Pith. sign in

REVIEW 3 major objections 5 minor 77 references

Large deviations for light-tailed L\'evy bridges on short time scales

T0 review · 3 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 This paper proves a sample-path large deviations principle for short-time light-tailed Lévy bridges, with an entropy-type rate function whose unique minimizer is the straight-line chord from 0 to x.

desk verdict The main sample-path LDP as stated is false: the rate function assigns zero cost to singular monotone paths, while the proof's finite-dimensional rate assigns infinite cost. read the letter →

arxiv 2505.23972 v1 pith:BPKXXXNU submitted 2025-05-29 math.PR

Michael A. Högele , Torsten Wetzel This is my paper
classification math.PR MSC 60F1060G5160J75
keywords largedeviationsprincipleLévybridgeshorttimescalinglight-tailedjumpssmoothregularvariationentropyratefunctionsamplepathLDPcompoundPoissonprocess
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 proves a sample-path large deviations principle for bridges of a Lévy process whose jumps have light tails $e^{-f(|y|)}$ with $f$ smoothly regularly varying of index $\alpha>1$, after spatial rescaling by $\varepsilon$ and temporal rescaling by $r_\varepsilon=o(\varepsilon^{-1})$. In this short-time regime the unconditioned rescaled process collapses to zero, so standard large-deviations machinery yields nothing; the paper instead conditions the process to hit $x\neq 0$ at time $T$ and proves that the conditioned bridge obeys a genuine LDP with speed $S(\varepsilon)=\varepsilon\,g(\varepsilon^{-1}r_\varepsilon^{-1})$ and an entropy-type rate function. The rate is finite only for continuous, nondecreasing radial parametrizations of the straight segment from $0$ to $x$, and its unique minimizer is the linear path, so all paths that wander off the chord are exponentially negligible. The same method yields an LDP for the number of jumps and asymptotic normality of jump increments, and is stable under addition of Brownian motion, drift, or lighter-tailed independent Lévy perturbations.

What carries the argument

The load-bearing object is the auxiliary function $g$, defined as the unique pointwise solution of the nonlinear functional equation (2.12). It is slowly varying, $g(e^\Lambda)\in SR_{1/\alpha}$, and its decisive property is the asymptotic cancellation relation (2.15): for $y$ in a logarithmic window around $1$, $g(\Lambda)[f'(g(y\Lambda))-f'(g(\Lambda))]=\ln y+o(|\ln y|)$. This relation is what converts the sharp convolution-density estimates of Theorem 2 into the entropy integral: after conditioning on the bridge endpoint, the log-density at an interior point $y$ telescopes into $|y|\ln(|y|/t)+|x-y|\ln(|x-y|/(T-t))-|x|\ln(|x|/T)$ divided by $S(\varepsilon)$. The full path LDP is then assembled from finite-dimensional LDPs for these telescoped densities, exponential tightness on Skorokhod space, and a transfer from the $J_1$ topology to the supremum norm.

What would settle it

Take $n=1$, $f(r)=r^2$, $T=1$, $x=1$, $r_\varepsilon\equiv 1$, and simulate the bridge $Y^{\varepsilon}$ for very small $\varepsilon$. Compute $S(\varepsilon)\ln\mathbb{P}(|Y_{1/2}^{\varepsilon}-1/4|<\delta)$ and $S(\varepsilon)\ln\mathbb{P}(\inf_t |Y_t^\varepsilon-[[0,1]]|>\kappa)$. The paper predicts the first converges to $\frac14\ln(1/2)+\frac34\ln(3/2)\approx 0.131$ and the second to $-\infty$; a finite limit for the second, or a different constant for the first, would falsify the LDP.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that conditioning a rescaled Lévy process with light-tailed jumps on a fixed endpoint changes the large-deviation regime completely: whereas the unconditioned process has only the trivial rate function, the bridge $(Y^{x,\varepsilon})_{t\in[0,T]}$ satisfies a genuine sample-path LDP on $D([0,T],\mathbb{R}^n)$ with the uniform topology. The speed is $S(\varepsilon)=\varepsilon\,g(\varepsilon^{-1}r_\varepsilon^{-1})$, where $g$ solves the nonlinear functional equation (2.12), and the rate is $I_x(\phi)=\int_0^T |\phi|'\ln|\phi|'\,dt-|x|\ln(|x|/T)$ for monotone continuous radial parametrizations of the segment $[[0,x]]$, and $I_x(\phi)=\infty$ otherwise. Hence the asymptotic energy-minimizing path is the linear path $\phi(t)=tx/T$, and paths leaving the segment are exponentially negligible at speed $S(\varepsilon)$.

Load-bearing premise

The whole construction rests on one sharp cancellation: for large scales $\Lambda$ and ratios $y$ close to $1$, the difference $f'(g(\Lambda y))-f'(g(\Lambda))$ must equal $\ln y / g(\Lambda)$ up to small errors; this is the step that turns density estimates into the entropy rate, and if it failed the stated form of the rate function would break.

Editorial extensions

If this is right

  • Paths that leave the segment $[[0,x]]$ or whose distance from $0$ decreases are exponentially negligible at speed $S(\varepsilon)$; in the limit, the bridge lives on continuous monotone parametrizations of the chord.
  • The linear path has rate zero, so the bridge converges in the large-deviation sense to uniform motion along the straight line from $0$ to $x$.
  • The number of jumps $N^{x,\varepsilon}$, centered at $m_{x,\varepsilon}=|x|\varepsilon^{-1}g(|x|\varepsilon^{-1}/r_\varepsilon T)^{-1}$ and scaled by $\sqrt{S(\varepsilon)k_\varepsilon}$, obeys an LDP with quadratic rate $y^2/2$, giving a law of large numbers with Gaussian-type fluctuations.
  • After rescaling by $\sqrt{f''(g(|x|\varepsilon^{-1}/r_\varepsilon T))}$, a single jump increment of $Y^{x,\varepsilon}/\varepsilon$ converges in distribution to a standard normal, with different variances along the target direction and its orthogonal complement.
  • The density estimates and the LDP survive when the compound Poisson component is supplemented by Brownian motion, drift, or an independent lighter-tailed Lévy perturbation satisfying Hypothesis II.

Reading between the lines

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

  • An implicit consequence is that the bridge's large-deviation behavior is governed by the radial entropy of the path rather than by the detailed shape of the jump measure; only the index $\alpha$ enters through the speed $S$. This suggests the same rate function may describe other conditioned light-tailed processes once a matching function $g$ is available.
  • Because the unconditioned process has a trivial LDP, the nontrivial rate emerges only through conditioning; this points to a general principle for short-time light-tailed bridges: conditioning acts as a filter that removes the trivial concentration at zero, exposing the entropy cost of the path's radial speed.
  • The speed $S(\varepsilon)=\varepsilon g(\varepsilon^{-1}r_\varepsilon^{-1})$ behaves like $\varepsilon(\ln(1/\varepsilon))^{1/\alpha}$, which is far slower than the Brownian $\varepsilon^2$; one could test whether transition probabilities and first-passage asymptotics between distant points inherit this logarithmic speed in related models.
  • The finite-dimensional rate function matches the infimum of the path rate over interpolations, so the LDP is consistent with a variational (Schrödinger-bridge-like) principle for the bridge; this may allow numerical computation of rare bridge probabilities by solving a one-dimensional entropy minimization on the chord.
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

3 major / 5 minor

Summary. The paper studies a rescaled multivariate Lévy process X^ε_t = ε L_{r_ε t} with r_ε = o(ε^{-1}) and light-tailed, rotationally invariant Lévy measure exp(-f(|y|)) dy for f smoothly regularly varying of index α>1. It conditions the process on the event X^ε_T = x and derives: (i) exponential density estimates for the marginals of L via convolution estimates and a nonlinear functional equation defining an auxiliary function g; (ii) a sample-path LDP for the bridge Y^{x,ε} on D[0,T] with the uniform norm, speed S(ε)=ε g(ε^{-1} r_ε^{-1}), and an entropy-type rate function I_x; (iii) an LDP for the number of jumps; and (iv) asymptotic normality of jump increments. The proofs proceed through finite-dimensional LDPs, a Skorokhod-J1 exponential tightness argument, and a transfer to the uniform topology via Feng–Kurtz theorems. The paper also proves sufficient conditions for perturbations by additional Lévy components and includes an appendix showing that no classical LDP holds for the un-conditioned rescaled process on short time scales.

Significance. If the main results were correct as stated, the paper would fill a genuine gap in the large-deviations literature: the short-time regime r_ε ≪ ε^{-1} for light-tailed Lévy bridges, where classical sample-path LDP methods fail. The approach via sharp convolution density estimates is original and technically substantial: the auxiliary function g is derived from a functional equation rather than fitted, the density estimates in Theorem 2 are explicit and parameter-free, and the paper identifies the optimal path as the linear parametrization of the segment [[0,x]] with entropy cost. The paper also contains useful robustness results for Brownian and Lévy perturbations. However, the main theorem as stated is not internally consistent: the rate function assigns value zero to singular continuous paths such as the Cantor function, while the finite-dimensional rate used in the proof assigns infinite cost to such paths. This is a load-bearing flaw, although it is localizable and appears fixable by restricting the admissible paths to absolutely continuous ones.

major comments (3)
  1. [§2.2.3, Theorem 3 and Eq. (2.21)–(2.22)] The rate function I_x is incorrect on singular continuous paths, and this invalidates the statement of Theorem 3. The set D_{x,T} defined in (2.21) contains the path φ(t) = x C(t)/|x|, where C is the Cantor–Lebesgue function on [0,T] with T=|x|=1. This φ is continuous, nondecreasing in |φ|, and satisfies φ([0,T])=[[0,x]], so φ∈D_{x,T}; since C'=0 a.e., Eq. (2.22) gives I_x(φ)=0. But for the level-k ternary partition, the finite-dimensional rate I_{x,τ} of Proposition 3.1, Eq. (3.52), equals k ln(3/2)→∞. Hence the central identification (3.56), on which the proof of Theorem 3 relies, fails for a path explicitly admitted by the definitions. Consequently, for a uniform open neighborhood A of φ, the lower bound required by the LDP would demand liminf S(ε) ln P(Y^{x,ε}∈A) ≥ 0, whereas the finite-dimensional upper bound forces limsup S(ε) ln P(Y^{x,ε}∈A) = -∞. The definition of D_{x,T} must be restricted to paths for which |φ| is absolutely continuous (with I_x=∞ otherwise), or the rate must be defined through the Lebesgue decomposition of the monotone function |φ|, assigning infinite cost to any singular component.
  2. [Theorem 1, item 3, versus Theorem 4(i) and §3.4.1] The simplified theorem statement in the introduction is inconsistent with the main theorem and with the proof. In Theorem 1, item 3, the normalizing constant is defined as k_ε := α ε |x|^{-1} g(|x| ε^{-(1-ρ)}), whereas Theorem 4(i) defines k_ε := α |x|^{-1} ε g(|x| ε^{-1}/(r_ε T)) |ln ε|, and the proof in §3.4.1 requires the factor |ln ε|: for example, the estimate (3.106) identifies θ''_ε(m) with (1+4δ) m^{-1}_{x,ε} α |ln ε|, which equals (1+4δ) k_ε only if k_ε contains |ln ε|, and (3.111) uses the same factor. As printed, Theorem 1, item 3, would assert an LDP with the wrong scaling and is false.
  3. [§3.3, proof of Theorem 3, after Eq. (3.56)] The transfer from the Skorokhod J1 topology to the uniform norm is not justified as written. The proof states 'By Theorem 4.13 and Theorem 4.14 in Feng and Kurtz [32]' without verifying the hypotheses of those theorems in the present setting. In particular, it is not checked that the rate function satisfies the continuity or exponential-tightness conditions needed to pass from the J1-LDP to the sup-norm LDP. Since the rate function is infinite on discontinuous paths, one expects such a transfer to be true, but the argument needs to be supplied explicitly; this is a genuine gap in the proof of the main theorem's topology.
minor comments (5)
  1. [Eq. (2.22)] The phrase '|φ|′(t) = d/dt |φ(t)| the total derivative, whenever it exists and set it equal to 0 otherwise' is ambiguous for monotone functions with a singular continuous part; after the correction in the major comment, the derivative should be understood as the density of the absolutely continuous part of the Lebesgue decomposition.
  2. [Theorem 1, item 2] The simplified sample-path LDP stated in Theorem 1, item 2, inherits the same singular-path defect as Theorem 3 and should be corrected together with the main theorem.
  3. [Remark 2.5, item 1] The inequality 'For each δ < 0' appears to be a typo and should read 'δ > 0'.
  4. [Throughout] There are numerous typographical and rendering issues, e.g. 'c` adl` ag' and 'L´ evy'; these should be cleaned up in a revision.
  5. [Lemma A.2] The proof of Lemma A.2 is referred to an unpublished thesis; given that the result is used in the proof of Lemma 2.4, it would be preferable to include the short proof or state explicitly where it appears.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the rate function is derived from density estimates, and citations to earlier author work are not load-bearing.

full rationale

The claimed derivation chain is self-contained at the level relevant to circularity. The speed function S(ε) = ε g(ε^{-1} r_ε^{-1}) and the rate function I_x are not assumed or fitted: g is uniquely defined by the functional equation (2.12), Lemma 2.2 proves existence, uniqueness, regular variation, and the asymptotic cancellation relation (2.15) from the smooth regular variation of f, and Theorem 2 derives the marginal density estimates from the convolution estimates of Proposition 2.1 together with Hypotheses I and II. The bridge density identity (1.21) then yields the finite-dimensional LDP in Proposition 3.1, and the pathwise LDP in Theorem 3 is obtained via the S-exponential tightness of Lemma 3.1 combined with Feng-Kurtz Theorems 4.13/4.14 and the contraction principle; the rate function is identified through the variational identity rather than by fiat. Citations to earlier work of the same authors ([42], [68], [69]) occur for motivation, consistency checks, and one auxiliary sufficient-condition lemma in Appendix A; even where a proof is deferred to [69], the cited result is an external tail estimate for Lévy processes and not an assumption of Theorem 3. No prediction in the paper is statistically forced by a fitted parameter, and no uniqueness claim or ansatz is imported from a same-author citation to close the argument. The skeptical objection concerning singular continuous paths is a question about whether the equality (3.56) is mathematically valid on the full class D_{x,T}; that is a correctness issue, not circularity, because the disputed equality is proved from the finite-dimensional rates rather than assumed.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The central theorem is a new derivation. No constants are fitted to data and no new physical or mathematical entities are postulated. The auxiliary function g is defined by a functional equation and its properties are proved, not assumed. The main restrictions are the light-tailed form of nu_eta, Hypothesis II on nu_xi, and the extra convexity assumption used in Theorem 4.

assumptions (5)
  • standard math Standard Lévy process background: Lévy-Khintchine formula, Lévy-Itô decomposition, Skorokhod topology, and the Feng-Kurtz and Dembo-Zeitouni LDP frameworks.
    Invoked throughout Section 2 and the proofs; these are established external tools, not new content.
  • domain assumption The jump measure splits as nu = nu_eta + nu_xi, where nu_eta has density exp(-f(|y|)) with f in SR_alpha, alpha > 1, and nu_xi satisfies Hypothesis II, including the super-exponential domination condition (2.8).
    This defines the process class for which the theorems are stated. If the perturbation xi dominates the density tails, the convolution estimates and the bridge LDP could fail.
  • domain assumption The time scale r_epsilon is regularly varying with index rho > -1 and satisfies r_epsilon = o(epsilon^{-1}).
    This selects the short-time regime. The classical sample-mean regime r_epsilon ~ epsilon^{-1} is excluded and requires different methods.
  • ad hoc to paper For Theorem 4, f is additionally assumed convex and monotonically non-decreasing.
    This extra assumption is introduced immediately before Definition 2.3 and is used in the proof of Lemma 3.3 through the subadditivity estimate (3.10). It is not part of Hypothesis I and is not mentioned in the abstract.
  • standard math External results in Feng and Kurtz [32], in particular Theorems 4.13, 4.14 and 4.28, are applicable to the bridge family with the non-standard speed S.
    The proof of Theorem 3 uses these results to pass from finite-dimensional LDPs and exponential tightness to the path LDP, and to transfer from the Skorokhod J1 topology to the uniform norm.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Large deviations for light-tailed L\'evy bridges on short time scales." pith.science (2026). https://pith.science/paper/BPKXXXNU

@misc{pith2026250523972,
  author       = {Pith},
  title        = {Pith review of: Large deviations for light-tailed L\'evy bridges on short time scales},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BPKXXXNU}},
  note         = {Machine review of arXiv:2505.23972}
}
abstract

Let $L = (L(t))_{t\geq 0}$ be a multivariate L\'evy process with L\'evy measure $\nu(dy) = \exp(-f(|y|)) dy$ for a smoothly regularly varying function $f$ of index $\alpha>1$. The process $L$ is renormalized as $X^\varepsilon(t) = \varepsilon L(r_\varepsilon t)$, $t\in [0, T]$, for a scaling parameter $r_\varepsilon= o(\varepsilon^{-1})$, as $\varepsilon \to 0$. We study the behavior of the bridge $Y^{\varepsilon, x}$ of the renormalized process $X^\varepsilon$ conditioned on the event $X^\varepsilon(T) = x$ for a given end point $x\neq 0$ and end time $T>0$ in the regime of small $\varepsilon$. Our main result is a sample path large deviations principle (LDP) for $Y^{\varepsilon, x}$ with a specific speed function $S(\varepsilon)$ and an entropy-type rate function $I_{x}$ on the Skorokhod space in the limit $\varepsilon \rightarrow 0+$. We show that the asymptotic energy minimizing path of $Y^{\varepsilon, x}$ is the linear parametrization of the straight line between $0$ and $x$, while all paths leaving this set are exponentially negligible. We also infer a LDP for the asymptotic number of jumps and establish asymptotic normality of the jump increments of $Y^{\varepsilon, x}$. Since on these short time scales $r_\varepsilon = o(\varepsilon^{-1})$) direct LDP methods cannot be adapted we use an alternative direct approach based on convolution density estimates of the marginals $X^{\varepsilon}(t)$, $t\in [0, T]$,for which we solve a specific nonlinear functional equation.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

77 extracted references · 77 canonical work pages

  1. [75]

    F.: Large deviations for Markov bridges with jumps

    Yang, X. F.: Large deviations for Markov bridges with jumps. J. Math. Anal. Appl. 416 (1), (2014), 1-12

    work page 2014

  2. [4]

    Bazhba, M., Blanchet, J., Rhee, C.-H., and Zwart, B.: Sample path large deviations for L´ evy processes and random walks with Weibull incrementes Ann. Appl. Probab. 30(6): 2695-2739

    work page

  3. [32]

    G.: Large Deviations for Stochastic Processes

    Feng, J., and Kurtz, T. G.: Large Deviations for Stochastic Processes. Volume 131 of Mathematical Surveys and Monographs. American Mathematical Society, 2006

    work page 2006

  4. [1]

    Cambridge University Press, Cambridge, 2004

    Applebaum, D.: L´ evy processes and stochastic calculus, Volume 93 of Cambridge studies in advanced mathematics. Cambridge University Press, Cambridge, 2004

    work page 2004

  5. [2]

    De Acosta, A.: Large deviations for vector-valued L´ evy processes.Stochastic Processes and their Applications 51 (1994) 75-115

    work page 1994

  6. [3]

    Advances in Probability, Vol

    De Acosta, A., Araujo, A., and Gin´ e, E.: On Poisson measures, Gaussian measures and the central limit theorem in Banach spaces. Advances in Probability, Vol. 4 (Marcel Dekker, New York, 1978) pp. 1-68

    work page 1978

  7. [5]

    Random Oper

    Barczy, M., and Kern, P.: Representations of multidimensional linear process bridges. Random Oper. Stoch. Equ. 21 (2), (2013), 159–189. 67

    work page 2013

  8. [6]

    H., Goldie, C

    Bingham, N. H., Goldie, C. M., and Teugels, J. L.: Regular variation, Volume 27 of Encyclopedia of Mathematic and its Applications, Cambridge University Press, Cambridge etc., 1987

    work page 1987

Show all 77 references
  1. [7]

    A., and Mogulskii, A

    Borovkov, A. A., and Mogulskii, A. A.: On large deviation principles in metric spaces. Siberian Mathematical Journal 51, (2010), 989–1003

    2010

  2. [8]

    Journal of the European Mathematics Society, 6(4), (2004), 399–424

    Bovier, A., Eckhoff, M., Gayrard, V., and Klein, M.: Metastability in reversible diffu- sion processes: Sharp asymptotics for capacities and exit times. Journal of the European Mathematics Society, 6(4), (2004), 399–424

    2004

  3. [9]

    H., and Zhai, J

    Brze´ zniak, Z., Peng, X. H., and Zhai, J. L.: Well-posedness and large deviations for 2D stochastic Navier-Stokes equations with jumps. J. Eur. Math. Soc. (JEMS) 25 (8), (2023), 3093–3176

    2023

  4. [10]

    The Annals of Probability, 44(3), (2016), 1723–1775

    Budhiraja, A., Dupuis, P., Ganguly, A.: Moderate deviation principles for stochastic dif- ferential equations with jumps. The Annals of Probability, 44(3), (2016), 1723–1775

    2016

  5. [11]

    Stochastic Process

    Budhiraja, A., Chen, J., and Dupuis, P.: Large deviations for stochastic partial differential equations driven by a Poisson random measure. Stochastic Process. Appl. 123, (2013), 523–60

    2013

  6. [12]

    Probability Theory and Stochastic Modelling, Vol

    Budhiraja, A., and Dupuis, P.: Analysis and Approximation of Rare Events: Representa- tions and Weak Convergence Methods. Probability Theory and Stochastic Modelling, Vol. 94, Springer Verlag, 2019

    2019

  7. [13]

    Budhiraja, A., Dupuis, P., and Maroulas, V.: Variational representations for continuous time processes. Ann. Inst. H. Poincar´ eProbab. Statist. 47, (2011), 725–747

    2011

  8. [14]

    Budhiraja, A., and Nyquist, P.: Large deviations for multidimensional state-dependent shot-noise processes. J. Appl. Probab. 52, (2015), 1097–1114

    2015

  9. [15]

    Transactions of the American Mathematical Society 372(12), (2019), 8363–8421

    Budhiraja, A., Dupuis, P., and Salins, M.: Uniform large deviation principles for Banach space valued stochastic evolution equations. Transactions of the American Mathematical Society 372(12), (2019), 8363–8421

    2019

  10. [16]

    Cerrai, S., and Salins, M.: Smoluchowski-Kramers approximation and large deviations for infinite-dimensional nongradient systems with applications to the exit problem. Ann. Probab. 44(4) (2016) 2591-2642

    2016

  11. [17]

    Stochastics and Dynamics, 23 (07), 2023

    De Oliveira Gomes, A., and Catuogno, P.: Large deviations for L´ evy diffusions in the small noise regime. Stochastics and Dynamics, 23 (07), 2023

    2023

  12. [18]

    Discrete and Continuous Dynamical Systems 37(1), (2017), 33–76

    Cerrai, S., Freidlin, M., and Salins, M.: On the Smoluchowski-Kramers approximation for SPDEs and its interplay with large deviations and long time behavior. Discrete and Continuous Dynamical Systems 37(1), (2017), 33–76

    2017

  13. [19]

    Chaumont, L., and Uribe-Bravo, G.: Markovian bridges: Weak continuity and pathwise constructions. Ann. Probab. 39, (2011), 609–647. MR2789508

    2011

  14. [20]

    Reciprocal classes and duality formulas Electron

    Conforti, G., L´ eonard, Ch., R¨ udiger, M., and Roelly, S.: Bridges of Markov counting processes. Reciprocal classes and duality formulas Electron. Commun. Probab. 20 (18), (2015), 1–12. 68

    2015

  15. [21]

    Cram´ er, H.: On a new limit theorem of the theory of probability.Uspekhi Matematicheskikh Nauk (10), (1944), 166–178

    1944

  16. [22]

    Springer Jones and Bartlett, Boston, 1993

    Dembo, A., and Zeitouni, O.: Large Deviations Techniques and Applications. Springer Jones and Bartlett, Boston, 1993

    1993

  17. [23]

    S.: A Weak Convergence Approach to the Theory of Large Deviations

    Dupuis, P., and Ellis, R. S.: A Weak Convergence Approach to the Theory of Large Deviations. Wiley Series in Probability and Statistics, 1997

    1997

  18. [24]

    S.: Large Deviations for a General Class of Random Vectors

    Ellis, R. S.: Large Deviations for a General Class of Random Vectors. The Annals of Probability, 12(1), (1984), 1-12

    1984

  19. [25]

    N., and Kurtz, T

    Ethier, S. N., and Kurtz, T. G.: Markov Processes Characterization and Convergence. Wiley Series in Probability and Statistics, 1986

    1986

  20. [26]

    Debussche, A., H¨ ogele, M., and Imkeller, P.: The dynamics of nonlinear reaction-diffusion equations with small L´ evy noise.Lecture Notes in Mathematics. 2085. Springer, 2013

    2013

  21. [27]

    Deuschel, J.-D., and Stroock, D.: Large deviations, Pure and Applied Mathematics, Academic Press, 1989

    1989

  22. [28]

    S.: A Weak Convergence Approach to the Theory of Large Devi- ations

    Dupuis, P., and Ellis, R. S.: A Weak Convergence Approach to the Theory of Large Devi- ations. Wiley, New York, 1997

    1997

  23. [29]

    D., and Varadhan, S

    Donsker, M. D., and Varadhan, S. R. S.: Large deviations for Markov processes and the asymptotic evaluation of certain Markov process expectations for large times. Probabilistic methods in differential equations (Proc. Conf., Univ. Victoria, Victoria, B. C., 1974), pp. 82–88. ...

    1974

  24. [30]

    random variables

    Eichelsbacher, P., and L¨ owe, M.: Moderate deviations for i.i.d. random variables. ESAIM: Probability and Statistics 7, (2003), 209–218

    2003

  25. [31]

    Springer, Berlin, 2006

    Ellis, R.S.: Entropy, Large Deviations, and Statistical Mechanics. Springer, Berlin, 2006

    2006

  26. [33]

    Fitzsimmons, P., Pitman, J., and Yor, M.: Markovian bridges: Construction, Palm inter- pretation, and splicing. Progr. Probab. 33, (1992), 101–134

    1992

  27. [34]

    I., and Wentzell, A

    Freidlin, M. I., and Wentzell, A. D.: Random Perturbations of Dynamical Systems. Volume 260 of Grundlehren der mathematischen Wissenschaften, Springer, 1984

    1984

  28. [35]

    Theory Probab

    G¨ artner, J.: On large deviations from the invariant measure. Theory Probab. Appl. 22, (1977), 24439

    1977

  29. [36]

    in: Stochastic Analysis and Applications, in: Abel Symp., Springer, 2007

    Gasbarra, D., Sottinen, T., and Valkeila, E.: Gaussian bridges. in: Stochastic Analysis and Applications, in: Abel Symp., Springer, 2007

    2007

  30. [37]

    Gasteratos, I., Salins, M., and Spiliopoulos, K.: Moderate deviations for systems of slow- fast reaction-diffusion equations, Stochastics and Partial Differential Equations: Analysis and Compuations 11(2) (2023), pp. 503-598. arXiv:2101.00085. 69

    2023 arXiv

  31. [38]

    Stochastic Analysis and Applications, 32 (1), (2013), 163–190

    H¨ ogele, M., and Pavlyukevich, I.: The exit problem from the neighborhood of a global attractor for heavy-tailed L´ evy diffusions. Stochastic Analysis and Applications, 32 (1), (2013), 163–190

    2013

  32. [39]

    Fields Institute Monographs 14, American Mathematical Society, Providence RI, 2000

    Den Hollander, F.: Large Deviations. Fields Institute Monographs 14, American Mathematical Society, Providence RI, 2000

    2000

  33. [40]

    Bernoulli, 13 (1), (2007), 252–278

    Houdr´ e, C., Kawai, R.: On layered stable processes. Bernoulli, 13 (1), (2007), 252–278

    2007

  34. [41]

    P.: Brownian bridges on Riemannian manifolds, Probab

    Hsu, E. P.: Brownian bridges on Riemannian manifolds, Probab. Theory Related Fields 84, (1990), 103–118

    1990

  35. [42]

    The Annals of Probability, 37(2), (2009), 530–564

    Imkeller, P., Pavlyukevich, I., and Wetzel, T.: First Exit Times for L´ evy-Driven Diffusions with Exponentially Light Jumps. The Annals of Probability, 37(2), (2009), 530–564

    2009

  36. [43]

    Imkeller, P., and Pavlyukevich, I.: First exit times of SDEs driven by stable L´ evy processes. Stoch. Proc. Appl. 116(4), (2006), 611–642

    2006

  37. [44]

    Karatzas, I., and Shreve, S.: Brownian Motion and Stochastic Calculus Springer, 1998

    1998

  38. [45]

    A measure-theoretical point of view

    L´ eonard, Ch., Roelly, S., and Zambrini, J.-C.: Reciprocal processes. A measure-theoretical point of view. Probability Surveys 11, (2014), 237–269

    2014

  39. [46]

    L., Zhai, J

    Liu, W., Song,Y. L., Zhai, J. L., and Zhang, T. S.: Large and moderate deviation principles for McKean-Vlasov SDEs with jumps. Potential Anal. 59(3), (2023)

    2023

  40. [47]

    The Annals of Probability, 15(2), (1987), 610–627

    Lynch, J., and Sethuraman, J.: Large Deviations for Processes with Independent Incre- ments. The Annals of Probability, 15(2), (1987), 610–627

    1987

  41. [48]

    Matoussi, A., Sabbagh, W., and Zhang, T.S.: Large deviation principles of obstacle prob- lems for quasilinear stochastic PDEs. Appl. Math. Optim. 83(2), (2021), 849–879

    2021

  42. [49]

    A.; Large deviations for processes with indedendent increments

    Mogulskii, A. A.; Large deviations for processes with indedendent increments. The Annals of Probability, 21(1), (1993), 202–215

    1993

  43. [50]

    A.: The Kramers problem for SDEs driven by small, accelerated L´ evy noise with exponentially light jumps

    De Oliveira Gomes, A., and H¨ ogele, M. A.: The Kramers problem for SDEs driven by small, accelerated L´ evy noise with exponentially light jumps. Stochastics and Dynamics, 21 (04), (2021), 2150019

    2021

  44. [51]

    E.: Large deviations and Metastability

    Olivieri, E., and Vares, M. E.: Large deviations and Metastability. Cambridge University Press, 2005

    2005

  45. [52]

    The Annals of Mathematical Statistics, 39(3), (1968), 936–937

    Orey, S.: On continuity properties of infinitely divisible distribution functions. The Annals of Mathematical Statistics, 39(3), (1968), 936–937

    1968

  46. [53]

    Pavlyukevich, I.: First exit times of solutions of stochastic differential equations driven by multiplicative L´ evy noise with heavy tails.Stoch. and Dyn. 11(2-3), (2011), 495–519

    2011

  47. [54]

    Stochastic Processes and their Applications 126, (2016), 1285–1305

    Privault, N., Yang, X., and Zambrini, J.-C.: Large deviations for Bernstein bridges. Stochastic Processes and their Applications 126, (2016), 1285–1305

    2016

  48. [55]

    V.: Convergence of Random Processes and Limit Theorems in Probability Theory

    Prokhorov, Yu. V.: Convergence of Random Processes and Limit Theorems in Probability Theory. Theory of Probability and its Applications, 1 (2), (1956), 157–214. 70

    1956

  49. [56]

    Puhalskii, A.: Large deviations and Idempotent Probability, Volume 228 of Monographd and Surveys in Pure and Applied Mathematics, Chapman and Hall/CRC, New York, 2001

    2001

  50. [57]

    American Mathematical Society, Graduate Studies in Mathematics, 162, 2015

    Rassoul-Agha, F., and Sepp¨ al¨ ainen, T.: A course on large deviations with an introduction to Gibbs measures. American Mathematical Society, Graduate Studies in Mathematics, 162, 2015

    2015

  51. [58]

    Potential Anal

    R¨ ockner, M., and Zhang, T.: Stochastic evolution equations of jump type: existence, uniqueness and large deviation principles. Potential Anal. 26, (2007), 255–279

    2007

  52. [59]

    I.: L´ evy Processes and Infinitely Divisible Distributions, Cambridge studies in advanced mathematics, Cambridge University Press, 1999

    Sato, K. I.: L´ evy Processes and Infinitely Divisible Distributions, Cambridge studies in advanced mathematics, Cambridge University Press, 1999

    1999

  53. [60]

    Schr¨ odinger, E.: ¨Uber die Umkehrung der Naturgesetze. Von E. Schr¨ odinger. (Sonderaus- gabe a. d. Sitz.-Ber. d. Preuß. Akad. d. Wiss., Phys.-math. Klasse, 1931, IX.) Verlag W. de Gruyter, Berlin, 1931

    1931

  54. [61]

    1961, Springer Verlag 2009

    Siegert, W.: Local Lyapunov Exponents: Sublimating Growth Rates of Linear Random Differential Equations, Lecture notes in Mathematics, vol. 1961, Springer Verlag 2009

    1961

  55. [62]

    V.: Limit Theorems for Stochastic Processes

    Skorokhod, A. V.: Limit Theorems for Stochastic Processes. Theory Prob. Applications 1, (1956), 261–290

    1956

  56. [63]

    Swiech, A., and Zabczyk, J.: Large Deviations for Stochastic PDE with L´ evy Noise,Journal of Functional Analysis 260(3), (2010), 674–723. ´

    2010

  57. [64]

    Varadhan, S. R. S.: Large Deviations and Applications. CBMS-NSF Regional Conference Series in Applied Mathematics vol.46, Soc. for Ind. and Appl. Math. 1984

    1984

  58. [65]

    Varadhan, S. R. S.: Special invited paper: Large deviations The Annals of Probability 36(2), (2008), 397–419

    2008

  59. [66]

    Varadhan, S. R. S.: Diffusion processes in a small time interval. Comm. Pure Appl. Math. 20, (1967), 659–685

    1967

  60. [67]

    Varadhan, S. R. S.: . Large Deviations and Applications. SIAM, Philadelphia. Comm. Pure Appl. Math. 20, (1984), 659–685

    1984

  61. [68]

    Diploma Thesis, Humboldt-Universit¨ at zu Berlin, 2007

    Wetzel, T.: Austrittszeiten f¨ ur dynamische Systeme mit L´ evy’schem Rauschen. Diploma Thesis, Humboldt-Universit¨ at zu Berlin, 2007

    2007

  62. [69]

    Wetzel, T.: LDPs f¨ ur bedingte L´ evyprozesse, Verfeinerte Analyse von First Exit Problemen. Ph.D. Thesis, Humboldt-Universit¨ at zu Berlin, 2021

    2021

  63. [70]

    Wittich, O.: An explicit local uniform large deviation bound for Brownian bridges, Statist. Probab. Lett. 73 (1), (2005), 51–56

    2005

  64. [71]

    L.: Large deviations for stochastic porous media equation on general measure spaces

    Wu, W.N., and Zhai, J. L.: Large deviations for stochastic porous media equation on general measure spaces. J. Differential Equations 269(11), (2020), 10002–10036

    2020

  65. [72]

    L.: Large deviations for stochastic generalized porous media equa- tions driven by L´ evy noise.SIAM J

    Wu, W.N., and Zhai, J. L.: Large deviations for stochastic generalized porous media equa- tions driven by L´ evy noise.SIAM J. Math. Anal. 56(1), (2024), 1–42

    2024

  66. [73]

    L.: (2018)

    Xiong, J., and Zhai, J. L.: (2018). Large deviations for locally monotone stochastic partial differential equations driven by L´ evy noise.Bernoulli 24(4A), (2018), 2842–2874. 71

    2018

  67. [74]

    Xu, T., and Zhang, T. S.: . Large deviation principles for 2-D stochastic Navier-Stokes equations driven by L´ evy processes.J. Funct. Anal. 25 7(5), (2009), 1519-1545

    2009

  68. [76]

    L., and Zhang, T .S.: Large deviations for SPDEs of jump type

    Yang, X., Zhai, J. L., and Zhang, T .S.: Large deviations for SPDEs of jump type. Stoch. Dyn. 15(4), (2015), 1550026

    2015

  69. [77]

    L., and Zhang, T

    Zhai, J. L., and Zhang, T. S.: Large deviations for 2-D stochastic Navier-Stokes equations driven by multiplicative L´ evy noises.Bernoulli 21 (4), (2015), 2351-2392. 72

    2015

Pith tools

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