Pith. sign in

REVIEW 2 major objections 4 minor 82 references

Optimizing optimal transport: Role of final distributions in finite-time thermodynamics

T0 review · 2 major / 4 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read A variational framework shows that relaxing the final distribution lowers the minimal thermodynamic cost of finite-time tasks in overdamped Langevin systems, with the optimal final distribution determined by Lagrange multiplier equations ov

desk verdict Clean variational framework for optimizing final distributions, but the measurement/feedback results are Gaussian-ansatz optima, not proven global optima. read the letter →

arxiv 2509.11314 v2 pith:5ZPBAN5D submitted 2025-09-14 cond-mat.stat-mech

Kaito Tojo , Rihito Nagase , Ken Funo , Takahiro Sagawa This is my paper
classification cond-mat.stat-mech MSC 49Q2282C3160J70 PACS 05.70.Ln05.40.-a
keywords optimaltransportstochasticthermodynamicsfinite-timeprocessesentropyproductionLagrangemultipliersfinaldistributionoptimizationmeasurementandfeedbackthermalsqueezing
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

The paper argues that in finite-time stochastic thermodynamics, fixing the final distribution—as standard optimal transport theory does—is unnecessary and wasteful. Any task that can be expressed as an equality constraint on a functional of the final distribution leaves many admissible final distributions, and the paper proposes to optimize over them using Lagrange multipliers, yielding variational equations whose solution gives the true optimal transport map, the optimal final distribution, and the optimal protocol. The authors show that this unified framework reproduces known results for particle transport, information erasure, and quadratic-potential work optimization, while producing new analytic results for thermal squeezing with non-Gaussian distributions and for measurement and feedback processes. If correct, this provides design principles for fast, energy-efficient information-processing devices and thermodynamic machines.

What carries the argument

The key object is the transport map T that pushes the initial distribution p_ini forward to the final distribution p_fin. The paper replaces the standard fixed-final-distribution optimal transport problem with a variational problem over T, using Lagrange multipliers λ to enforce the task constraint A[T]=A_f. The central identity is the variation formula δA[T]/δT_i = ∂_{r_j}{ĝ(T(r))} J̃_{ji}(r) p_ini(r) (Eq. (22)), where ĝ is the derivative of the task functional's density with respect to the final density. This reduces the constrained optimization to solving a system of algebraic equations for T and λ, from which the optimal protocol follows via the Benamou–Brenier velocity field.

What would settle it

Construct a non-Gaussian joint distribution for (X,Y) with mutual information I_f = ln 2 and compare its required partial entropy production (obtained by numerically solving Eq. (41) without the Gaussian ansatz) against the Gaussian result (43). If a non-Gaussian solution gives a lower cost than Ξ_ini^YY ρ_f²/(Dτ), the optimality claim fails.

Watch

Extended reading notes

Core claim

The central claim is that the minimum thermodynamic cost for a task specified by a constraint A[p_fin]=A_f is obtained by solving the variational equations δC̃_τ/δT + λ·δA/δT = 0 and A[T]=A_f, where T is the optimal transport map from the initial to the final distribution, C̃_τ[T] is the cost functional (entropy production, work, or partial entropy production), and A[T] is the task functional evaluated after transport. The paper derives the general variation of any final-distribution functional (Eq. (22)), applies it to several tasks, and obtains closed-form optimal final distributions and minimal costs, including the analytic Gaussian results for measurement (Eqs. (42)–(43)) and feedback (E

Load-bearing premise

For the measurement and feedback results, the paper assumes the optimal final distributions are Gaussian (Appendix B); if a non-Gaussian final distribution with the same mutual information yields lower partial entropy production, the stated costs are upper bounds rather than true optima.

Editorial extensions

If this is right

  • If the framework is correct, any thermodynamic task that can be expressed as an equality constraint on a final-distribution functional has an explicit variational characterization of its minimal finite-time cost.
  • The optimal final distribution for thermal squeezing is generally non-Gaussian, and its minimal entropy production is given by Eq. (33), which extends prior Gaussian-only results.
  • For measurement and feedback, the optimal final distribution and minimal partial entropy production are given by Eqs. (42)–(45), enabling quantitative design of finite-time information processors.
  • The framework unifies and generalizes earlier results on optimal finite-time erasure and work in quadratic potentials, showing that these are special cases of the same variational scheme.
  • The method applies equally to entropy production, work, and partial entropy production, covering both conservative and nonconservative driving, and extends naturally to far-from-equilibrium regimes.

Reading between the lines

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

  • A direct testable extension is to numerically solve the full variational equation (41) for measurement in non-Gaussian settings and check whether a non-Gaussian final distribution with the same mutual information I_f achieves a lower partial entropy production than the Gaussian-ansatz result (43).
  • The framework suggests that any 'thermodynamic speed limit' that fixes both endpoints may be systematically tightened by optimizing the terminal distribution; a similar hierarchy might exist for Markov jump processes, where the optimal transport formulation is less straightforward.
  • The variational equation for mutual information in measurement is nonlinear and may admit multiple solutions; the paper's Gaussian ansatz picks one branch, but other branches could correspond to genuinely different—possibly better—protocols, a point the paper leaves open.
  • The result for free-energy control implies that the minimal entropy production to reach a target nonequilibrium free energy may be significantly lower than the cost of reaching any specific final state, a fact relevant for designing work-storage devices.
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

2 major / 4 minor

Summary. The paper develops a unified variational framework for finite-time thermodynamic optimization in overdamped Langevin systems. Instead of fixing the final distribution, as in the standard Benamou–Brenier/Wasserstein approach, the authors impose task-specific equality constraints A[p_fin]=A_f and minimize the thermodynamic cost over both the transport map and the final distribution. The central result is the Lagrange multiplier system in Eqs. (14)–(15), together with the general variation formula Eq. (22) for functionals of the final distribution. The framework is applied to particle transport, thermal squeezing, information erasure, work optimization in quadratic potentials, free-energy control, and measurement/feedback. For measurement and feedback, explicit formulas are obtained for Gaussian final distributions, Eqs. (42)–(45), using the ansatz in Appendix B.

Significance. If the main claims hold, the paper provides a useful unified treatment of finite-time task optimization that goes beyond fixing the final distribution, and it generalizes earlier results on erasure [39,40] and quadratic-potential work [41]. The derivation of the variational equations from first principles is clean, and the thermal-squeezing result is genuinely non-Gaussian and globally justified by the Gelbrich/Wasserstein lower bound. The general variation formula Eq. (22) is a valuable technical contribution. The information-processing applications are potentially significant, but their current proof of optimality is restricted to a Gaussian ansatz; without a global optimality argument, Eqs. (42)–(45) are only proven optimal within that family. The paper is therefore not yet at the level of its advertised 'truly optimal' claims for measurement and feedback.

major comments (2)
  1. [Appendix B and Sec. IV F, Eqs. (42)–(45)] The measurement and feedback results are derived under an explicit Gaussian restriction: 'We restrict the final distributions to be Gaussian' (Appendix B1) and the ansatz forms (B3), (B13) for the conditional transport maps. Equation (41) is only a first-order stationarity condition for a problem whose constraint I_fin=I_f is a level set of mutual information; such level sets are generally non-convex, and stationarity does not imply global optimality. No uniqueness, second-order, or non-Gaussian check is provided. Figures 3 and 4 compare only Gaussian candidates, so they do not test the ansatz. Consequently, Eqs. (42)–(45) are currently proven optimal only within the Gaussian family; if a non-Gaussian final distribution with the same mutual information has lower partial entropy production, these formulas are upper bounds rather than the claimed minima. This affects the load-bearing 'trul
  2. [Sec. IV F, Eq. (41) and stationarity] Even within the Gaussian family, the paper solves only the first-order condition (41) after postulating the form of the conditional map. For a non-convex constrained problem, multiple stationary points can exist, and the solved point could be a saddle point or local maximum. The authors should at least verify that the solution satisfies a second-order sufficient condition, or provide independent evidence that the Gaussian solution is the global minimizer. This is a separate, more restricted concern, but it reinforces the need to calibrate the optimality claims for measurement and feedback.
minor comments (4)
  1. [Eq. (18)] Equation (18) has 'F_fin[T] = β ∫ dr ...'; from the preceding definition F = β^{-1} D + F_eq, the prefactor should be β^{-1}. The following equation (19) uses β^{-1}, so this appears to be a typo.
  2. [Eq. (B13)] In Eq. (B13), the last term is written as (Ξ^{XY}_{fin}/Ξ^{YY}) x, but the conditional mean of X|Y should involve y, not x. Please check whether this is a typo; the same applies to the surrounding discussion.
  3. [Appendix A3] Equation (A23) applies the local formula (22) to the mutual information, which is a nonlocal functional of p_fin. The derivation is valid because Eq. (A21) reduces δI_fin to ∫ (δp_fin) ι_fin, but the text should state this explicitly. The current wording 'Using Eq. (22), we have' is too terse for a reader checking the validity of Eq. (41).
  4. [Sec. IV F, Eq. (41)] The index notation in Eq. (41) is ambiguous: T^Y_{meas,i} and r_i should be specified to run only over the Y components, since the X component of the cost variation is identically zero. A short sentence clarifying the index range would improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central variational derivation is self-contained; the Gaussian ansatz is a rigor gap, not a circular reduction.

full rationale

The core framework is derived in-paper from the Benamou-Brenier formula and calculus of variations. The optimization problem is reduced to minimizing \tilde{C}_\tau[T] subject to A[T]=A_f (Eqs. 11-12), and the first-order conditions (14)-(15) are obtained from the Lagrangian (13), not imported from a fit or from the desired result. The variation of the cost, Eq. (17), and the general variation of a final-distribution functional, Eq. (22), are derived in Appendix A from the push-forward relation (7) and Jacobi's formula. Applications are solved directly from these variational equations; where results match earlier work (e.g., Refs. [39-41,68]), those are external benchmark checks, not fitted inputs. The measurement/feedback results do cite the same group's prior paper [57] for the conditional-transport reduction in Eq. (37), but that is a parameter-free mathematical theorem with its own independent content, not an unverified premise unique to this paper, so it does not make the derivation circular. The Gaussian restriction in Appendix B ('We restrict the final distributions to be Gaussian' and 'We also adopt an ansatz') is an explicit limitation on the search space: it may mean Eqs. (42)-(45) are only optimal within the Gaussian family, and this is a potential correctness/optimality gap, not a reduction of the claimed result to its own inputs by construction. No fitted parameter is relabeled as a prediction, and no uniqueness claim is imported solely from the authors' prior work to force the conclusion. Therefore no significant circularity is present.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

No parameters are fitted to data. Task targets (mu_f, Xi_f, I_f) and Lagrange multipliers are inputs or are determined by constraints; the Gaussian ansatz parameters are solved from the variational equations, not tuned. No new physical entities are introduced; the optimal transport map and Lagrange multipliers are mathematical constructions.

assumptions (6)
  • domain assumption The minimum entropy production for fixed initial and final distributions equals W2(pini,pfin)^2/(D tau) (Eq. 5).
    Standard result for overdamped Langevin systems with fully controllable forces; used as the basis for reducing protocol optimization to transport-map optimization.
  • domain assumption Conservative and nonconservative forces are fully controllable (Sec. II).
    The optimal protocol velocity (Eq. 16) and the equality in Eq. (5) require this controllability; the conclusion section notes that complete control of nonconservative forces is demanding experimentally.
  • domain assumption Thermodynamic tasks are representable as equality constraints A[p_fin]=A_f on functionals of the final distribution (Sec. III A).
    The whole variational framework is built on this encoding; tasks such as erasure, squeezing, and measurement are assumed to be exactly characterized by such constraints.
  • standard math The transport map T is invertible with nonzero Jacobian where the variational formulas are applied (Appendix A2).
    Variations of T^{-1} and det J_T are used to derive Eq. (22); the paper does not state explicit regularity conditions on p_ini and T.
  • ad hoc to paper For measurement and feedback, the optimal final distributions are Gaussian and the conditional transport maps take the specified forms (Appendix B1, B2).
    This ansatz is imposed to make the variational equation (41) solvable analytically; it is not derived from optimality, so the resulting cost may be only an upper bound among all distributions.
  • domain assumption Ideal measurement and feedback assume the untouched marginal is inert (v_X=0 or v_Y=0) for the whole process (Sec. IV F 1).
    This makes the partial entropy production of that subsystem zero and defines the cost; it is a standard idealization in information thermodynamics.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Optimizing optimal transport: Role of final distributions in finite-time thermodynamics." pith.science (2026). https://pith.science/paper/5ZPBAN5D

@misc{pith2026250911314,
  author       = {Pith},
  title        = {Pith review of: Optimizing optimal transport: Role of final distributions in finite-time thermodynamics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5ZPBAN5D}},
  note         = {Machine review of arXiv:2509.11314}
}
read the original abstract

Performing thermodynamic tasks within finite time while minimizing thermodynamic costs is a central challenge in stochastic thermodynamics. Here, we develop a unified framework for optimizing the thermodynamic cost of performing various tasks in finite time for overdamped Langevin systems. Conventional optimization of thermodynamic cost based on optimal transport theory leaves room for varying the final distributions according to the intended task, enabling further optimization. Taking advantage of this freedom, we use Lagrange multipliers to derive the optimal final distribution that minimizes the thermodynamic cost. Our framework applies to a wide range of thermodynamic tasks, including particle transport, thermal squeezing, and information processing such as information erasure, measurement, and feedback. Our results are expected to provide design principles for information-processing devices and thermodynamic machines that operate at high speed with low energetic costs.

Figures

Figures reproduced from arXiv: 2509.11314 by the authors.

Figure 1
Figure 1. FIG. 1. Summary of the main results and their applications. (a) Schematic illustration indicating that further optimization is [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Non-Gaussian example of thermodynamically optimal thermal squeezing in finite time. (a) Initial distribution before [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. An example of thermodynamically optimal finite-time measurement. (a) Schematic illustration of measurement process. [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: FIG. 4. An example of thermodynamically optimal finite-time feedback. (a) Schematic illustration of feedback process. The [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

82 extracted references · 1 canonical work pages

  1. [41]

    Oikawa, Y

    S. Oikawa, Y. Nakayama, S. Ito, T. Sagawa, and S. Toyabe, Experimentally achieving minimal dissipa- tion via thermodynamically optimal transport (2025), arXiv:2503.01200 [cond-mat.stat-mech]

    arXiv 2025

  2. [1]

    We consider a bipartite systemXY, withX being the system of interest andYthe memory

    Setup for measurement and feedback We now introduce the setup of information thermo- dynamics. We consider a bipartite systemXY, withX being the system of interest andYthe memory. The sub- systemsXandYare eachd-dimensional, and the total system decomposes asr= (r X ,r Y )⊤, wherer X andr Y are the position of the subsystemXandY, respectively. In the measu...

  3. [2]

    dWzyXotTeR5ckqSQheWlLrS7gf0=

    Measurement We consider a process in which the subsystemYmea- suresXwithin finite timeτto generate more mutual information than the initial distributionI f > Iini. Here, we take the thermodynamic cost asC τ = Σ Y τ |ΣXτ =0 ( ˜Cτ = ˜ΣY τ,meas[T Y|X meas]), and the constraint on the final distributions to beI fin[T Y|X meas] =I f . We have to obtain the var...

  4. [3]

    wlMIVkfyPZwM0K+/eFJ/I47p1b0=

    Feedback We consider a process in which the subsystemYper- forms feedback onXin finite timeτto consume an amountI ini −I f >0 of mutual information. Here, we take the thermodynamic cost asC τ = Σ X τ |ΣYτ =0 ( ˜Cτ = ˜ΣX τ,fb[T X|Y fb ]). The variational equation to be solved is obtained from Eq. (41) by exchangingXand Y. Figure 4 shows an example of therm...

  5. [4]

    (20), we need the vari- ation of the nonequilibrium free energy in Eq

    V ariation of free energy To obtain the variational formula of the work with re- spect to the transport map in Eq. (20), we need the vari- ation of the nonequilibrium free energy in Eq. (19). In this subsection, we give the details of that derivation. For the statep fin and the potentialV τ at timet=τ, the nonequilibrium free energy is Ffin =β −1D(pfin||p...

  6. [5]

    V ariation of functional of final distributions We give the details of the derivation of the variation of Eq. (22). We have δ Z dr g(pfin(r),r) = Z dr ∂g(X,r) ∂X X=p fin(r) δpfin(r), (A7) where, using Eq. (7), δpfin(r) =δ pini(T −1(r)) detJ T (T −1(r)) .(A8) We then need to expressδT −1(r) as a variation with respect toT(r). In Sec. A 2 a below, we expres...

  7. [6]

    meas” or “fb

    V ariation of mutual information In this subsection, we provide the details of the deriva- tion of Eq. (A23). We omit the subscript “meas” or “fb” for simplicity. Variation of the mutual information at final time is given by δIfin =δ Z drp XY fin (r) ln pXY fin (r) pX fin(rX )pY fin(rY ) = Z dr δpXY fin (r) ιfin(r) + Z drδp XY fin (r) +δp X fin(rX ) +δp f...

  8. [7]

    Since the measurement does not change the marginal distri- bution ofX, the mean and variance ofxare fixed as µX t = 0 and Ξ XX t = Ξ XX for allt∈[0, τ]

    Measurement As the initial distribution, we take pXY ini (x, y) = 1 2π p ΞXX ΞY Y ini exp − x2 2ΞXX − y2 2ΞY Y ini , (B2) which has zero mean and zero mutual information. Since the measurement does not change the marginal distri- bution ofX, the mean and variance ofxare fixed as µX t = 0 and Ξ XX t = Ξ XX for allt∈[0, τ]. We restrict the final distributio...

Show all 82 references
  1. [8]

    Since an ideal feedback does not change the marginal distribution ofY, the mean and variance ofyare fixed asµ Y t = 0 and Ξ Y Y t = Ξ Y Y for allt∈[0, τ]

    F eedback As the initial distribution, we take pXY ini (x, y) = 1r 2π ΞXX ini ΞY Y− ΞXY ini 2 exp − x y ΞXX ini ΞXY ini ΞXY ini ΞY Y −1 x y ! ,(B12) which has zero mean and finite mutual information if Ξ XY ini >0. Since an ideal feedback does not change the marginal distribut...

  2. [9]

    Villani,Optimal Transport: Old and New, Vol

    C. Villani,Optimal Transport: Old and New, Vol. 338 (Springer, 2009)

    2009

  3. [10]

    Santambrogio,Optimal Transport for Applied Mathe- maticians: Calculus of Variations, PDEs, and Modeling (Birkh¨ auser Cham, 2015)

    F. Santambrogio,Optimal Transport for Applied Mathe- maticians: Calculus of Variations, PDEs, and Modeling (Birkh¨ auser Cham, 2015)

    2015

  4. [11]

    Peyr´ e and M

    G. Peyr´ e and M. Cuturi, Computational optimal trans- port: With applications to data science, Foundations and Trends®in Machine Learning11, 355 (2019)

    2019

  5. [12]

    Arjovsky, S

    M. Arjovsky, S. Chintala, and L. Bottou, Wasserstein generative adversarial networks, inProceedings of the 34th International Conference on Machine Learning, Proceedings of Machine Learning Research, Vol. 70, edited by D. Precup and Y. W. Teh (PMLR, 2017) pp. 214–223

    2017

  6. [13]

    Rubner, C

    Y. Rubner, C. Tomasi, and L. J. Guibas, The earth mover’s distance as a metric for image retrieval, Inter- national Journal of Computer Vision40, 99 (2000)

    2000

  7. [14]

    Schiebinger, J

    G. Schiebinger, J. Shu, M. Tabaka, B. Cleary, V. Sub- ramanian, A. Solomon, J. Gould, S. Liu, S. Lin, P. Berube, L. Lee, J. Chen, J. Brumbaugh, P. Rigollet, K. Hochedlinger, R. Jaenisch, A. Regev, and E. S. Lan- der, Optimal-transport analysis of single-cell gene expres- sion ...

    2019

  8. [15]

    Sekimoto,Stochastic Energetics, Vol

    K. Sekimoto,Stochastic Energetics, Vol. 799 (Springer Berlin, Heidelberg, 2010)

    2010

  9. [16]

    Seifert, Stochastic thermodynamics, fluctuation the- orems and molecular machines, Reports on Progress in Physics75, 126001 (2012)

    U. Seifert, Stochastic thermodynamics, fluctuation the- orems and molecular machines, Reports on Progress in Physics75, 126001 (2012)

    2012

  10. [17]

    Luca and P

    P. Luca and P. Simone,Stochastic Thermodynamics: An Introduction(Princeton University Press, 2021)

    2021

  11. [18]

    Shiraishi,An Introduction to Stochastic Thermody- namics(Springer Singapore, 2023)

    N. Shiraishi,An Introduction to Stochastic Thermody- namics(Springer Singapore, 2023)

    2023

  12. [19]

    Ciliberto, Experiments in stochastic thermodynamics: Short history and perspectives, Phys

    S. Ciliberto, Experiments in stochastic thermodynamics: Short history and perspectives, Phys. Rev. X7, 021051 (2017)

    2017

  13. [20]

    Shiraishi, K

    N. Shiraishi, K. Saito, and H. Tasaki, Universal trade-off relation between power and efficiency for heat engines, Phys. Rev. Lett.117, 190601 (2016)

    2016

  14. [21]

    Van den Broeck, Thermodynamic efficiency at maxi- mum power, Phys

    C. Van den Broeck, Thermodynamic efficiency at maxi- mum power, Phys. Rev. Lett.95, 190602 (2005)

    2005

  15. [22]

    Schmiedl and U

    T. Schmiedl and U. Seifert, Efficiency at maximum power: An analytically solvable model for stochastic heat engines, Europhysics Letters81, 20003 (2007)

    2007

  16. [23]

    Esposito, R

    M. Esposito, R. Kawai, K. Lindenberg, and C. Van den Broeck, Efficiency at maximum power of low-dissipation carnot engines, Phys. Rev. Lett.105, 150603 (2010)

    2010

  17. [24]

    Schmiedl and U

    T. Schmiedl and U. Seifert, Optimal finite-time pro- 16 cesses in stochastic thermodynamics, Phys. Rev. Lett. 98, 108301 (2007)

    2007

  18. [25]

    G. E. Crooks, Measuring thermodynamic length, Phys. Rev. Lett.99, 100602 (2007)

    2007

  19. [26]

    D. A. Sivak and G. E. Crooks, Thermodynamic metrics and optimal paths, Phys. Rev. Lett.108, 190602 (2012)

    2012

  20. [27]

    Blaber and D

    S. Blaber and D. A. Sivak, Optimal control in stochastic thermodynamics, Journal of Physics Communications7, 033001 (2023)

    2023

  21. [28]

    Shiraishi, K

    N. Shiraishi, K. Funo, and K. Saito, Speed limit for clas- sical stochastic processes, Phys. Rev. Lett.121, 070601 (2018)

    2018

  22. [29]

    Ito, Stochastic thermodynamic interpretation of infor- mation geometry, Phys

    S. Ito, Stochastic thermodynamic interpretation of infor- mation geometry, Phys. Rev. Lett.121, 030605 (2018)

    2018

  23. [30]

    Ito and A

    S. Ito and A. Dechant, Stochastic time evolution, infor- mation geometry, and the cram´ er-rao bound, Phys. Rev. X10, 021056 (2020)

    2020

  24. [31]

    A. C. Barato and U. Seifert, Thermodynamic uncertainty relation for biomolecular processes, Phys. Rev. Lett.114, 158101 (2015)

    2015

  25. [32]

    T. R. Gingrich, J. M. Horowitz, N. Perunov, and J. L. England, Dissipation bounds all steady-state current fluctuations, Phys. Rev. Lett.116, 120601 (2016)

    2016

  26. [33]

    J. M. Horowitz and T. R. Gingrich, Thermodynamic uncertainty relations constrain non-equilibrium fluctua- tions, Nature Physics16, 15 (2020)

    2020

  27. [34]

    D. H. Wolpert, J. Korbel, C. W. Lynn, F. Tasnim, J. A. Grochow, G. Karde¸ s, J. B. Aimone, V. Balasub- ramanian, E. D. Giuli, D. Doty, N. Freitas, M. Marsili, T. E. Ouldridge, A. W. Richa, P. Riechers, ´E. Rold´ an, B. Rubenstein, Z. Toroczkai, and J. Paradiso, Is stochastic t...

    2024 doi

  28. [35]

    S. Dago, J. Pereda, N. Barros, S. Ciliberto, and L. Bel- lon, Information and thermodynamics: Fast and precise approach to landauer’s bound in an underdamped mi- cromechanical oscillator, Phys. Rev. Lett.126, 170601 (2021)

    2021

  29. [36]

    Scandi, D

    M. Scandi, D. Barker, S. Lehmann, K. A. Dick, V. F. Maisi, and M. Perarnau-Llobet, Minimally dissipative in- formation erasure in a quantum dot via thermodynamic length, Phys. Rev. Lett.129, 270601 (2022)

    2022

  30. [37]

    S. Dago, S. Ciliberto, and L. Bellon, Adiabatic computing for optimal thermodynamic efficiency of information processing, Proceedings of the Na- tional Academy of Sciences120, e2301742120 (2023), https://www.pnas.org/doi/pdf/10.1073/pnas.2301742120

    2023 doi

  31. [38]

    Barros, S

    N. Barros, S. Whitelam, S. Ciliberto, and L. Bellon, Learning efficient erasure protocols for an underdamped memory, Phys. Rev. E111, 044114 (2025)

    2025

  32. [39]

    Ma, R.-X

    Y.-H. Ma, R.-X. Zhai, J. Chen, C. P. Sun, and H. Dong, Experimental test of the 1/τ-scaling entropy generation in finite-time thermodynamics, Phys. Rev. Lett.125, 210601 (2020)

    2020

  33. [40]

    S. A. M. Loos, S. Monter, F. Ginot, and C. Bechinger, Universal symmetry of optimal control at the microscale, Phys. Rev. X14, 021032 (2024)

    2024

  34. [42]

    S. Pal, S. Saryal, D. Segal, T. S. Mahesh, and B. K. Agar- walla, Experimental study of the thermodynamic uncer- tainty relation, Phys. Rev. Res.2, 022044 (2020)

    2020

  35. [43]

    Aurell, C

    E. Aurell, C. Mej ´ ıa-Monasterio, and P. Muratore- Ginanneschi, Optimal protocols and optimal transport in stochastic thermodynamics, Phys. Rev. Lett.106, 250601 (2011)

    2011

  36. [44]

    Aurell, K

    E. Aurell, K. Gaw¸ edzki, C. Mej ´ ıa-Monasterio, R. Mo- hayaee, and P. Muratore-Ginanneschi, Refined second law of thermodynamics for fast random processes, Jour- nal of Statistical Physics147, 487 (2012)

    2012

  37. [45]

    Dechant and Y

    A. Dechant and Y. Sakurai, Thermodynamic interpre- tation of wasserstein distance (2019), arXiv:1912.08405 [cond-mat.stat-mech]

    2019 arXiv

  38. [46]

    Nakazato and S

    M. Nakazato and S. Ito, Geometrical aspects of en- tropy production in stochastic thermodynamics based on wasserstein distance, Phys. Rev. Res.3, 043093 (2021)

    2021

  39. [47]

    Proesmans, J

    K. Proesmans, J. Ehrich, and J. Bechhoefer, Finite-time landauer principle, Phys. Rev. Lett.125, 100602 (2020)

    2020

  40. [48]

    Proesmans, J

    K. Proesmans, J. Ehrich, and J. Bechhoefer, Optimal finite-time bit erasure under full control, Phys. Rev. E 102, 032105 (2020)

    2020

  41. [49]

    Y. Chen, T. T. Georgiou, and A. Tannenbaum, Stochas- tic control and nonequilibrium thermodynamics: Funda- mental limits, IEEE Transactions on Automatic Control 65, 2979 (2020)

    2020

  42. [50]

    Chennakesavalu and G

    S. Chennakesavalu and G. M. Rotskoff, Unified, geomet- ric framework for nonequilibrium protocol optimization, Phys. Rev. Lett.130, 107101 (2023)

    2023

  43. [51]

    Van Vu and Y

    T. Van Vu and Y. Hasegawa, Geometrical bounds of the irreversibility in markovian systems, Phys. Rev. Lett. 126, 010601 (2021)

    2021

  44. [52]

    A. Dechant, Minimum entropy production, detailed balance and wasserstein distance for continuous-time markov processes, Journal of Physics A: Mathematical and Theoretical55, 094001 (2022)

    2022

  45. [53]

    Yoshimura, A

    K. Yoshimura, A. Kolchinsky, A. Dechant, and S. Ito, Housekeeping and excess entropy production for general nonlinear dynamics, Phys. Rev. Res.5, 013017 (2023)

    2023

  46. [54]

    Van Vu and K

    T. Van Vu and K. Saito, Thermodynamic unification of optimal transport: Thermodynamic uncertainty relation, minimum dissipation, and thermodynamic speed limits, Phys. Rev. X13, 011013 (2023)

    2023

  47. [55]

    Nagayama, K

    R. Nagayama, K. Yoshimura, A. Kolchinsky, and S. Ito, Geometric thermodynamics of reaction-diffusion sys- tems: Thermodynamic trade-off relations and optimal transport for pattern formation, Phys. Rev. Res.7, 033011 (2025)

    2025

  48. [56]

    Sagawa and M

    T. Sagawa and M. Ueda, Fluctuation theorem with in- formation exchange: Role of correlations in stochastic thermodynamics, Phys. Rev. Lett.109, 180602 (2012)

    2012

  49. [57]

    J. M. Horowitz and M. Esposito, Thermodynamics with continuous information flow, Phys. Rev. X4, 031015 (2014)

    2014

  50. [58]

    J. M. R. Parrondo, J. M. Horowitz, and T. Sagawa, Thermodynamics of information, Nature Physics11, 131 (2015)

    2015

  51. [59]

    M. L. Rosinberg and J. M. Horowitz, Continuous infor- mation flow fluctuations, Europhysics Letters116, 10007 (2016)

    2016

  52. [60]

    Taghvaei, O

    A. Taghvaei, O. M. Miangolarra, R. Fu, Y. Chen, and T. T. Georgiou, On the relation between information and power in stochastic thermodynamic engines, IEEE Con- trol Systems Letters6, 434 (2022)

    2022

  53. [61]

    Fujimoto and S

    Y. Fujimoto and S. Ito, Game-theoretical approach to 17 minimum entropy productions in information thermody- namics, Phys. Rev. Res.6, 013023 (2024)

    2024

  54. [62]

    Nagase and T

    R. Nagase and T. Sagawa, Thermodynamically optimal information gain in finite-time measurement, Phys. Rev. Res.6, 033239 (2024)

    2024

  55. [63]

    Nagase and T

    R. Nagase and T. Sagawa, Thermodynamic optimization of finite-time feedback protocols for markov jump sys- tems, Phys. Rev. E112, 024118 (2025)

    2025

  56. [64]

    Kamijima, K

    T. Kamijima, K. Funo, and T. Sagawa, Finite-time ther- modynamic bounds and trade-off relations for informa- tion processing, Phys. Rev. Res.7, 013329 (2025)

    2025

  57. [65]

    Kamijima, A

    T. Kamijima, A. Takatsu, K. Funo, and T. Sagawa, Op- timal finite-time maxwell’s demons in langevin systems, Phys. Rev. Res.7, 023159 (2025)

    2025

  58. [66]

    N. G. van Kampen,Stochastic Processes in Physics and Chemistry, 3rd ed. (North Holland, 2007)

    2007

  59. [67]

    Risken,The Fokker-Planck Equation: Methods of So- lution and Applications, 2nd ed

    H. Risken,The Fokker-Planck Equation: Methods of So- lution and Applications, 2nd ed. (Springer, 1989)

    1989

  60. [68]

    Benamou and Y

    J.-D. Benamou and Y. Brenier, A computational fluid mechanics solution to the monge-kantorovich mass trans- fer problem, Numerische Mathematik84, 375 (2000)

    2000

  61. [69]

    Kogelnik and T

    H. Kogelnik and T. Li, Laser beams and resonators, Appl. Opt.5, 1550 (1966)

    1966

  62. [70]

    Klaers, Landauer’s erasure principle in a squeezed ther- mal memory, Phys

    J. Klaers, Landauer’s erasure principle in a squeezed ther- mal memory, Phys. Rev. Lett.122, 040602 (2019)

    2019

  63. [71]

    Roßnagel, O

    J. Roßnagel, O. Abah, F. Schmidt-Kaler, K. Singer, and E. Lutz, Nanoscale heat engine beyond the carnot limit, Phys. Rev. Lett.112, 030602 (2014)

    2014

  64. [72]

    Klaers, S

    J. Klaers, S. Faelt, A. Imamoglu, and E. Togan, Squeezed thermal reservoirs as a resource for a nanomechanical en- gine beyond the carnot limit, Phys. Rev. X7, 031044 (2017)

    2017

  65. [73]

    Rugar and P

    D. Rugar and P. Gr¨ utter, Mechanical parametric amplifi- cation and thermomechanical noise squeezing, Phys. Rev. Lett.67, 699 (1991)

    1991

  66. [74]

    Rashid, T

    M. Rashid, T. Tufarelli, J. Bateman, J. Vovrosh, D. Hempston, M. S. Kim, and H. Ulbricht, Experimen- tal realization of a thermal squeezed state of levitated optomechanics, Phys. Rev. Lett.117, 273601 (2016)

    2016

  67. [75]

    Kamba, N

    M. Kamba, N. Hara, and K. Aikawa, Quantum squeez- ing of a levitated nanomechanical oscillator (2025), arXiv:2504.17944 [quant-ph]

    2025 arXiv

  68. [76]

    Gelbrich, On a formula for the L 2 wasserstein met- ric between measures on euclidean and hilbert spaces, Mathematische Nachrichten147, 185 (1990)

    M. Gelbrich, On a formula for the L 2 wasserstein met- ric between measures on euclidean and hilbert spaces, Mathematische Nachrichten147, 185 (1990)

    1990

  69. [77]

    Landauer, Information is physical, Physics Today44, 23 (1991), https://pubs.aip.org/physicstoday/article- pdf/44/5/23/8304036/23 1 online.pdf

    R. Landauer, Information is physical, Physics Today44, 23 (1991), https://pubs.aip.org/physicstoday/article- pdf/44/5/23/8304036/23 1 online.pdf

    1991

  70. [78]

    B´ erut, A

    A. B´ erut, A. Arakelyan, A. Petrosyan, S. Ciliberto, R. Dillenschneider, and E. Lutz, Experimental verifica- tion of landauer’s principle linking information and ther- modynamics, Nature483, 187 (2012)

    2012

  71. [79]

    Y. Jun, M. c. v. Gavrilov, and J. Bechhoefer, High- precision test of landauer’s principle in a feedback trap, Phys. Rev. Lett.113, 190601 (2014)

    2014

  72. [80]

    Pi˜ nero, R

    J. Pi˜ nero, R. Sol´ e, and A. Kolchinsky, Optimization of nonequilibrium free energy harvesting illustrated on bac- teriorhodopsin, Phys. Rev. Res.6, 013275 (2024)

    2024

  73. [81]

    Kolchinsky, I

    A. Kolchinsky, I. Marvian, C. Gokler, Z.-W. Liu, P. Shor, O. Shtanko, K. Thompson, D. Wolpert, and S. Lloyd, Maximizing free energy gain, Entropy27, 10.3390/e27010091 (2025)

    2025 doi

  74. [82]

    H. N. Jan R. Magnus,Matrix Differential Calculus with Applications in Statistics and Econometrics, 3rd ed. (Wi- ley, 2019)

    2019

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.