Pith. sign in

REVIEW 3 major objections 3 minor 1 cited by

Universal Reinforcement Learning in Coalgebras: Asynchronous Stochastic Computation via Conduction

T0 review · 3 major / 3 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read Reinforcement learning recast as a single coinductive problem

desk verdict The abstract promises a coalgebraic RL framework, but the supplied text is an unrelated pandemic-modeling paper, so there is nothing here to referee. read the letter →

arxiv 2508.15128 v1 pith:BYWGZYRW submitted 2025-08-20 cs.LG cs.AI

Sridhar Mahadevan This is my paper · ORCID
classification cs.LGcs.AI MSC 18B2518C2068T05
keywords universalreinforcementlearningcoalgebrafinalcoinductiontopostheoryasynchronousconvergenceMarkovdecisionprocesspredictivestaterepresentation
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

Universal reinforcement learning (URL) is proposed as a categorical generalization of RL: MDPs, POMDPs, predictive state representations, and linear dynamical systems are all coalgebras, and the space of RL algorithms is a functor category whose codomain is a topos. The classical RL problem of computing the exact or approximate (action) value function as a fixed point is restated as the problem of determining a final coalgebra. The paper connects metric coinduction with the classical asynchronous convergence theorem for distributed minimization, so asynchronous parallel RL algorithms are treated as coinductive constructions. The payoff, if the encoding is faithful, is a single coalgebraic scheme covering model classes and solution methods that are usually developed separately.

What carries the argument

The final coalgebra is the load-bearing object: it is the terminal object in the category of coalgebras for a functor, and the paper identifies it with the value-function fixed point that RL algorithms seek. Around it, metric coinduction transfers convergence arguments from the classical asynchronous convergence theorem for distributed minimization, and the topos-valued functor category supplies the limits, colimits, subobject classifier, and exponentials needed to treat algorithms as objects of the same categorical setting.

What would settle it

Take a small finite MDP with a known optimal value function and run the proposed asynchronous coalgebraic iteration with stale component updates under the theorem's asynchrony conditions; if the iteration converges to a value different from the Bellman optimal value, or if a valid instance of the theorem cannot be constructed, the identification fails. Alternatively, exhibit a coalgebra in the proposed category whose final coalgebra exists but does not match the Bellman fixed point of the corresponding MDP.

Watch

Extended reading notes

Core claim

The paper's central claim is that the core problem of RL—computing the fixed point that determines the exact or approximate (action) value function—is a special case of a more general problem: determining the final coalgebra asynchronously, in a parallel distributed manner. It asserts that dynamical models used in RL, including MDPs, POMDPs, PSRs, and linear dynamical systems, are types of coalgebras, and that the space of algorithms for MDPs or PSRs can be modeled as a functor category whose codomain category is a topos. The expected payoff is that coinduction, especially metric coinduction, supplies the convergence mechanism for asynchronous distributed RL, generalizing the classical theor

Load-bearing premise

That MDPs, POMDPs, PSRs, LDSs, and RL algorithms can be faithfully encoded as coalgebras in a topos-valued functor category, so that the final coalgebra is exactly the Bellman value-function fixed point and metric coinduction carries the asynchronous convergence theorem over to this setting.

Editorial extensions

If this is right

  • If the identification holds, every iterative RL algorithm that converges to a value-function fixed point can be viewed as a step in a coinductive construction of a final coalgebra.
  • Asynchronous parallel RL algorithms inherit a uniform convergence guarantee from the classical asynchronous convergence theorem for distributed minimization.
  • MDPs, POMDPs, PSRs, and LDSs—models usually treated with separate solution theories—fall under one coalgebraic framework, so insights transfer between them.
  • Because the algorithm space forms a topos-valued functor category, the design space of RL algorithms has categorical structure that can be used to compose and compare algorithms.

Reading between the lines

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

  • The full text supplied with this submission is a different paper, a review of pandemic models; it contains none of the categorical development promised in the abstract. The central claim therefore rests on the abstract alone in the submitted material.
  • A testable consequence is that a standard asynchronous value-iteration algorithm on a small, explicitly defined MDP should be representable as a metric-coinductive construction of a final coalgebra; exhibiting that representation would make the abstract claim concrete.
  • If the final coalgebra in the relevant topos fails to coincide with the Bellman fixed point—for example, because the functor's coalgebraic behavior encodes transitions but not the contraction structure—then the unification would hold only for a restricted class of RL problems.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 3 minor

Summary. The abstract of arXiv:2508.15128 announces a categorical generalization of reinforcement learning ('universal reinforcement learning'), in which the value-function fixed point is replaced by the asynchronous computation of a final coalgebra in a topos-valued functor category, with convergence inherited from the Bertsekas–Tsitsiklis Asynchronous Convergence Theorem via metric coinduction. The submitted full text, however, is not the paper described in the abstract. It is the manuscript 'Modeling pandemics' by Dawson, Cooper, and Charalampidis, a review of SEIR/SIR epidemic models, data-fitting methods, geographic reaction–diffusion models, and stochastic simulation. The body contains no definitions or theorems about coalgebras, toposes, functor categories, coinduction, reinforcement learning, MDPs, PSRs, or asynchronous convergence. The central claim of the abstract is therefore entirely unsupported by the submitted text.

Significance. If the announced framework were developed and proved, it could represent a substantive unification: classical RL value iteration and Q-learning would become a special case of computing a final coalgebra, and asynchronous distributed RL algorithms would inherit convergence guarantees from a general categorical theorem. Such a result would be of broad interest to the RL and categorical semantics communities. However, none of that development appears in the manuscript. There are no machine-checked proofs, reproducible code, parameter-free derivations, or falsifiable predictions in the supplied text. The only evidence for the announced framework is the abstract itself, which is insufficient for evaluation.

major comments (3)
  1. [Abstract vs. full text] The body of the submission is not the paper described in the abstract. Sections I–IV and Appendices A–D contain an unrelated pandemic-modeling manuscript with SEIR/SIR equations, curve fitting, and Langevin simulation. The terms 'coalgebra', 'topos', 'functor category', 'coinduction', 'conduction', 'MDP', 'PSR', and 'Bertsekas' do not appear in the body. The abstract's central claim—that RL fixed-point problems generalize to asynchronous final-coalgebra computation—therefore has no supporting derivation or proof in the submitted text.
  2. [No definitions or theorem statements] The abstract mentions 'universal coalgebras', 'metric coinduction', and the 'Bertsekas–Tsitsiklis Asynchronous Convergence Theorem', but the manuscript does not define any of these objects or state any related theorem. There is no specification of the functor whose final coalgebra is claimed to represent the value function, no proof that the value function coincides with that final coalgebra, and no statement of the conditions under which metric coinduction transfers the asynchronous convergence theorem. The central mathematical assertion is thus unverifiable from the submitted text.
  3. [Claim that MDPs/POMDPs/PSRs/LDSs are coalgebras] The abstract asserts that MDPs, POMDPs, PSRs, and LDSs are special types of coalgebras. The full text's only dynamical models are compartmental epidemic equations (e.g., Eq. (3.3)) and stochastic reaction-diffusion models (Eq. (3.12)); no coalgebraic encoding is given for any RL model. Consequently, the load-bearing premise that the RL algorithm space admits a faithful functor-category/topos encoding is not established, and the claimed generalization to asynchronous parallel computation cannot be assessed.
minor comments (3)
  1. [Title and framing] The title 'Universal Reinforcement Learning in Coalgebras' does not match the content of the submitted body, which is a review of pandemic models. The abstract promises a two-part paper ('In the first half... In the second half...'), but the body has no such structure.
  2. [Metadata inconsistency] The full text bears the arXiv identifier 2508.15125 and is dated 'September 2, 2025, 7:00pm PST', while the submission is identified as arXiv:2508.15128. This inconsistency should be resolved before any resubmission.
  3. [Appendices and references] The extensive appendices on linearized SEIR solutions, Doi-shifted many-body formalisms, Langevin equations, and the Gillespie algorithm are unrelated to the abstract's RL/coalgebra content. If the intended paper exists, the submitted file appears to be the wrong manuscript.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity can be identified because the claimed URL derivation is not present in the supplied text; the body is an unrelated pandemic-modeling manuscript.

full rationale

The abstract claims that RL value-function fixed points generalize to asynchronous final-coalgebra computation in a topos-valued functor category, with convergence inherited from Bertsekas-Tsitsiklis via metric coinduction. However, the supplied full text is 'Modeling pandemics' by Dawson, Cooper, and Charalampidis, an unrelated review of SEIR-type epidemic models, curve-fitting, and stochastic simulation. It contains no coalgebras, toposes, functor categories, coinduction, or RL equations. There is therefore no derivation chain whose premises and conclusions can be compared, and no equation or fitted parameter that can be shown to reduce to the paper's own inputs. Under the hard rules, circularity requires quoting a specific reduction (e.g., Eq. X = Eq. Y by construction); none exists here. The abstract's central claim is unsupported—the promised 'second half ... universal coalgebras' is absent from the body—but unsupportedness is a completeness/integrity failure, not circularity. Consequently, the circularity score is 0, with the caveat that this says nothing about the correctness or evidentiary value of the URL claim.

Assumptions & free parameters 0 free parameters · 3 assumptions · 2 invented entities

The abstract rests on strong structural assumptions and introduces at least two terms of art (universal coalgebras, conduction) with no definitions in the supplied text. Because the full text is an unrelated pandemic-modeling paper, the ledger reflects only what the abstract asserts; independent evidence for any of the claimed structures is absent.

assumptions (3)
  • domain assumption MDPs, POMDPs, PSRs, and LDSs are faithfully representable as coalgebras in a category supporting a final coalgebra.
    Invoked in the abstract's second half: 'Dynamical system models... are all special types of coalgebras.' This faithfulness is load-bearing for the claim that RL fixed points generalize to final coalgebra computation, and is not proven in any supplied text.
  • ad hoc to paper The space of RL algorithms for MDPs or PSRs forms a functor category whose codomain category is a topos (admitting all (co)limits, subobject classifier, exponentials).
    From the abstract: 'The space of algorithms for MDPs or PSRs can be modeled as a functor category, where the co-domain category forms a topos.' This is a strong structural assumption with no derivation or citation chain provided in the supplied text.
  • domain assumption The Bertsekas-Tsitsiklis asynchronous convergence theory transfers to the coalgebraic setting via metric coinduction.
    Abstract: 'describe the relationship between metric coinduction and their proof of the Asynchronous Convergence Theorem.' The transfer of a classical distributed-computation convergence theorem to a coinductive setting is assumed without proof in the supplied text.
invented entities (2)
  • Universal reinforcement learning (URL) / universal coalgebras
    purpose: The proposed generalization of RL: reframing value-function fixed points as computation of a final coalgebra in a topos-valued functor category.
    Introduced in the abstract with no formal definition, no falsifiable prediction, and no matching content in the supplied full text.
  • Conduction (asynchronous stochastic computation)
    purpose: The mechanism by which the final coalgebra is computed asynchronously in parallel, linked to metric coinduction and the Bertsekas-Tsitsiklis theorem.
    Term appears only in the title and abstract; no mathematical definition is assessable in the supplied text.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Universal Reinforcement Learning in Coalgebras: Asynchronous Stochastic Computation via Conduction." pith.science (2026). https://pith.science/paper/BYWGZYRW

@misc{pith2026250815128,
  author       = {Pith},
  title        = {Pith review of: Universal Reinforcement Learning in Coalgebras: Asynchronous Stochastic Computation via Conduction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BYWGZYRW}},
  note         = {Machine review of arXiv:2508.15128}
}
read the original abstract

In this paper, we introduce a categorial generalization of RL, termed universal reinforcement learning (URL), building on powerful mathematical abstractions from the study of coinduction on non-well-founded sets and universal coalgebras, topos theory, and categorial models of asynchronous parallel distributed computation. In the first half of the paper, we review the basic RL framework, illustrate the use of categories and functors in RL, showing how they lead to interesting insights. In particular, we also introduce a standard model of asynchronous distributed minimization proposed by Bertsekas and Tsitsiklis, and describe the relationship between metric coinduction and their proof of the Asynchronous Convergence Theorem. The space of algorithms for MDPs or PSRs can be modeled as a functor category, where the co-domain category forms a topos, which admits all (co)limits, possesses a subobject classifier, and has exponential objects. In the second half of the paper, we move on to universal coalgebras. Dynamical system models, such as Markov decision processes (MDPs), partially observed MDPs (POMDPs), a predictive state representation (PSRs), and linear dynamical systems (LDSs) are all special types of coalgebras. We describe a broad family of universal coalgebras, extending the dynamic system models studied previously in RL. The core problem in finding fixed points in RL to determine the exact or approximate (action) value function is generalized in URL to determining the final coalgebra asynchronously in a parallel distributed manner.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Consciousness as a Functor

    cs.AI 2025-08 reject novelty 4.0 of 10

    The paper recasts Baars' Global Workspace Theory as a categorical framework in which unconscious processes form a 'topos of coalgebras' and the language of thought is its internal Mitchell-Benabou language.

Reference graph

Works this paper leans on

58 extracted references · 54 canonical work pages · cited by 1 Pith paper

  1. [1]

    Geographical models 9

    Machine learning models 9 C. Geographical models 9

    work page

  2. [2]

    Homogeneous and steady state solutions 12

    work page

  3. [3]

    Stability analysis 13

    work page

  4. [4]

    Temporal stability 14

    work page

  5. [5]

    Stochastic models 16

    Spacial stability 15 D. Stochastic models 16

    work page

  6. [6]

    Applications 16 A

    Langevin equations 16 IV. Applications 16 A. Fitting models to data 16

    work page

  7. [7]

    Application to the 2014 Ebola epidemic in West Africa 19 C

    SIR model and the 2014 Ebola epidemic in West Africa 19 B. Application to the 2014 Ebola epidemic in West Africa 19 C. Application to the US COVID-19 pandemic 21 Acknowledgments 21 A. Solution of the linearized SEIR equations 21 B. Derivation of the Langevin equations 23

    work page 2014

  8. [8]

    The master equation 23

    work page

Show all 58 references
  1. [9]

    Many-body formulation 24

  2. [10]

    The continuum limit 27

  3. [11]

    progressively fall through the same steps

    Path integral 28 C. Gaussian random variables 30 D. Microscopic equations 30 References 31 2 I. INTRODUCTION As a result of the COVID-19 pandemic, numerous mathematical models of the pandemic have recently been proposed and debated in both the press and in scientific publicati...

  4. [12]

    chooses a model curve, inspired by renormalization group techniques, and fits the curve to the observed number of cases available to date. The model from the Institute for Health Metrics and Evaluation (IHME) at UW appears to belong to this first category [2], although apparen...

    2020

  5. [13]

    2-layer densely connected network with 10 units in the hidden layer and the ReLU activation function

    Machine learning models The model of Dandekar and Barbastathis [3] (referred to as the MIT model) uses a neural network to find the best fit for a quarantine function. They start with a simplified SIR model, eliminating the incubation period state E, so as to reduce the number...

    2000

  6. [14]

    red,” and those in (3.15b) “blue

    Homogeneous and steady state solutions Ignoring the recovered population, homogeneous and steady state solutions exist for the rest of the SIR model. Setting ϕα(x, t) = ϕα, they are given by solutions of the equations: ν ϕI − λ ϕS ϕI = 0 . (3.14a) ν ϕS + λ ϕS ϕI = f , (3.14b) ...

  7. [15]

    Then the stability of the solutions of the rate equations can be studied by setting Φ(x, t) = Φ0 + δΦ ei(kx−ωt) , (3.18) and expanding the result to first order in δΦ

    Stability analysis Let us define Φ(x, t) = ϕI(x, t) ϕS(x, t) , Φ0 = ϕI ϕS , δ Φ = δϕI δϕS , (3.17) where Φ 0 are solutions of the homogeneous and steady state solutions (3.14). Then the stability of the solutions of the rate equations can be studied by setting Φ(x, t) = Φ0 + δ...

  8. [16]

    From (3.21), this requires ν + µ + λ ( ϕI − ϕS ) = 0 , (3.23a) µ ν+ λ (µ ϕI − ν ϕS) ≤ 0

    Temporal stability Homogeneous and oscillatory solutions are found for k = 0 when B0 = 0 and C0 < 0. From (3.21), this requires ν + µ + λ ( ϕI − ϕS ) = 0 , (3.23a) µ ν+ λ (µ ϕI − ν ϕS) ≤ 0 . (3.23b) For the red state, ϕS = f /νand ϕI = 0, so that (3.23a) required that λf /ν= ν...

  9. [17]

    These patterns are called Turing patterns — the condition under which the patters emerge are called Turing bifurcations

    Spacial stability Inhomogeneous and stable steady state patterns can be set up if certain conditions are met. These patterns are called Turing patterns — the condition under which the patters emerge are called Turing bifurcations. For patterns to emerge, we must have Ck ≤ 0. T...

  10. [18]

    Solving Langevin- type equations are an alternate way to simulating the master equation directly

    Langevin equations In appendix B, we derive Langevin equations for the SIR model which incorporate noise generated internally by the stochastic nature of the microscopic model. Solving Langevin- type equations are an alternate way to simulating the master equation directly. Fo...

  11. [19]

    Here we consider I0 and N to be fixed

    Example As an example, let us take the SIR model with two model populations y1(t) = I(t) and y2(t) = S(t), and with two model (unknown) parameters, p1 = β and p2 = γ, satisfying the differential equations, ˙y1 = f1(y, p) = p1 y1 y2/N − p2 y1 , (4.7a) ˙y2 = f2(y, p) = −p1 y1 y2...

  12. [20]

    SIR model and the 2014 Ebola epidemic in West Africa We try here to fit the simpler SIR model to the 2014 Ebola epidemic in West Africa. The strict SIR model we use has two parameters β and γ and is given by the equations, dS dt = −β S(t)I(t)/N , (4.13a) dI dt = β S(t)I(t)/N −...

    2014

  13. [21]

    Schr¨ odinger

    Many-body formulation The master equation suggests introducing an occupation number algebra with annihila- tion and creation operators 3 ˆaα(i) and ˆa† α(i) for each site and for each species, where the index i labels the site and the index α the species. The operators obey th...

  14. [22]

    For the reaction S + I k1 − →2I, the stoichiometric vector is S = ( −1, +1) and the reaction rate is a = k1 S I. So this event contributes a factor of k1 (S + 1) (I − 1) P (S + 1, I− 1) − S I P(S, I) , (B7) to the master equation at each site i, and a factor k1 X i [ ˆa† S(i) ...

  15. [23]

    For the reaction I k2 − →0, the stoichiometric vector is S = (0, −1), and the reaction rate is a = k2 I. So this event contributes a factor of k2 (I + 1) P (S, I+ 1) − I P(S, I) , (B9) to the master equation at each site i, and a factor k2 X i [ ˆa† I(i) − 1 ] ˆaI(i) (B10) to ...

  16. [24]

    For the reaction S k3 − →0, the stoichiometric vector is S = ( −1, 0), and the reaction rate is a = k3 S. So this event contributes a factor of k3 (S + 1) P (S + 1, I) − S P(S, I) , (B11) to the master equation at each site i, and a factor k3 X i [ ˆa† S(i) − 1 ] ˆaS(i) (B12) ...

  17. [25]

    So this event contributes a factor of k4 P (S − 1, I) − P (S, I) , (B13) to the master equation at each site i, and a factor k4 X i [ 1 − ˆa† S(i) ] (B14) to the Hamiltonian

    For the reaction 0 k4 − →S, the stoichiometric vector is S = (1, 0), and the reaction rate is a = k4. So this event contributes a factor of k4 P (S − 1, I) − P (S, I) , (B13) to the master equation at each site i, and a factor k4 X i [ 1 − ˆa† S(i) ] (B14) to the Hamiltonian

  18. [26]

    Finally, the diffusion process is generated by hopping of like species between nearest neighbor sites 1 and 2 at a rate dα. If n1 is the number of particles of species α at site 1 and n2 the number of particles of species α at site 2, this event contributes a factor dα (n1 + 1...

  19. [27]

    The Doi shift A simple trick, discovered by Doi [21], provides a probabilistic interpretation of the state vector. Doi found that if all creation operators are translated by one unit leaving the annihilation operators unchanged, the overlap of the transformed state vector with...

  20. [28]

    Set the time t = 0 and the initial occupation values for the reactants n1(0)

  21. [29]

    Generate two random numbers r1 and r2 uniformly distributed in (0 , 1)

  22. [30]

    Compute the propensity function ai(n) for each reaction and the sum over all reactions: a(n) = qX i=1 ai(n) . (D3)

  23. [31]

    Using (D2), compute the time ∆ t when the next reaction takes place and update the time, t = t + ∆t

  24. [32]

    (D4) Then the jth reaction takes place

    Compute which reaction takes place by finding the value of j where 1 a(n) j−1X i=1 ai(n) ≤ r2 < 1 a(n) jX i=1 ai(n) . (D4) Then the jth reaction takes place. Since r2 is a random number it doesn’t matter how the reactions are ordered. Then update the occupation numbers of reac...

  25. [33]

    Go back to step 2 and continue to the end time. There is only one reaction per time step so one does not have to recompute all the propensity functions at each time step, but can just update only those that are changed by the reaction selected. Reaction parameters for the SEIR...

  26. [34]

    Della Morte, D

    M. Della Morte, D. Orlando, and F. Sannino, Renormalization group approach to pandemics: The covid-19 case, Frontiers in Physics 8, 144 (2020). 31

    2020

  27. [35]

    N. P. Jewell, J. A. Lewnard, and B. L. Jewell, Caution Warranted: Using the Institute for Health Metrics and Evaluation Model for Predicting the Course of the COVID-19 Pandemic, Annals of Internal Medicine 10.7326/M20-1565 (2020), https://annals.org/acp/content public/journal/...

    2020 doi

  28. [36]

    Dandekar and G

    R. Dandekar and G. Barbastathis, Quantifying the effect of quarantine control in covid- 19 infectious spread using machine learning, medRxiv 10.1101/2020.04.03.20052084 (2020), https://www.medrxiv.org/content/early/2020/04/06/2020.04.03.20052084.full.pdf

    2020 doi

  29. [37]

    Girona, Confinement time required to avoid a quick rebound of covid-19: Predictions from a monte carlo stochastic model, Frontiers in Physics 8, 186 (2020)

    T. Girona, Confinement time required to avoid a quick rebound of covid-19: Predictions from a monte carlo stochastic model, Frontiers in Physics 8, 186 (2020)

    2020

  30. [38]

    Berezovskaya, G

    F. Berezovskaya, G. Karev, B. Song, and C. Castillo-Chavez, A Simple Epidemic Model with Surprising Dynamics, Mathematical Biosciences and Engineering 2, 133 (2005)

    2005

  31. [39]

    W. Wang, Y. Cai, M. Wu, K. Wang, and Z. Li, Complex dynamics of a reaction–diffusion epidemic model, Nonlinear Analysis: Real World Applications 13, 2240 (2012)

    2012

  32. [40]

    Brauer, Early estimates of epidemic final sizes, Journal of Biological Dynamics 13, 23 (2019), pMID: 29742981, https://doi.org/10.1080/17513758.2018.1469792

    F. Brauer, Early estimates of epidemic final sizes, Journal of Biological Dynamics 13, 23 (2019), pMID: 29742981, https://doi.org/10.1080/17513758.2018.1469792

    2019

  33. [41]

    Diekmann, J

    O. Diekmann, J. A. P. Heesterbeek, and M. G. Roberts, The construction of next-generation matrices for compartmental epidemic models, Journal of The Royal Society Interface 7, 873 (2010), https://royalsocietypublishing.org/doi/pdf/10.1098/rsif.2009.0386

    2010

  34. [42]

    J. C. Miller, A note on the derivation of epidemic final sizes, Bulletin of Mathematical Biology 74, 2125 (2012)

    2012

  35. [43]

    Y. Cao, S. Li, L. Petzold, and R. Serban, Adjoint sensitivity analysis for differential-algebraic equations: The adjoint dae system and its numerical solution, SIAM Journal on Scientific Computing 24, 1076 (2003), https://doi.org/10.1137/S1064827501380630

    2003 doi

  36. [44]

    Rackauckas, Y

    C. Rackauckas, Y. Ma, J. Martensen, C. Warner, K. Zubov, R. Supekar, D. Skinner, and A. Ramadhan, Universal differential equations for scientific machine learning (2020), arXiv:2001.04385 [cs.LG]

    2020 arXiv

  37. [45]

    Rackauckas, M

    C. Rackauckas, M. Innes, Y. Ma, J. Bettencourt, L. White, and V. Dixit, DiffEqFlux.jl — A Julia Library for Neural Differential Equations, arXiv: 1902.02376v1 (2019)

    1902 arXiv

  38. [46]

    Cox and P

    S. Cox and P. Matthews, Exponential time differencing for stiff systems, Journal of Compu- tational Physics 176, 430 (2002)

    2002

  39. [47]

    A. K. Kassam, Solving reaction-diffusion equations 10 times faster, Elsevier Science (2003)

    2003

  40. [48]

    A. K. Kassam and L. N. Trefethen, Fourth-order time stepping for stiff PDEs, SIAM J. Sci. Comput. 26, 1214 (2005)

    2005

  41. [49]

    Li and L

    S. Li and L. Petzold, Software and algorithms for sensitivity analysis of large-scale differential algebraic systems, Journal of Computational and Applied Mathematics 125, 131 (2000), numerical Analysis 2000. Vol. VI: Ordinary Differential Equations and Integral Equations

    2000

  42. [50]

    C. L. Althaus, Estimating the Reproduction Number of Ebola Virus (EBOV) Dur- ing the 2014 Outbreak in West Africa, PLOS Currents Outbreaks 1, 10.1371/cur- rents.outbreaks.91afb5e0f279e7f29e7056095255b288 (2014)

    2014 doi

  43. [51]

    B. Tang, X. Wang, Q. Li, N. Bragazzi, S. Tang, Y. Xiao, and J. Wu, Estimation of the Transmission Risk of the 2019-nCoV and Its Implication for Public Health Interventions, J. Clin. Med. 9, 462 (2020)

    2019

  44. [52]

    Smirnova, L

    A. Smirnova, L. deCamp, and G. Chowell, Forecasting Epidemics Through Nonparametric Estimation of Time-Dependent Transmission Rates Using the SEIR Model., Bull Math Biol 81, 4343 (2019)

    2019

  45. [53]

    M. R. Roussel, The chemical master equation (2007), lecture notes, University of Lethbridge. 32

    2007

  46. [54]

    Doi, Second quantization representation for classical many-particle system, Journal of Physics A: Mathematical and General 9, 1465 (1976)

    M. Doi, Second quantization representation for classical many-particle system, Journal of Physics A: Mathematical and General 9, 1465 (1976)

    1976

  47. [55]

    D. T. Gillespie, A general method for numerically simulating the stochastic time evolution of coupled chemical reactions, Journal of Computational Physics 22, 403 (1976)

    1976

  48. [56]

    D. T. Gillespie, Exact stochastic simulation of coupled chemical reactions, The Journal of Physical Chemistry, The Journal of Physical Chemistry 81, 2340 (1977)

    1977

  49. [57]

    Erban, J

    R. Erban, J. Chapman, and P. Maini, A practical guide to stochastic simulations of reaction- diffusion processes (2007), arXiv:0704.1908 [q-bio.SC]

    2007 arXiv

  50. [58]

    Erban and S

    R. Erban and S. J. Chapman, Stochastic Modelling of Reaction–Diffusion Processes , Cam- bridge Texts in Applied Mathematics (Cambridge University Press, 2020). 33

    2020

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.