Pith. sign in

REVIEW 4 major objections 5 minor 32 references

The Memory Engine: Self-Organized Coherence from Internal Feedback

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A Brownian particle coupled to its own decaying memory field can spontaneously switch from diffusion to coherent, phase-locked motion.

desk verdict A cleanly written model with a broken bifurcation analysis and unverifiable numerics; worth review but not acceptance as is. read the letter →

arxiv 2505.21711 v1 pith:CVTCKA7N submitted 2025-05-27 cond-mat.stat-mech

Aranyak Sarkar This is my paper
classification cond-mat.stat-mech
keywords memoryengineself-organizedcoherenceinternalfeedbacknon-MarkoviandynamicsBrownianparticleviscoelasticsubstratetransferentropyburst-trap
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

This paper argues that feedback alone can convert stochastic diffusion into structured, phase-locked motion. It builds a minimal continuous-space model, the Coupled Memory Graph Process, in which a single Brownian particle writes its trajectory into a decaying scalar memory field and is in turn steered by the gradient of that field. Simulations show a sharp transition from unstructured diffusion to coherent burst–trap cycles as substrate stiffness increases, with an intermediate coherence island in stiffness–viscosity space. The paper locates this transition as a bifurcation of straight-line motion: transverse perturbations grow when the memory decay rate exceeds a critical value set by the field curvature. The authors conclude that coherence in memory-coupled systems is not a fragile, tuned condition but a generic outcome of internal feedback dynamics.

What carries the argument

The load-bearing mechanism is the closed feedback loop between the particle trajectory and the memory field. The field equation is $\partial_t S = -\alpha_s S + A\int_0^t \Theta_s(t-\tau)G_\sigma(\mathbf{r}-\mathbf{r}(\tau))\,d\tau$, and the particle motion is the Volterra equation $\dot{\mathbf{r}} = \int_0^t \Theta_m(t-\tau)\dot{\mathbf{r}}(\tau)\,d\tau - \kappa\nabla S + \xi(t)$. Linearizing around a straight trajectory and projecting on the transverse direction gives the effective kernel $\Lambda(s)=\Theta_s(s)\,\hat{\mathbf{n}}^\top \partial_{\hat{\mathbf{n}}}\nabla G_\sigma(v_0 s)\cdot\hat{\mathbf{n}}$, and the Laplace-transformed dispersion relation $s[1-\tilde{\Theta}_m(s)]+\kappa[1-\tilde{\Lambda}(s)]=0$. Evaluating at $s=0$ yields the threshold $\alpha_s > \alpha_c = 1/K$; this threshold is the piece that connects the analytic stability analysis to the numerical phase diagram.

What would settle it

One concrete test is to solve the exact linearized perturbation equation numerically, keeping the full Gaussian feedback kernel without the exponential-kernel replacement, and check whether the straight-line trajectory really destabilizes at $\alpha_s = 1/K$; an instability exactly at the predicted threshold would confirm the mechanism, while a stable straight line for $\alpha_s > 1/K$ would refute it.

Watch

Extended reading notes

Core claim

The paper's central discovery is that a single memoryless Brownian particle in two dimensions can spontaneously transition from ordinary diffusion to coherent, phase-locked motion when it is coupled to a scalar field that records its own past trajectory. The field $S(\mathbf{r},t)$ evolves as a spatiotemporal convolution of the path with an exponential memory $\Theta_s(t)=\alpha_s e^{-\alpha_s t}$ and a Gaussian imprint $G_\sigma$, while the particle velocity responds to the gradient $\nabla S$. In simulations, the straight-line solution becomes transversally unstable precisely when the memory decay rate exceeds $\alpha_c = 1/K$, where $K$ is the transverse curvature of the imprint at the origin; this instability coincides with peak transfer entropy from field to particle and with saturation of the field energy. The paper interprets this triple coincidence as a 'memory engine': coherence is a bifurcation-driven outcome of internal feedback, not a consequence of external forcing, optimization, or fine tuning.

Load-bearing premise

The predicted bifurcation boundary assumes that the memory feedback felt by a perturbed trajectory can be reduced to a simple exponential kernel with a constant curvature coefficient, so that the Laplace-transform identity in the paper's Eq. (17) holds; if the position-dependent Gaussian kernel must be kept, the threshold $\alpha_s = 1/K$ is not established.

Editorial extensions

If this is right

  • If the central claim is correct, the same minimal ingredients—a decaying scalar memory field and a gradient feedback force—should produce phase-locked motion in any realization of the model, because coherence arises at a bifurcation threshold rather than at a specially tuned point.
  • The explicit condition $\alpha_s > 1/K$ gives a quantitative prediction: the diffusion-to-coherence transition should occur when the substrate relaxation rate exceeds the curvature of the Gaussian imprint, so increasing stiffness first enhances and then destroys coherence through memory saturation.
  • The triple alignment of energy saturation, transfer-entropy asymmetry, and transverse instability provides three independent diagnostics that should identify the coherence point without fitting parameters.
  • Because the transition is a symmetry-breaking instability, it should persist at finite noise levels, with noise broadening the threshold rather than eliminating it; the simulations with thermal forcing support this reading.
  • The observed burst–trap cycles and the tri-modal speed distribution are concrete signatures that could be sought in experiments with viscoelastic substrates or in walking-droplet analogues.

Reading between the lines

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

  • The paper does not derive a nonlinear amplitude equation near the bifurcation; an editorial extension is to compute one, which would predict the amplitude of the phase-locked oscillations and the width of the coherence island without full simulation.
  • Because the feedback kernel is written with a Gaussian spatial imprint, the same threshold argument should carry over to other localized imprint functions; whether $\alpha_c = 1/K$ remains exact for non-Gaussian imprints is a testable extension.
  • The paper's 'memory engine' language suggests a thermodynamic accounting: an editorial next step is to compute the entropy production of the loop and ask whether the coherent regime coincides with minimal dissipation, connecting this model to information-engine results.
  • A multi-particle version, with one shared memory field written by many trajectories, would test whether the single-particle coherence island is a building block for collective synchronization; the paper mentions this extension but does not analyze it.
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 / 5 minor

Summary. The manuscript proposes a continuous-space realization of the Coupled Memory Graph Process (CMGP): a single Brownian particle moves in two dimensions and deposits a decaying scalar memory field S(r,t) (Eq. 1), while its velocity is governed by an integro-differential equation with intrinsic velocity memory and a gradient-feedback term −κ∇S(r(t),t) (Eq. 3). The paper claims that this closed feedback loop produces a sharp transition from diffusion to coherent burst–trap motion as substrate stiffness E is increased, and it reports three aligned signatures: (i) saturation of the memory-field energy balance, (ii) a peak in transfer-entropy asymmetry, and (iii) a linear stability bifurcation of straight-line trajectories at α_s > 1/K (Section II.F). Numerical simulations in the (E,η) plane show a coherence 'island' near (7.2,5.1), with phase-locked trajectories, oscillatory velocity autocorrelation, superdiffusive MSD, and spectral entrainment. The paper generalizes this into a 'memory engine' design principle. No code, data, or error bars are provided.

Significance. The idea that a minimal memory-mediated feedback loop can spontaneously organize a noisy particle into coherent, phase-locked motion is genuinely interesting and well-motivated by experimental analogues such as walking droplets and viscoelastic active matter. The governing equations are clearly stated, and the attempt to derive an analytic instability threshold is a strength in principle. However, the central analytic derivation is not valid: the Laplace transform step in Eq. (17) is incorrect, and the Gaussian curvature K in Eq. (14) has the wrong sign for the stated imprint. Consequently, the predicted bifurcation threshold α_c = 1/K is not established. In addition, the energy-balance 'signature' is a definitional steady-state identity, so the claimed three-axis alignment reduces to the transfer-entropy peak and a numerical stability diagram that is not independently checkable. If the result were correct, it would be a significant conceptual contribution to non-Markovian active matter; with the present evidence, the central claim is unsupported.

major comments (4)
  1. [II.F, Eq. (17)] Equation (17) is not a valid Laplace transform identity. With Λ(u) defined as in Eq. (14), the feedback term in Eq. (15) equals δy(t)∫_0^t Λ(u)du − ∫_0^t Λ(u)δy(t−u)du, so its transform contains L{C(t)δy(t)} with C(t)=∫_0^t Λ(u)du, not simply \tildeΛ(s)\tilde y(s). For the exponential kernel Λ(t)=α_s K e^{−α_s t}, the exact transform is K[s/(s+α_s)\tilde y(s) − \tilde y(s+α_s)], not [1−α_sK/(s+α_s)]\tilde y(s). Therefore Eq. (18) does not follow, and the zero-frequency evaluation κ(1−α_sK) is not a valid stability criterion. The threshold α_c = 1/K is thus not derived.
  2. [II.F, Eq. (14)] For the Gaussian imprint in Eq. (2), the curvature K = \hat n^T ∂_{\hat n}∇G_σ(0)\hat n evaluates to −1/(2πσ^4) < 0, so the predicted critical rate α_c = 1/K is negative. No sign convention or rescaling is given in the text that would make K positive. The subsequent replacement of Λ(s) by α_s K e^{−α_s s} with K evaluated at the origin also discards the explicit dependence of the kernel on v0s; the Gaussian factor e^{−∥v0s∥²/(2σ²)} cannot be reduced to a pure exponential with the same decay rate. These errors are load-bearing because the α_c = 1/K threshold is the analytical basis for the claimed bifurcation and for the line in Fig. 7(a).
  3. [II.D and III.C, Eq. (8), Fig. 3(b)] The energy balance dε_s/dt = I(t) − D(t) is a definitional identity obtained by differentiating Eq. (4); it contains no dynamical content. For any trajectory with bounded field energy, the infinite-time average of dε_s/dt is zero, so ⟨I(t)⟩ ≈ ⟨D(t)⟩ in steady state for every parameter set, not just at the coherence point. The 'intersection' of ⟨I⟩ and ⟨D⟩ in Fig. 3(b) is therefore not an independent signature of a transition, and the claimed alignment of energy saturation with the transfer-entropy peak does not constitute evidence for a special coherence point.
  4. [III.H, Fig. 7] The numerical stability phase diagram in Fig. 7(a) is presented without error bars, without the number of realizations, and without the numerical scheme used to integrate Eqs. (1) and (3). More importantly, the solid black 'analytical' line is based on the invalid threshold from Section II.F, so the apparent agreement between theory and simulation cannot be assessed. Figure 7(b) shows the expected qualitative growth or decay of δy(t), but this is a direct simulation of the linearized equation and does not validate the specific bifurcation condition.
minor comments (5)
  1. [III.A] The numerical discretization is underspecified: the text states a square grid, periodic domain, Δt = 1, and non-dimensional units, but it does not give the grid spacing, box size, integration scheme for the Volterra integral, or the number of ensemble members, which prevents reproducibility of the phase boundaries in Figs. 2–4 and 7.
  2. [III.H, Fig. 7(a)] Figure 7(a) labels the analytical threshold as α_c ∼ E/η, but Section II.F derives α_c = 1/K with K having units of inverse area; the mapping from K to E/η is asserted rather than derived.
  3. [III.C] The term 'superinformal state' is used without a quantitative definition; please define it or remove it.
  4. [References] Reference [17] is an unpublished manuscript with no arXiv or journal identifier; since it is the foundation of the CMGP, the citation should be completed or the dependence on it stated as unpublished work.
  5. [III.G] The 'memory towers' and the thirteen flights in Section III.G appear to be identified by inspection of a single trajectory; a quantitative recurrence or clustering criterion would strengthen the claim of repeated structure.

Circularity Check

2 steps flagged · score 6.0 of 10

The claimed three-axis alignment is partly built from definitions: the energy-saturation axis is an identity, and the stability threshold is not independently derived from Eq. (15), with Fig. 7 confirming a self-consistency check rather than a prediction.

  1. self definitional [Section II.D, Eq. (8); Section III.C, Fig. 3(b)]
    "so that the net energy balance reads: d/dt εs(t) = I(t) − D(t). (8) ... At intermediate stiffness, the two rates intersect, yielding a vanishing net energy flow ⟨dεs/dt⟩ ≈ 0—a dynamically saturated regime where the field remains energetically balanced yet functionally active."

    Eq. (8) is an exact identity obtained by differentiating the definition ε_s = (1/2)∫S^2 dr and substituting the field equation; it does not encode any dynamical mechanism. For any statistically stationary or time-averaged simulation, ⟨dεs/dt⟩ ≈ 0 automatically implies ⟨I⟩ ≈ ⟨D⟩. Therefore the 'energetic crossover' is not an independent signature of coherence. Presenting it as one of the three aligned axes makes the alignment partly true by construction, not by the physics of the memory engine.

  2. fitted input called prediction [Section II.F, Eqs. (17)-(18); Section III.H, Fig. 7(a)]
    "L{∫_0^t Λ(s) (δy(t) − δy(t − s)) ds} = [1 − Λ̃(s)] ỹ(s). (17) ... Instability occurs when this expression crosses zero, yielding the critical condition: α_s > α_c := 1/K. ... The solid black line represents the critical feedback threshold αc ∼ E/η, derived analytically, and closely tracks the numerically observed transition."

    The critical condition is obtained from an asserted Laplace identity that is not valid: the term δy(t)∫_0^t Λ(s)ds is a product and does not transform to ỹ(s), so Eq. (18) and α_c = 1/K are not derived consequences of Eq. (15). The figure then plots the 'analytic' boundary as α_c ∼ E/η, which is the same stiffness–viscosity mapping used to set the simulations, with no reported value of K or conversion between K and E/η. The numerical stability diagram is generated from the same linearized feedback dynamics, so the agreement of the line with the numerical boundary is a self-consistency check rather than an independent confirmation of a bifurcation prediction.

full rationale

The paper's continuous-space model and numerical phenomenology are presented self-containedly: the coupled integro-differential equations are stated explicitly, and the observed burst–trap cycles, transfer-entropy peak, and phase map do not depend on the prior CMGP paper for their content. The self-citation to [17] is not load-bearing for the derivation. However, two pieces of the central 'three-axis alignment' claim do not survive scrutiny. First, the energy-saturation axis is tautological: since dε/dt = I − D follows from definitions, any stationary simulation satisfies ⟨I⟩ ≈ ⟨D⟩, so the coincidence of the energetic crossover with the transfer-entropy peak is not independent evidence for coherence. Second, the analytic stability threshold α_c = 1/K is not established: Eq. (17) is not a valid Laplace transform identity, because δy(t)∫_0^t Λ(s)ds is a product whose transform is not ỹ(s); the exact characteristic equation for the exponential kernel does not reduce to Eq. (18). The figure's 'analytic' boundary is drawn using the E/η parameter mapping without specifying K, and the numerical stability diagram is computed from the same linearized feedback model, so the agreement is a self-consistency check. These issues make the central 'bifurcation-driven, not tuned' conclusion partially circular, even though the raw numerical observations could be reproducible if the original equations are integrated faithfully. Correctness concerns about the invalid transform and the sign of K are noted separately from the circularity assessment.

Assumptions & free parameters 6 free parameters · 5 assumptions · 4 invented entities

The paper relies on hand-chosen parameters (E, eta, sigma, A, noise amplitude, transfer entropy settings), unstated mathematical assumptions in the stability derivation, and new named objects (memory field, memory towers, memory engine and superinformal state) without an independent measurement protocol. The most consequential input is the exponential-memory ansatz together with the unvalidated viscoelastic mapping used to connect parameters to substrate stiffness.

free parameters (6)
  • feedback strength kappa (via substrate stiffness E) = E swept from 1 to 30; coherence peak near E approximately 7.2
    Chosen by hand over a range to locate the coherence island; the central result depends on this scan.
  • unified memory decay rate alpha = 1/tau_s = 1/tau_m = alpha approximately E/eta with eta fixed at 5; alpha approximately 1.44 at the coherence point
    Set by the viscoelastic mapping and used in the stability threshold; not independently measured.
  • imprint width sigma = sigma = R (particle radius)
    Chosen as a modeling convenience; no sensitivity analysis is reported.
  • deposition amplitude A = not specified
    Controls the memory field magnitude but its numerical value, and any dependence on it, are not reported.
  • noise amplitude = not specified
    Gaussian noise strength is not given; the diffusion-to-coherence transition cannot be reproduced without it.
  • transfer entropy estimation parameters = not specified
    Embedding dimensions l, k and binning are not stated, so the transfer entropy peak at E = 7 is not reproducible.
assumptions (5)
  • domain assumption Exponential kernels Theta_s and Theta_m are normalized memory kernels with characteristic times tau_s, tau_m.
    The entire model and the claimed physical mapping depend on exponential memory; no derivation from a viscoelastic constitutive law is given (Sec. II B-C).
  • domain assumption The viscoelastic mapping tau_s approximately eta/E, kappa approximately E*sigma^2 holds.
    Asserted in Sec. II C by analogy with Maxwell relaxation and microrheology; not validated for this model or used to derive an experimental prediction.
  • ad hoc to paper A straight-line trajectory with constant speed v0 is an admissible base state for transverse stability analysis.
    The noise-free base state and the choice of v0 are introduced ad hoc; the relationship between v0 and the model parameters is unspecified.
  • ad hoc to paper The Gaussian imprint curvature K in Eq. (14) may be evaluated at the origin and is positive so that alpha_c = 1/K makes sense.
    Direct differentiation of G_sigma gives a negative transverse curvature; the paper does not define K's sign or justify the origin evaluation, yet the threshold depends on it.
  • domain assumption Simulation on a periodic grid with dt = 1 faithfully represents the continuum model.
    No grid-convergence or time-step-convergence tests are reported (Sec. III A).
invented entities (4)
  • scalar memory field S(r,t)
    purpose: Records the particle's Gaussian imprints and exerts gradient forces, turning the substrate into an active memory store.
    The field is a model construct; no direct experimental measurement of S in this system is proposed, although it is inspired by wave fields in walking droplets.
  • memory towers
    purpose: Localized regions where the trajectory repeatedly enters and stalls, used to explain burst-trap cycles.
    Descriptive feature of specific simulated trajectories; no falsifiable experimental handle is given (Sec. III G).
  • memory engine and superinformal state
    purpose: Conceptual labels for the proposed self-organizing feedback regime.
    Not formally defined and not tied to a measurable prediction; the superinformal state appears once with no equations.
  • continuous-space Coupled Memory Graph Process (CMGP)
    purpose: The framework being introduced; organizes the model and its interpretation.
    This is the paper's own prior framework [17]; the continuous-space version is the paper itself, so no external support is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The Memory Engine: Self-Organized Coherence from Internal Feedback." pith.science (2026). https://pith.science/paper/CVTCKA7N

@misc{pith2026250521711,
  author       = {Pith},
  title        = {Pith review of: The Memory Engine: Self-Organized Coherence from Internal Feedback},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CVTCKA7N}},
  note         = {Machine review of arXiv:2505.21711}
}
abstract

We present a continuous-space realization of the Coupled Memory Graph Process (CMGP), a minimal non-Markovian framework in which coherence emerges through internal feedback. A single Brownian particle evolves on a viscoelastic substrate that records its trajectory as a scalar memory field and exerts local forces via the gradient $\nabla$ of accumulated imprints. This autonomous, closed-loop dynamics generates structured, phase-locked motion without external forcing. The system is governed by coupled integro-differential equations: the memory field evolves as a spatiotemporal convolution of the particle's path, while its velocity responds to the gradient of this evolving field. Simulations reveal a sharp transition from unstructured diffusion to coherent burst-trap cycles, controlled by substrate stiffness and marked by multimodal speed distributions, directional locking, and spectral entrainment. This coherence point aligns across three axes: (i) saturation of memory energy, (ii) peak transfer entropy, and (iii) a bifurcation in transverse stability. We interpret this as the emergence of a \textit{memory engine} -- a self-organizing mechanism converting stored memory into predictive motion -- illustrating that coherence arises not from tuning, but from coupling.

Figures

Figures reproduced from arXiv: 2505.21711 by the authors.

Figure 1
Figure 1. FIG. 1. Schematic of surface-mediated memory and feedback [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 4
Figure 4. FIG. 4 [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figures from the paper (3 more)
Figure 5
Figure 5. Figure 5: FIG. 5 [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6 [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: illustrates this transition. Panel (a) shows a numerically computed phase diagram in the (E, η) plane, classifying each parameter pair as stable or unstable de￾pending on whether δy(t) grows or decays. The boundary aligns closely with the analytical prediction αc = E/η…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

32 extracted references · 31 canonical work pages

  1. [1]

    An energy landscape, where memory injection and dissipation reach a steady balance

    work page

  2. [2]

    An entropy landscape, where directional transfer entropy peaks, marking maximal feedback flow

    work page

  3. [3]

    Together, these features position CMGP as a gen- eral framework for modeling self-organization in memory- active systems

    A stability landscape, where straight trajectories destabilize and give way to phase-locked motion. Together, these features position CMGP as a gen- eral framework for modeling self-organization in memory- active systems. While the present work focuses on a sin- gle particle coupled to a scalar field, the formulation read- ily extends to multiple interact...

    work page

  4. [4]

    Stochastic resonance

    Luca Gammaitoni, Peter H¨ anggi, Peter Jung, and Fabio Marchesoni. Stochastic resonance. Reviews of Modern Physics, 70(1):223–287, 1998

    work page 1998

  5. [5]

    Synchronization: A Universal Concept in Non- linear Sciences

    Arkady Pikovsky, Michael Rosenblum, and J¨ urgen Kurths. Synchronization: A Universal Concept in Non- linear Sciences. Cambridge University Press, Cambridge, UK, 2001

    work page 2001

  6. [6]

    Effects of noise in excitable systems

    Benjamin Lindner, Jordi Garc ´ ıa-Ojalvo, Alexander Neiman, and Lutz Schimansky-Geier. Effects of noise in excitable systems. Physics Reports , 392(6):321–424, 2004

    work page 2004

  7. [7]

    Eric Lauga and Thomas R. Powers. The hydrodynam- ics of swimming microorganisms. Reports on Progress in Physics, 72(9):096601, 2009

    work page 2009

  8. [8]

    Camley and Wouter-Jan Rappel

    Brian A. Camley and Wouter-Jan Rappel. Physical mod- els of collective cell motility: from cell to tissue. Journal of Physics D: Applied Physics , 50(11):113002, 2017

    work page 2017

Show all 32 references
  1. [9]

    Programming soft robots with flexible mechanical meta- materials

    Ahmad Rafsanjani, Arda Akcay, and Cagdas Daraio. Programming soft robots with flexible mechanical meta- materials. Science Robotics, 4(34):eaav7874, 2019

    2019

  2. [10]

    Single-particle diffrac- tion and interference at a macroscopic scale

    Yves Couder and Emmanuel Fort. Single-particle diffrac- tion and interference at a macroscopic scale. Physical Review Letters, 97(15):154101, 2006

    2006

  3. [11]

    Path-memory induced quantization of classical orbits

    Emmanuel Fort, Antonin Eddi, Arezki Boudaoud, Julien Moukhtar, and Yves Couder. Path-memory induced quantization of classical orbits. Proceedings of the Na- tional Academy of Sciences , 107(41):17515–17520, 2010

    2010

  4. [12]

    Self-organization into quantized eigenstates of a classical wave-driven particle

    St´ ephane Perrard, Matthieu Labousse, Marc Miskin, Emmanuel Fort, and Yves Couder. Self-organization into quantized eigenstates of a classical wave-driven particle. Nature Communications, 5:3219, 2014

    2014

  5. [13]

    Overload wave-memory induces amnesia of a self-propelled particle

    Maxime Hubert, St´ ephane Perrard, Nicolas Vandewalle, and Matthieu Labousse. Overload wave-memory induces amnesia of a self-propelled particle. Nature Communica- tions, 13(1):4357, 2022

    2022

  6. [14]

    M. C. Cross and P. C. Hohenberg. Pattern forma- tion outside of equilibrium. Reviews of Modern Physics , 65(3):851–1112, 1993

    1993

  7. [15]

    Wiley-VCH, 2008

    Eckehard Sch¨ oll and Heinz Georg Schuster.Handbook of Chaos Control. Wiley-VCH, 2008

    2008

  8. [16]

    The mechanics and statistics of active matter

    Sriram Ramaswamy. The mechanics and statistics of active matter. Annual Review of Condensed Matter Physics, 1:323–345, 2010

    2010

  9. [17]

    R. Kubo. The fluctuation-dissipation theorem. Reports on Progress in Physics , 29(1):255, 1966

    1966

  10. [18]

    Angelini, Edouard Hannezo, Xavier Trepat, Miguel Marquez, Jeffrey J

    Thomas E. Angelini, Edouard Hannezo, Xavier Trepat, Miguel Marquez, Jeffrey J. Fredberg, and David A. Weitz. Glass-like dynamics of collective cell migra- tion. Proceedings of the National Academy of Sciences , 108(12):4714–4719, 2011

    2011

  11. [19]

    Learning self-driven collective dynamics with graph net- works

    Rui Wang, Feiteng Fang, Jiamei Cui, and Wen Zheng. Learning self-driven collective dynamics with graph net- works. Scientific Reports, 12(1):500, 2022

    2022

  12. [20]

    A non-markovian route to coherence in heterogeneous diffusive systems, 2025

    Aranyak Sarkar. A non-markovian route to coherence in heterogeneous diffusive systems, 2025

    2025

  13. [21]

    Efficiency of an au- tonomous, dynamic information engine operating on a single active particle

    Lorenzo Cocconi and Keyan Chen. Efficiency of an au- tonomous, dynamic information engine operating on a single active particle. Physical Review E , 109(4):044116, 2024

    2024

  14. [22]

    Gripenberg, S.-O

    G. Gripenberg, S.-O. Londen, and O. Staffans. Volterra Integral and Functional Equations. Cambridge University Press, 1990

    1990

  15. [23]

    Birkh¨ auser, 1993

    Jan Pr¨ uss.Evolutionary Integral Equations and Applica- tions. Birkh¨ auser, 1993

    1993

  16. [24]

    Nonequilibrium Statistical Mechanics

    Robert Zwanzig. Nonequilibrium Statistical Mechanics . Oxford University Press, 2001

    2001

  17. [25]

    Kupferman

    R. Kupferman. Fractional kinetics in kac–zwanzig heat bath models. Journal of Statistical Physics , 114:291–326, 2004

    2004

  18. [26]

    J. D. Ferry. Viscoelastic Properties of Polymers . Wiley, 1980

    1980

  19. [27]

    T. G. Mason and D. A. Weitz. Optical measurements of frequency-dependent linear viscoelastic moduli of com- plex fluids. Physical Review Letters , 74(7):1250–1253, 1995

    1995

  20. [28]

    T. A. Waigh. Microrheology of complex fluids. Reports on Progress in Physics , 68(3):685–742, 2005

    2005

  21. [29]

    J. W. M. Bush. The new wave of pilot-wave theory. Physics Today, 68(8):47–52, 2015

    2015

  22. [30]

    Measuring information transfer

    Thomas Schreiber. Measuring information transfer. Physical Review Letters, 85(2):461–464, 2000

    2000

  23. [31]

    Strogatz

    Steven H. Strogatz. Nonlinear Dynamics and Chaos: With Applications to Physics, Biology, Chemistry, and Engineering. CRC Press, 2nd edition, 2018

    2018

  24. [32]

    Bobet and H

    A. Bobet and H. H. Einstein. Fracture coalescence in rock-type materials under uniaxial and biaxial compres- sion. International Journal of Rock Mechanics and Min- ing Sciences, 35(7):863–888, 1998

    1998

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.