migrating cells and in complex high-dimensional systems, including flocks with Viscek-like alignment interactions. Our method is robust to experimental measurement errors, and includes a self-consistent estimate of the inference error.

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

0 major / 2 minor

Summary. The manuscript derives a principled framework for inferring the parameters of underdamped stochastic dynamics (governed by the underdamped Langevin equation) from trajectories that are sampled at discrete times and corrupted by measurement noise. The resulting operational method, Underdamped Langevin Inference (ULI), is reported to recover the dynamics accurately on experimental single-cell migration trajectories and on high-dimensional simulations of flocks with Vicsek-like alignment interactions; the method is claimed to be robust to measurement errors and to supply a self-consistent estimate of inference error.

Significance. If the central derivation holds, the work supplies a concrete, noise-robust procedure for extracting physical parameters from realistic experimental data in underdamped systems. This is relevant to biophysics and collective behavior, where underdamped Langevin descriptions are common yet inference from discrete noisy observations has been challenging. The inclusion of an internal error estimate and tests on both real cell data and Vicsek-type flocks are positive features that would increase the method's utility if validated.

minor comments (2)
  1. [Abstract] Abstract: the phrase 'Viscek-like alignment interactions' is a typographical error for the standard 'Vicsek-like' model; this should be corrected.
  2. [Abstract] The abstract states that the framework 'yields an operational method' and 'performs well' but supplies no equations, no explicit form of the inferred force or friction terms, and no quantitative error metrics; a short statement of the key inferential relation would improve readability for readers who do not immediately consult the full text.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive assessment of the manuscript and for recommending minor revision. The recognition of ULI's utility for experimental trajectories and high-dimensional systems such as Vicsek-like flocks is appreciated.

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper presents a derivation of the ULI framework for recovering underdamped Langevin parameters from discrete noisy trajectories. The abstract and provided context describe an operational method that performs on cell and flock data, with no quoted equations or steps showing that any claimed prediction or result reduces by construction to fitted inputs, self-definitional relations, or load-bearing self-citations. The central inference procedure appears independent of the target quantities and self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract supplies no concrete information on free parameters, background axioms, or new postulated entities.

pith-pipeline@v0.9.0 · 5670 in / 948 out tokens · 25790 ms · 2026-05-24T14:33:30.540352+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

42 extracted references · 42 canonical work pages

  1. [1]

    author author J. P. \ Crutchfield \ and\ author B. McNamara ,\ http://www.complex-systems.com/pdf/01-3-3.pdf journal journal Complex Systems \ volume 1 ,\ pages 417 ( year 1987 ) NoStop

  2. [2]

    author author B. C. \ Daniels \ and\ author I. Nemenman ,\ 10.1038/ncomms9133 journal journal Nature Communications \ volume 6 ,\ pages 1 ( year 2015 ) NoStop

  3. [3]

    author author S. L. \ Brunton , author J. L. \ Proctor , \ and\ author J. N. \ Kutz ,\ 10.1073/pnas.1517384113 journal journal Proc Natl Acad Sci U S A \ volume 113 ,\ pages 3932 ( year 2016 ) NoStop

  4. [4]

    Siegert , author R

    author author S. Siegert , author R. Friedrich , \ and\ author J. Peinke ,\ 10.1016/S0375-9601(98)00283-7 journal journal Phys. Lett. A \ volume 243 ,\ pages 275 ( year 1998 ) NoStop

  5. [5]

    Ragwitz \ and\ author H

    author author M. Ragwitz \ and\ author H. Kantz ,\ 10.1103/PhysRevLett.87.254501 journal journal Physical Review Letters \ volume 87 ,\ pages 254501 ( year 2001 ) NoStop

  6. [6]

    author author M. E. \ Beheiry , author M. Dahan , \ and\ author J. B. \ Masson ,\ 10.1038/nmeth.3441 journal journal Nature Methods \ volume 12 ,\ pages 594 ( year 2015 ) NoStop

  7. [7]

    P \' e rez Garc \' i a , author J

    author author L. P \' e rez Garc \' i a , author J. Donlucas P \' e rez , author G. Volpe , author A. V. Arzola , \ and\ author G. Volpe ,\ 10.1038/s41467-018-07437-x journal journal Nature Communications \ volume 9 ,\ pages 5166 ( year 2018 ) NoStop

  8. [8]

    Frishman \ and\ author P

    author author A. Frishman \ and\ author P. Ronceray ,\ 10.1103/PhysRevX.10.021009 journal journal Physical Review X \ volume 10 ,\ pages 21009 ( year 2020 ) NoStop

  9. [9]

    Selmeczi , author S

    author author D. Selmeczi , author S. Mosler , author P. H. \ Hagedorn , author N. B. \ Larsen , \ and\ author H. Flyvbjerg ,\ 10.1529/biophysj.105.061150 journal journal Biophysical journal \ volume 89 ,\ pages 912 ( year 2005 ) NoStop

  10. [10]

    Li , author E

    author author L. Li , author E. C. \ Cox , \ and\ author H. Flyvbjerg ,\ 10.1088/1478-3975/8/4/046006 journal journal Physical Biology \ volume 8 ,\ pages 046006 ( year 2011 ) NoStop

  11. [11]

    Sep \' u lveda , author L

    author author N. Sep \' u lveda , author L. Petitjean , author O. Cochet , author E. Grasland-Mongrain , author P. Silberzan , \ and\ author V. Hakim ,\ 10.1371/journal.pcbi.1002944 journal journal PLoS Computational Biology \ volume 9 ( year 2013 ),\ 10.1371/journal.pcbi.1002944 NoStop

  12. [12]

    D'alessandro , author A

    author author J. D'alessandro , author A. P. \ Solon , author Y. Hayakawa , author C. Anjard , author F. Detcheverry , author J. P. \ Rieu , \ and\ author C. Rivi \` e re ,\ 10.1038/nphys4180 journal journal Nature Physics \ volume 13 ,\ pages 999 ( year 2017 ) NoStop

  13. [13]

    u ckner , author A. Fink , author C. Schreiber , author P. J. F. \ R \

    author author D. B. \ Br \" u ckner , author A. Fink , author C. Schreiber , author P. J. F. \ R \" o ttgermann , author J. O. \ R \" a dler , \ and\ author C. P. \ Broedersz ,\ 10.1038/s41567-019-0445-4 journal journal Nature Physics \ volume 15 ,\ pages 595 ( year 2019 ) NoStop

  14. [14]

    author author G. J. \ Stephens , author B. Johnson-Kerner , author W. Bialek , \ and\ author W. S. \ Ryu ,\ 10.1371/journal.pcbi.1000028 journal journal PLoS Comput Biol \ volume 4 ,\ pages e1000028 ( year 2008 ) NoStop

  15. [15]

    author author G. J. \ Stephens , author M. Bueno , author D. Mesquita , author W. S. \ Ryu , \ and\ author W. Bialek ,\ 10.1073/pnas.1007868108 journal journal Proc. Natl. Acad. Sci. USA \ volume 108 ,\ pages 7286 ( year 2011 ) NoStop

  16. [16]

    Gautrais , author F

    author author J. Gautrais , author F. Ginelli , author R. Fournier , author S. Blanco , author M. Soria , author H. Chat \' e , \ and\ author G. Theraulaz ,\ 10.1371/journal.pcbi.1002678 journal journal PLoS Computational Biology \ volume 8 ( year 2012 ),\ 10.1371/journal.pcbi.1002678 NoStop

  17. [17]

    Katz , author K

    author author Y. Katz , author K. Tunstr m , author C. C. \ Ioannou , author C. Huepe , \ and\ author I. D. \ Couzin ,\ 10.1073/pnas.1107583108 journal journal Proceedings of the National Academy of Sciences of the United States of America \ volume 108 ,\ pages 18720 ( year 2011 ) NoStop

  18. [18]

    Jhawar , author R

    author author J. Jhawar , author R. G. \ Morris , author U. R. \ Amith-Kumar , author M. Danny Raj , author T. Rogers , author H. Rajendran , \ and\ author V. Guttal ,\ 10.1038/s41567-020-0787-y journal journal Nature Physics \ volume 16 ,\ pages 488 ( year 2020 ) NoStop

  19. [19]

    Cavagna , author A

    author author A. Cavagna , author A. Cimarelli , author I. Giardina , author G. Parisi , author R. Santagati , author F. Stefanini , \ and\ author M. Viale ,\ 10.1073/pnas.1005766107 journal journal Proceedings of the National Academy of Sciences \ volume 107 ,\ pages 11865 ( year 2010 ) NoStop

  20. [20]

    Attanasi , author A

    author author A. Attanasi , author A. Cavagna , author L. Del Castello , author I. Giardina , author T. S. \ Grigera , author A. Jelic , author S. Melillo , author L. Parisi , author O. Pohl , author E. Shen , \ and\ author M. Viale ,\ 10.1038/NPHYS3035 journal journal Nature Physics \ volume 10 ,\ pages 691 ( year 2014 a ) NoStop

  21. [21]

    Buhl , author D

    author author J. Buhl , author D. J. T. \ Sumpter , author I. D. \ Couzin , author J. J. \ Hale , author E. Despland , author E. R. \ Miller , \ and\ author S. J. \ Simpson ,\ 10.1126/science.1125142 journal journal Science \ volume 312 ,\ pages 1402 ( year 2006 ) NoStop

  22. [22]

    Attanasi , author A

    author author A. Attanasi , author A. Cavagna , author L. Del Castello , author I. Giardina , author S. Melillo , author L. Parisi , author O. Pohl , author B. Rossaro , author E. Shen , author E. Silvestri , \ and\ author M. Viale ,\ 10.1371/journal.pcbi.1003697 journal journal PLoS Computational Biology \ volume 10 ,\ pages e1003697 ( year 2014 b ) NoStop

  23. [23]

    Gogia \ and\ author J

    author author G. Gogia \ and\ author J. C. \ Burton ,\ 10.1103/PhysRevLett.119.178004 journal journal Physical Review Letters \ volume 119 ,\ pages 178004 ( year 2017 ) NoStop

  24. [24]

    author author A. E. \ Brown \ and\ author B. de Bivort ,\ 10.1101/220855 journal journal Nature Physics \ volume 14 ,\ pages 653 ( year 2018 ) NoStop

  25. [25]

    author author J. N. \ Pedersen , author L. Li , author C. Gradinaru , author R. H. \ Austin , author E. C. \ Cox , \ and\ author H. Flyvbjerg ,\ 10.1103/PhysRevE.94.062401 journal journal Physical Review E \ volume 94 ,\ pages 062401 ( year 2016 ) NoStop

  26. [26]

    Ferretti , author V

    author author F. Ferretti , author V. Chard \` e s , author T. Mora , author A. M. \ Walczak , \ and\ author I. Giardina ,\ http://arxiv.org/abs/1912.10491 journal journal Phys. Rev. X \ volume in press ( year 2020 ) NoStop

  27. [27]

    Lehle \ and\ author J

    author author B. Lehle \ and\ author J. Peinke ,\ 10.1103/PhysRevE.91.062113 journal journal Physical Review E - Statistical, Nonlinear, and Soft Matter Physics \ volume 91 ,\ pages 062113 ( year 2015 ) NoStop

  28. [28]

    @ ( eqn:process \@@italiccorr ) in the It\^o-sense, the inferred force field F_ ( x , v ) corresponds to this convention

    note Since we interpret eqn. @ ( eqn:process \@@italiccorr ) in the It\^o-sense, the inferred force field F_ ( x , v ) corresponds to this convention. Stop

  29. [29]

    note See Supplemental Material at [URL will be inserted by publisher] for detailed derivations of the correction terms and estimators. Stop

  30. [30]

    @ ( eqn:guesstimator \@@italiccorr ) changes from 1/6 to 1/2

    note Note that due to the change of definition of hat 05E v , the prefactor of the correction term in Eq. @ ( eqn:guesstimator \@@italiccorr ) changes from 1/6 to 1/2 . Stop

  31. [31]

    Kruse \ and\ author F

    author author K. Kruse \ and\ author F. J \" u licher ,\ 10.1016/j.ceb.2004.12.007 journal journal Current Opinion in Cell Biology \ volume 17 ,\ pages 20 ( year 2005 ) NoStop

  32. [32]

    u ckner , author C. Schreiber , author P. J. \ R \

    author author A. Fink , author D. B. \ Br \" u ckner , author C. Schreiber , author P. J. \ R \" o ttgermann , author C. P. \ Broedersz , \ and\ author J. O. \ R \" a dler ,\ 10.1016/j.bpj.2019.11.3389 journal journal Biophysical Journal \ volume 118 ,\ pages 552 ( year 2020 ) NoStop

  33. [33]

    u ckner , author A. Fink , author J. O. \ R \

    author author D. B. \ Br \" u ckner , author A. Fink , author J. O. \ R \" a dler , \ and\ author C. P. \ Broedersz ,\ http://dx.doi.org/10.1098/rsif.2019.0689 journal journal J. R. Soc. Interface \ volume 17 ,\ pages 20190689 ( year 2020 ) NoStop

  34. [34]

    Friedrich , author J

    author author R. Friedrich , author J. Peinke , \ and\ author C. Renner ,\ 10.1103/PhysRevLett.84.5224 journal journal Physical Review Letters \ volume 84 ,\ pages 5224 ( year 2000 ) NoStop

  35. [35]

    Lukeman , author Y

    author author R. Lukeman , author Y. X. \ Li , \ and\ author L. Edelstein-Keshet ,\ 10.1073/pnas.1001763107 journal journal Proceedings of the National Academy of Sciences of the United States of America \ volume 107 ,\ pages 12576 ( year 2010 ) NoStop

  36. [36]

    Cavagna , author I

    author author A. Cavagna , author I. Giardina , \ and\ author T. S. \ Grigera ,\ 10.1016/j.physrep.2017.11.003 journal journal Physics Reports \ volume 728 ,\ pages 1 ( year 2017 ) NoStop

  37. [37]

    Bialek , author A

    author author W. Bialek , author A. Cavagna , author I. Giardina , author T. Mora , author E. Silvestri , author M. Viale , \ and\ author A. M. \ Walczak ,\ 10.1073/pnas.1118633109 journal journal Proceedings of the National Academy of Sciences \ volume 109 ,\ pages 4786 ( year 2012 ) NoStop

  38. [38]

    Czirok , author E

    author author Tamas Vicsek , author A. Czirok , author E. Ben-Jacob , author I. Cohen , \ and\ author O. Shochet ,\ https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.75.1226 journal journal Phys. Rev. Lett. \ volume 74 ( year 1995 ) NoStop

  39. [39]

    Gr \' e goire , author H

    author author G. Gr \' e goire , author H. Chat \' e , \ and\ author Y. Tu ,\ 10.1016/S0167-2789(03)00102-7 journal journal Physica D: Nonlinear Phenomena \ volume 181 ,\ pages 157 ( year 2003 ) NoStop

  40. [40]

    Chat \' e , author F

    author author H. Chat \' e , author F. Ginelli , author G. Gr \' e goire , author F. Peruani , \ and\ author F. Raynaud ,\ 10.1140/epjb/e2008-00275-9 journal journal European Physical Journal B \ volume 64 ,\ pages 451 ( year 2008 ) NoStop

  41. [41]

    note A readily usable Python package to perform Underdamped Langevin Inference is available at https://github.com/ronceray/UnderdampedLangevinInference. Stop

  42. [42]

    Baldovin , author A

    author author M. Baldovin , author A. Puglisi , \ and\ author A. Vulpiani ,\ 10.1371/journal.pone.0212135 journal journal PLoS ONE \ volume 14 ,\ pages 1 ( year 2019 ) NoStop