Pith. sign in

REVIEW 2 major objections 4 minor 15 references

Temperature chaos in directed polymers

T0 review · 2 major / 4 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read The paper proves that in a canonical random polymer model, free energies at two inverse temperatures become independent directed landscapes when the temperatures are far apart.

desk verdict Genuinely new decoupling theorem and black-noise proof, with a repairable L2 gap in the independence argument and a load-bearing influence bound I could not fully verify in the visible text. read the letter →

arxiv 2607.18194 v1 pith:SI5Q7XPS submitted 2026-07-20 math.PR math-phmath.MP

Shirshendu Ganguly , Victor Ginsburg , Zoe Himwich This is my paper · ORCID
classification math.PRmath-phmath.MP MSC 60K3560H1582B44
keywords temperaturechaoscontinuumdirectedrandompolymerlandscapeKPZuniversalityclassnoisesensitivityblackGaussianmultiplicativewhiteresampling
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 aims to prove temperature chaos: a random system's optimal-energy landscape reorganizes completely when temperature is changed, not just slightly. It studies the continuum directed random polymer (CDRP), the canonical continuum model in the Kardar–Parisi–Zhang universality class, at two inverse temperatures beta1 and beta2 driven by the same white-noise environment. The main theorem states that when beta1 is much smaller than beta2 and both are large, the two free-energy profiles converge jointly to two independent copies of the directed landscape, the universal zero-temperature limit. This gives the first rigorous instance of energetic decoupling across temperatures in a KPZ model. A byproduct proves that the directed landscape itself is a two-dimensional black noise.

What carries the argument

The central mechanism is the thin-strip influence bound. For a spatial strip S of half-width epsilon beta2^{2/3}, the L2 sensitivity of the beta2 free energy to resampling the white noise inside S is at most epsilon^alpha for small epsilon. This is proved via an Efron–Stein inequality for white noise: the strip is tiled into boxes and the total influence is bounded by the sum of per-box influences. Each per-box influence is small because the polymer crosses a given small box with probability about the box's width in transversal units, and when it does, the energy change is controlled by the KPZ fluctuation on that box's scale. Brownian comparison estimates for the Airy line ensemble (or its

What would settle it

Simulate two coupled CDRP free-energy profiles H_beta1 and H_beta2 with beta1=o(beta2), and measure the covariance of their finite-dimensional distributions; if the covariance does not vanish as beta2/beta1 →∞, Theorem 1.7 is false. A sharper check targets the localization input: measure the quenched transversal fluctuation exponent of the CDRP polymer at large beta; if the typical width is not of order beta^{2/3}, the strip argument fails.

Watch

Extended reading notes

Core claim

Coupled through the same space-time white noise, the CDRP free energy profiles H_beta1 and H_beta2, scaled to their natural KPZ units, converge in distribution as beta1,beta2 →∞ with beta1=o(beta2) to (L1,L2), where L1 and L2 are independent directed landscapes. The theorem is proved by showing that the beta1-polymer is localized inside a thin spatial strip of width of order beta2^{2/3} around its typical path, that the beta1 free energy is nearly measurable with respect to the noise in that strip, and that resampling the noise inside any such thin strip changes the beta2 free energy negligibly. Thus the two free-energy profiles become functions of essentially independent pieces of the noise

Load-bearing premise

The argument stands on the CDRP polymer localization estimate imported from an earlier work: at inverse temperature beta and over a unit time interval, the polymer's transversal displacement has typical scale beta^{2/3} with stretched-exponential tails. If this scale is not exactly beta^{2/3}, the strip either fails to contain the beta1 polymer or the decoupling regime beta1=o(beta2) is not the correct one.

Editorial extensions

If this is right

  • If Theorem 1.7 holds, temperature chaos is rigorous in the CDRP: the free-energy profiles at well-separated temperatures are asymptotically independent, not merely decorrelated.
  • The same machinery proves the directed landscape is a two-dimensional black noise, making it the third known example after critical planar percolation and the Brownian web.
  • Polymer measures at any two distinct inverse temperatures are mutually singular almost surely, even when free energies are strongly correlated; the path-space Gibbs measures live on different energy-level sets.
  • For temperatures that are polynomially close, the coupled pair converges to two copies of the same directed landscape, so the chaos transition must occur at a nontrivial separation scale.
  • A sharp threshold is conjectured at beta2 - beta1 ~ beta1^{1/3}; the paper proves stability and decoupling on either side of that scale but leaves the threshold itself open.

Reading between the lines

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

  • One plausible extension is that the same thin-strip mechanism transfers from the CDRP to discrete lattice directed polymers in the strong-disorder regime, predicting the experimentally relevant n^{-1/6} temperature-perturbation scale; the paper only sketches this connection.
  • If the directed landscape is a two-dimensional black noise, then any finite-resolution white-noise approximation of the landscape sees it only through high-frequency components, a feature that could matter for numerical simulations of interface growth.
  • The same influence method may show that the critical two-dimensional stochastic heat flow is a three-dimensional black noise, which would be the first known example in dimension three or higher.
  • The relation between temperature and disorder chaos sketched in the paper suggests a testable equivalence: the critical temperature-perturbation scale should match the disorder-perturbation scale under a reparameterization, a prediction that finite-size simulations could probe.
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 manuscript studies temperature chaos for the continuum directed random polymer (CDRP). Its central result, Theorem 1.7, states that when two CDRP free-energy profiles at inverse temperatures β1 and β2 are coupled through the same white noise and β1,β2→∞ with β1=o(β2), the pair converges in distribution to two independent directed landscapes. The proof is built on a thin-strip influence bound (Theorem 5.2), proved via a white-noise version of the Efron–Stein inequality, together with imported convergence and localization results for the CDRP and the directed landscape. As a byproduct, the authors prove that the directed landscape is a two-dimensional black noise (Theorem 1.9), resolving a conjecture of Virág. They also prove mutual singularity of CDRP polymer measures at different temperatures and formulate a conjecture for the sharp temperature-chaos exponent.

Significance. If completed, this would be the first rigorous energetic temperature-chaos result for a canonical KPZ-universality model, with the limiting pair explicitly identified as independent copies of the universal zero-temperature object. The measure-theoretic framework for restricted CDRP partition functions and the influence estimates are likely to be reusable. The black-noise theorem resolves an open conjecture and provides a third known example of a two-dimensional black noise. The paper is unusually careful about measurability and resampling of white noise, and the overall proof architecture is coherent and detailed.

major comments (2)
  1. [§5, proof of Theorem 1.7 / Lemma 5.4] Lemma 5.4 is invoked with X_n=Hbar^ξ_{β1}(u), Z_n=Hbar^ξ_{β1,S}(u), but the lemma requires limsup E||X_n−Z_n||^2 ≤ ε. Lemma 5.1 only establishes convergence in probability. Proposition 4.14 controls the quenched probability of polymer exit on a high-probability event, but it does not, as stated, control the size of |log p| on the exceptional event where the exit probability is not small. An additional integrability statement for the restricted free energy, or a modification of Lemma 5.4 requiring only convergence in probability, is needed. As written, the independence conclusion does not formally follow from the stated lemmas.
  2. [§5, proof of Theorem 1.7] There is a second formal mismatch with Lemma 5.4. The lemma requires ||Y_n−W_n||_p → 0 for the sequences associated with each ε>0. Theorem 5.2 gives only limsup E|Hbar^ξ_{β2}−Hbar^{η_S}_{β2}|^2 ≤ ε^α for a fixed ε>0. Since ε is fixed in the application, this bound does not vanish. To obtain the desired tensorization one must either diagonalize over ε→0 or state a variant of Lemma 5.4 in which both approximation errors satisfy limsup ≤ ε. The current proof does not supply this step, leaving the central claim incomplete.
minor comments (4)
  1. [§5, Eq. (5.3)] The notation H^ξ_{β1,S} for the restricted free energy is very close to H^ξ_{β1}; consider a different symbol (e.g., H^{ξ,strip}_{β1}) to avoid confusion in a long proof.
  2. [§5, Lemma 5.1] The endpoint scaling in the strip uses factors 2^{1/3}β^{2/3}; it would help to state explicitly how these factors are absorbed by the constants in Proposition 4.14 and the convergence in Theorem 1.5.
  3. [§1.5 / §7.9] Section 7.9 is a sketch of a possible SHF result rather than a theorem. This is acceptable, but it should be labeled more prominently as heuristic/sketch so that it is not read as a claim.
  4. [References] Some citations, e.g., [QR V25], contain a space in the author field. Please check the reference formatting.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the decoupling theorem follows from external convergence/localization inputs and a new influence estimate; self-citations are background only.

full rationale

The derivation of Theorem 1.7 does not reduce to its own inputs. Marginal convergence to the directed landscape is an external input ([Wu26, Theorem 1.6], stated as Theorem 1.5); the polymer localization entering Lemma 5.1 is [DZ24, Corollary 3.5] (Proposition 4.14); and the new thin-strip influence bound (Theorem 5.2) is proved in Section 6 from the Efron–Stein inequality for white noise together with those external scaling and regularity estimates. The restricted free energy H_{β1,S} is not defined in terms of H_{β2}; the approximation H_{β1,S} ≈ H_{β1} is a consequence of the external localization estimate in the regime β1 = o(β2), not an identity imposed by construction. The independence of the approximating vectors in Lemma 5.4 comes from Proposition 2.2 (measurability of the restricted free energy) and the independent-resampling coupling, not from assuming the conclusion. Theorem 1.9 is proved through a separate discrete LPP proxy (Sections 7.5–7.6), not by assuming Theorem 1.7. The only caveat is a written-proof gap: Lemma 5.1 establishes convergence in probability for the restricted-energy difference, while Lemma 5.4 asks for an L^2 bound; this is likely repairable from the uniform fourth-moment bound (Lemma 4.5) plus localization tails, but it is a completeness/correctness issue, not circularity. Self-citations such as [GH24a, GH23a] and [GG] appear only in background or outlook and are not load-bearing.

Assumptions & free parameters 2 free parameters · 9 assumptions · 1 invented entities

The central theorem rests on three external pillars cited as theorems: (i) [AJRAS22] construction of the CDRP partition function process; (ii) [Wu26, Theorem 1.6] zero-temperature convergence of the CDRP free energy to the directed landscape, used both to identify the limits L1, L2 and to import zero-temperature estimates into Theorem 5.2; (iii) [DZ24] polymer transversal fluctuation bounds and annealed geodesic convergence. These are legitimate domain assumptions from cited prior work, not invented here, but they concentrate correctness risk. The only hand-chosen numbers are the proof exponents alpha and gamma in Section 6. No new particles, forces, or physical entities are postulated; the restricted length is a defined functional of an existing object.

free parameters (2)
  • alpha (influence exponent, Theorem 5.2 / Theorem 6.1) = unspecified; only existence of some alpha > 0 is claimed (Remark 5.3: true order expected epsilon^{1/2})
    The key estimate limsup E|H_beta2 − H_beta2^{eta_S}|^2 <= epsilon^alpha holds for some proof-dependent alpha whose value is never pinned down. This does not affect the qualitative statement of Theorem 1.7, but it makes the quantitative content of the central estimate weak and uncheckable in the truncated portion of Section 6.
  • gamma (mesh boundary exponent, Section 6.2) = any value in (0, 1/6)
    The edge boxes of height epsilon^gamma n in the Efron–Stein decomposition are a hand-chosen mesh parameter; the text states the precise value does not affect the argument. It is a proof-artifact constant, not fitted to data.
assumptions (9)
  • domain assumption [AJRAS22, Theorem 2.2]: existence and regularity of the CDRP partition function process Z^xi_beta(s,x;t,y) with chaos expansion (3.16), strict positivity, Chapman–Kolmogorov (Theorem 3.12 here)
    Used throughout Sections 2–6 as the foundation of the CDRP model; imported without proof.
  • domain assumption [Wu26, Theorem 1.6]: zero-temperature convergence of the CDRP free energy profile to the directed landscape (Theorem 1.5 here)
    Identifies the marginal limits in Theorem 1.7 and enables importing directed-landscape regularity into the proof of Theorem 5.2 (reformulated as Theorem 6.2).
  • domain assumption [DZ24, Corollary 3.5 and Theorem 1.10]: polymer transversal fluctuation bounds and annealed polymer convergence to the directed landscape geodesic (Propositions 4.14, 4.15)
    Lemma 5.1 (localization of the beta1 polymer in the thin strip) and the high-probability events in Section 6.3 rest directly on these bounds.
  • domain assumption [CHH23, Theorem 1.1]: local Brownianity of parabolic Airy/KPZ line ensembles (Proposition 4.6)
    Used in Proposition 4.8 (geodesic anticoncentration) and in the Section 6.5 box-visit probability estimates.
  • domain assumption [DOV22] and [GZ22, Lemma 3.11]: construction and regularity of the directed landscape (metric composition, scaling, Holder continuity, geodesic transversal fluctuations), Propositions 4.1–4.3, 4.12–4.13
    Core inputs for the Section 6.3 favorable events and for the restricted-length machinery of Section 7.
  • domain assumption [CG20a, CG20b]: stretched-exponential upper/lower tail bounds for the CDRP free energy (Lemma 4.5)
    Used in Lemma 6.7 to control the contribution from the complement of the Good event via Cauchy–Schwarz and a fourth-moment bound.
  • domain assumption [HP24]: the directed landscape is a one-dimensional black noise in the time direction
    Theorem 1.9 strengthens this to two-dimensional; the 1D result is used as a component of the 2D noise framework.
  • domain assumption [QRV25]: Gaussian multiplicative chaos framework for CDRP (used in Section 8)
    Theorem 1.8 (mutual singularity) is proved through this perspective; Section 8 details were not visible in the available text.
  • standard math Standard Gaussian analysis and measure theory: Wiener chaos expansion, martingale convergence, Efron–Stein inequality on product spaces, Doob–Dynkin factorization (Section 3)
    Classical background, quoted with references; not new content.
invented entities (1)
  • Restricted length L_R and sigma-algebras F_R of the directed landscape independent evidence
    purpose: Defines the 2D noise structure used in Theorem 1.9: F_R = sigma(L_R(s,x;t,y) for endpoints in rectangle R). The join property N4 requires these restricted lengths to reconstruct the landscape on unions of rectangles.
    Not a postulated physical entity: L_R is constructed from the directed landscape itself (Definition 4.9, restricted). Its properties are internally checkable and the construction is falsifiable within the mathematics; no external experimental handle is needed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Temperature chaos in directed polymers." pith.science (2026). https://pith.science/paper/SI5Q7XPS

@misc{pith2026260718194,
  author       = {Pith},
  title        = {Pith review of: Temperature chaos in directed polymers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SI5Q7XPS}},
  note         = {Machine review of arXiv:2607.18194}
}
abstract

Disordered systems such as spin glasses and polymers characteristically exhibit random energy landscapes with many macroscopically separated energetic valleys corresponding to near-ground states. This high complexity renders these systems extremely sensitive to perturbations of external parameters. For instance, the support of associated Gibbs measures may change macroscopically under such perturbations, a phenomenon known as chaos in the literature. In experiments, chaotic phenomena are typically studied via temperature perturbations. In this article, we initiate the rigorous study of temperature-chaotic properties of the continuum directed random polymer (CDRP), a canonical model in the KPZ universality class. The CDRP is driven by white noise and is parametrized by inverse temperature $\beta$, and is known [Wu '26, Das-Zhu '24] to converge in the zero-temperature limit $\beta \to \infty$ to the directed landscape constructed in [Dauvergne-Ortmann-Vir\'ag '22], the putative universal scaling limit of models in the KPZ universality class. The main result of this article considers the CDRP free energies coupled through the same white noise at a pair of inverse temperatures $(\beta_1, \beta_2)$, and shows that they decouple in the limit $\beta_2 \gg \beta_1 \gg 1$, converging to a pair of independent directed landscapes. This is the first such "energetic de-correlation across temperatures" result. Our key estimate measures the "pivotality" or "influence" of spatially thin strips in models of last passage percolation. As a byproduct, the proof strategy also allows to show that the directed landscape is a two-dimensional black noise (in the sense of [Tsirelson-Vershik '98]), previously conjectured by Vir\'ag. This provides the third known example of a two-dimensional black noise after critical planar percolation [Schramm-Smirnov '11] and the Brownian web [Ellis-Feldheim '16].

Figures

Figures reproduced from arXiv: 2607.18194 by the authors.

Figure 1
Figure 1. A simulation of lattice directed polymers on Z 2 (rotated by 45◦ ) and their free energies at three different inverse temperatures, coupled through a sin￾gle instance of an i.i.d Gaussian environment. The polymers all start at (0, 0) and travel upwards to height n with n = 214 taking north-east (1, 1) or north-west (−1, 1) steps. In the plots, the colors orange, blue, pink correspond respectively to inverse temperat… view at source ↗
Figure 2
Figure 2. Left: Exponential LPP on Z 2 . The vertices of Z 2 carry i.i.d. Exponential(1) weights. The last passage time is the maximum sum of weights along any directed (north-east/north-west) lattice path from (0, 0) to (n, 0). The blue path depicts the geodesic (maximizing path). Middle: The geodesic from the left panel is redrawn on top of a thin strip S (shaded) of height n and width εn2/3 . The key bound (1.12) asserts t… view at source ↗
Figure 3
Figure 3. Left: Temporally adjacent boxes. Right: Spatially adjacent boxes. In either case, let R3 be the rectangle comprised of R1, R2, and the interface between them. The task is to reconstruct the restricted lengths in R3 from the restricted lengths in R1, R2. These restricted lengths are respectively depicted using orange and blue “restricted geodesics.” R2. Let Gε be the σ-algebra generated by restricted lengths of paths… view at source ↗
Figures from the paper (14 more)
Figure 4
Figure 4. Figure 4: The limiting case of spatially adjacent half-infinite strips. Given the data of the restricted lengths in R1 and R2 (depicted via blue “restricted geodesics”), we seek to reconstruct the length of the orange geodesic, which can cross the interface between R1 and R2 (ve…
Figure 5
Figure 5. Figure 5: The setting of Proposition 3.19. A continuous path from (s, x) to (t, y) is drawn in black. The path belongs to some fixed cylinder set S = Tℓ j=1{X(sj ) ∈ Sj} ⊂ C([s, t]) (this constrains the horizontal locations of the blue points, but these constraints are not depic…
Figure 6
Figure 6. Figure 6: The proof of Proposition 3.19. Several times are depicted as horizontal dashed lines: black lines for s ⩽ s ′ < t′ ⩽ t (drawn here as s < s′ < t′ < t), blue lines for s1 < · · · < sℓ from Proposition 3.19, and orange lines for the δ-mesh Rδ := (s ′ , t′ )∩δZ. The blue/…
Figure 7
Figure 7. Figure 7: The vertical strip is B = [0, n] × [−εn2/3 , εn2/3 ]. As in (6.6), B is subdivided into boxes B0, B1, . . . , BM, where B1, . . . , BM−1 have height ⩽ ε 3/2n, and B0, BM have height ≍ ε γn for some γ ∈ (0, 1 6 ). Theorem 6.1 asserts a bound for the L 2 -influence of th…
Figure 8
Figure 8. Figure 8: The proof of Lemma 6.4. Shaded gray is the box [rk, rk+1] × [−ε, ε]. The KPZ-scaled polymer X(n) is conditioned to land in [−ε, ε] at times rk and rk+1, and we seek to bound the conditional probability of it wandering further than ε log(1/ε) from X(n) (rk) during [rk, …
Figure 9
Figure 9. Figure 9: Depicted are the rectangles Rδ ⊂ R, with Rδ shaded and R \ Rδ un￾shaded. We fix any u ∈ R 4 ↑ (Rδ ) and draw its endpoints in orange. By geodesic transversal fluctuation estimates (Proposition 4.13), there exists a random δ∗ > 0 such that all geodesics of height ⩽ δ 3/…
Figure 10
Figure 10. Figure 10: The proof of Property N3 (disjoint independence). We consider disjoint open rectangles R1 (unshaded) and R2 (shaded), and aim to prove independence of LR1 and LR2 . Depicted is the case where R1, R2 share a face and have the same bottom/top coordinates. By combining t…
Figure 11
Figure 11. Figure 11: ): (1) (Temporally adjacent boxes). t1 = s2 and (x1, y1) = (x2, y2), and hence R3 = (s1, t2) × (x1, y2); (2) (Spatially adjacent boxes). (s1, t1) = (s2, t2) and y1 = x2, and hence R3 = (s1, t1)×(x1, y2). R1 R2 R1 R2 a 0 b T 0 [PITH_FULL_IMAGE:figures/full_fig_p089_11.png]
Figure 12
Figure 12. Figure 12: The large black rectangles are R1 = (0, T) × (a, 0) and R2 = (0, T) × (0, b). We fix a point u∗ = (s∗, x∗;t∗, y∗) ∈ R 4 ↑ (R3), drawn in blue, where R3 = (0, T) × (a, b). The orange interval is J δ = (a + δ log4 (1/δ), b − δ log4 (1/δ)). The horizontal solid black lin…
Figure 13
Figure 13. Figure 13: The two levels of discretization. Left: The scale-δ discretization scheme from [PITH_FULL_IMAGE:figures/full_fig_p095_13.png]
Figure 14
Figure 14. Figure 14: Part of the strip [0, δ3/2 ] × R. The horizontal lines correspond to the times t0, t1, . . . , tn+1 defined in (7.24). Each horizontal line carries a copy of ε 50Z depicted as black points (figure not to scale). In solid blue is a directed path π from the bottom of th…
Figure 15
Figure 15. Figure 15: As in [PITH_FULL_IMAGE:figures/full_fig_p104_15.png]
Figure 16
Figure 16. Figure 16: Resampling in the discrete LPP model. Top: Two mesh geodesics with different endpoints are drawn (black paths with orange/blue endpoints) with the segments highlighted as in [PITH_FULL_IMAGE:figures/full_fig_p106_16.png]
Figure 17
Figure 17. Figure 17: Coarse-graining of the CDRP. Left: A trajectory of the length-n CDRP is drawn as a continuous blue path. The space-time plane is divided into unit blocks, and the blocks the CDRP passes through (shaded blue) form a coarse version of its trajectory. Right: The CDRP is …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

15 extracted references · 10 linked inside Pith

  1. [7]

    Enhanced noise sensitivity, 2D directed polymers and stochas- tic heat flow.arXiv preprint arXiv:2507.10379,

    [CD25] Francesco Caravenna and Anna Donadini. Enhanced noise sensitivity, 2D directed polymers and stochas- tic heat flow.arXiv preprint arXiv:2507.10379,

    arXiv

  2. [8]

    A class of d-dimensional directed polymers in a Gaussian environment.arXiv preprint arXiv:2603.06574,

    [COTX26] Le Chen, Cheng Ouyang, Samy Tindel, and Panqiu Xia. A class of d-dimensional directed polymers in a Gaussian environment.arXiv preprint arXiv:2603.06574,

    arXiv

  3. [11]

    Fractal geometry of the space-time difference profile in the directed landscape via construction of geodesic local times.arXiv preprint arXiv:2204.01674,

    [GZ22] Shirshendu Ganguly and Lingfu Zhang. Fractal geometry of the space-time difference profile in the directed landscape via construction of geodesic local times.arXiv preprint arXiv:2204.01674,

    arXiv

  4. [13]

    Introduction to KPZ

    [Qua12] Jeremy Quastel. Introduction to KPZ. InCurrent developments in mathematics, 2011, pages 125–194. Int. Press, Somerville, MA,

    2011

  5. [1963]

    The heat and the landscape I.arXiv preprint arXiv:2008.07241,

    [Vir20] B´ alint Vir´ ag. The heat and the landscape I.arXiv preprint arXiv:2008.07241,

    arXiv 2008

  6. [1996]

    Last passage percolation with a defect line and the solution of the slow bond problem.arXiv preprint arXiv:1408.3464,

    [BSS14] Riddhipratim Basu, Vladas Sidoravicius, and Allan Sly. Last passage percolation with a defect line and the solution of the slow bond problem.arXiv preprint arXiv:1408.3464,

    arXiv

  7. [2006]

    [Tsa24] Li-Cheng Tsai

    Unabridged republication of the 1967 original. [Tsa24] Li-Cheng Tsai. Stochastic heat flow by moments.arXiv preprint arXiv:2410.14657,

    arXiv 1967

  8. [2010]

    The directed landscape is a black noise.arXiv preprint arXiv:2404.16801,

    [HP24] Zoe Himwich and Shalin Parekh. The directed landscape is a black noise.arXiv preprint arXiv:2404.16801,

    arXiv

Show all 15 references
  1. [2014]

    Rotationally invariant first passage percolation: Concentration and scaling relations.arXiv preprint arXiv:2312.14143,

    [BSS23] Riddhipratim Basu, Vladas Sidoravicius, and Allan Sly. Rotationally invariant first passage percolation: Concentration and scaling relations.arXiv preprint arXiv:2312.14143,

    arXiv

  2. [2015]

    Stochastic heat flow is a black noise.arXiv preprint arXiv:2506.16484,

    [GT25] Yu Gu and Li-Cheng Tsai. Stochastic heat flow is a black noise.arXiv preprint arXiv:2506.16484,

    arXiv

  3. [2019]

    Geometry of the Gaussian multiplicative chaos in the Wiener space.arXiv preprint arXiv:2008.04290,

    [BM20] Yannic Br¨ oker and Chiranjib Mukherjee. Geometry of the Gaussian multiplicative chaos in the Wiener space.arXiv preprint arXiv:2008.04290,

    2008 arXiv

  4. [2021]

    The scaling limit of the longest increasing subsequence.arXiv preprint arXiv:2104.08210,

    [DV21] Duncan Dauvergne and B´ alint Vir´ ag. The scaling limit of the longest increasing subsequence.arXiv preprint arXiv:2104.08210,

    arXiv

  5. [2024]

    Geodesic switches and exceptional times in dynamical Brownian last passage percolation

    [Bha25a] Manan Bhatia. Geodesic switches and exceptional times in dynamical Brownian last passage percolation. arXiv preprint arXiv:2510.27589,

  6. [2025]

    Near-existence of bigeodesics in dynamical exponential last passage percolation.arXiv preprint arXiv:2504.12293,

    [Bha25b] Manan Bhatia. Near-existence of bigeodesics in dynamical exponential last passage percolation.arXiv preprint arXiv:2504.12293,

  7. [2026]

    The Green’s function of the parabolic Anderson model and the continuum directed polymer.arXiv preprint arXiv:2208.11255,

    [AJRAS22] Tom Alberts, Christopher Janjigian, Firas Rassoul-Agha, and Timo Sepp¨ al¨ ainen. The Green’s function of the parabolic Anderson model and the continuum directed polymer.arXiv preprint arXiv:2208.11255,

    arXiv

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.