The singular Gaussian scaling limit for a tagged particle in the infinite-dimensional Dyson model at beta=2
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.
Ada: We're looking at the paper: "The singular Gaussian scaling limit for a tagged particle in the infinite-dimensional Dyson model at beta=2".
Emmy: Here's the gist of the paper. The paper proves that a logarithmic central limit theorem exists for a tagged particle in infinite-dimensional Dyson Brownian motion at inverse temperature β = 2, showing its fluctuation is subdiffusive and slower than classical short-range systems The main results establish the logarithmic CLT and the corresponding multi-time scaling limit The infinite-dimensional Dyson Brownian motion (DBM) is the bulk scaling limit of finite DBM and has the sine point process sine2 as its invariant distribution To be more precise, we consider an infinite system of Brownian particles on R with logarithmic repulsion at inverse temperature β = 2 The drift is the repulsive force associated with the logarithmic pair potential − log x Logarithmic Asymptotics for the Variance The exact variance formula is now established for the physical current, and the logarithm comes from frequencies between order t − 1 and a fixed positive scale Theorem 2.11 states that as t → ∞, Var(J(t)) = 1/π squared log t + O(1) This result is derived by isolating the small frequencies and extracting the logarithm from the remaining band CLT for the Current The CLT for the current J(t) as t → ∞ is established, stating that J(t) p Var(J(t)) ⇒ N (0, 1) The proof involves controlling higher cumulants, showing that κm(Xt) ≤ m!
Ada: C m log t σm/t = O (log t) 1−m/2 −→ 0 This leads to the standard Gaussian limit via the method of moments CLT for the Tagged Particle The paper deduces a central limit theorem for the displacement of a tagged particle from Theorem 3.1, which states that as t → ∞, π (X0(t) − X0(0)) √ log t ⇒ N (0, 1) This is achieved by showing that the normalized statistic Xt/σt converges to a standard Gaussian process via Slutsky’s theorem Finite-Dimensional Process CLT The one-time CLT is extended to finite-dimensional convergence under the homogenised time scale, proving that for every finite set of times, (Jα(u1),..., Jα(un)) ⇒ (G(u1),..., G(un)) This is achieved by controlling the cumulant generating function and using the Cramér–Wold theorem Tagged-Particle Process: Multi-Time Scaling Limit The final result is that for every finite collection u1,..., un > 0, (Xα(u1),..., Xα(un)) ⇒ (G(u1),..., G(un)) This is proven by applying the same arguments to each fixed linear combination of currents at times u1/α,..., un/α The covariance structure of the limit G has covariance (5.3) References
AS26: T.
Emmy: Assiotis and K.
Ada: Suzuki, Collision and non-collision for diffusions on configuration space, preprint, arXiv:2607.22333, 2026
Bil95: P.
Paper summary: Ada: Now, to break down what this paper "The singular Gaussian scaling limit for a tagged particle in the infinite-dimensional Dyson model at beta=two" actually does, it establishes a logarithmic central limit theorem for the movement of one specific particle in an infinite system of Brownian particles <ref:2611.08365#pg1,The singular Gaussian scaling limit for a tagged particle in the infinite>.
Emmy: The thesis is that under these conditions, the fluctuation of that tagged particle has this specific scaling behavior, which is order the square root of log t as time goes to infinity. They start from the sine two point process as their starting point for this analysis.
Kolya: What’s interesting here for me is how they define things. They consider an infinite system of Brownian particles on the real line with a logarithmic repulsion force at inverse temperature beta equal to two, and the drift is that repulsive force associated with the logarithmic pair potential, which is negative log x.
Ada: Right. And their main result establishes this logarithmic central limit theorem and identifies it as a multi-time scaling limit of a tagged particle with the logarithmic scale of square root log t existing in finite-dimensional distribution convergence.
Emmy: They further identify this limit with the singular Gaussian field that has zero Hurst parameter, which is really interesting because it provides a realization of that singular Gaussian field as the scaling limit of a tagged particle raised by D¨urr, Goldstein and Lebowitz.
Kolya: So you’re telling me they’ve managed to pin down the exact nature of this limiting process by connecting it to a known structure in other contexts. What does that connection tell us about the physical intuition?
Ada: It tells us that even in these highly complex, infinite-dimensional systems, the long-time behavior simplifies to a very specific type of Gaussian field dictated by zero Hurst parameter.
Emmy: It suggests there is a deep structural principle governing how these particles interact over long periods that manifests as this specific scaling limit.
Conclusion: Ada: Looking at the title, "The singular Gaussian scaling limit for a tagged particle in the infinite-dimensional Dyson model at beta=two", it really captures the essence of what they found <ref:2611.08365#pg1,The singular Gaussian scaling limit for a tagged particle in the infinite>. It’s about finding that specific Gaussian behavior under these particular conditions.
Emmy: The paper shows that despite the complexity of an infinite system with logarithmic repulsion, we can pin down a precise scaling limit for a single particle's movement, which is governed by log t growth.
Kolya: For someone who only listens to this show, what does this mean in practical terms? It means if you’re studying systems with long-range repulsive interactions, the fluctuations you expect are much slower than standard models suggest.
Ada: It means the system exhibits a kind of rigidity where the particle's spread is constrained by that logarithmic repulsion. The fact that it scales like log t instead of something faster tells us about the specific constraints on motion in this infinite setting.
Emmy: The authors are using this result to identify a particular singular Gaussian field, which gives them a concrete structure to describe what’s happening at the limit. It moves the discussion from just observing slow growth to defining the exact mathematical shape of that limit.
Kolya: I think the implication is that we can use this identification with the zero Hurst parameter field to understand other related infinite systems where those same types of long-range interactions are present.
Ada: That’s right. It’s not just about a single result; it's about finding a universal structure in how these infinite interacting particle systems behave at long times.
Chubu University · Durham University
math.PR, math-ph, math.MP
Submitted: 2026-11-05
Updated: 2026-11-09
Comments: V2: added two references of Dandekar-Krapivsky-Mallick and streamlined minor presentations
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 92/100
The gist: The paper proves that a logarithmic central limit theorem exists for a tagged particle in infinite-dimensional Dyson Brownian motion at inverse temperature β = 2…
Key concepts
- Infinite-dimensional Dyson Brownian Motion (DBM)
- This model describes an infinite system of Brownian particles on the real line where particles repel each other logarithmically. It is a key object studied in this paper, serving as the bulk scaling limit for finite DBM systems and possessing a sine2 invariant distribution.
- Logarithmic Asymptotics for Variance
- The variance of the physical current $J(t)$ in this system grows logarithmically with time: $\text{Var}(J(t)) \sim 1/\pi^2 \log t$. This logarithmic growth is crucial because it dictates the slower, subdiffusive nature of the particle's fluctuations compared to classical systems.
- Logarithmic Central Limit Theorem (CLT)
- The paper proves that the normalized displacement of a tagged particle converges to a standard Gaussian distribution as time $t \to \infty$. This convergence is established by controlling higher-order cumulants, showing they decay sufficiently fast relative to the logarithmic variance growth.
- Tagged Particle Displacement Scaling
- For a tagged particle $X_0(t)$, the normalized displacement converges to a standard Gaussian process: $\pi (X_0(t) - X_0(0)) / \sqrt{\log t} \Rightarrow N (0, 1)$. This result is derived by relating the tagged particle's variance to the logarithmic growth rate of the current's variance.
Terminology
Summary
The paper proves that a logarithmic central limit theorem exists for a tagged particle in infinite-dimensional Dyson Brownian motion at inverse temperature β = 2, showing its fluctuation is subdiffusive and slower than classical short-range systems
Main Results
The main results establish the logarithmic CLT and the corresponding multi-time scaling limit The infinite-dimensional Dyson Brownian motion (DBM) is the bulk scaling limit of finite DBM and has the sine point process sine2 as its invariant distribution To be more precise, we consider an infinite system of Brownian particles on R with logarithmic repulsion at inverse temperature β = 2 The drift is the repulsive force associated with the logarithmic pair potential − log x
Logarithmic Asymptotics for the Variance
The exact variance formula is now established for the physical current, and the logarithm comes from frequencies between order t − 1 and a fixed positive scale Theorem 2.11 states that as t → ∞, Var(J(t)) = 1/π squared log t + O(1) This result is derived by isolating the small frequencies and extracting the logarithm from the remaining band
CLT for the Current
The CLT for the current J(t) as t → ∞ is established, stating that J(t) p Var(J(t)) ⇒ N (0, 1) The proof involves controlling higher cumulants, showing that κm(Xt) ≤ m! C m log t σm/t = O (log t) 1−m/2 −→ 0 This leads to the standard Gaussian limit via the method of moments
CLT for the Tagged Particle
The paper deduces a central limit theorem for the displacement of a tagged particle from Theorem 3.1, which states that as t → ∞, π (X0(t) − X0(0)) √ log t ⇒ N (0, 1) This is achieved by showing that the normalized statistic Xt/σt converges to a standard Gaussian process via Slutsky’s theorem
Finite-Dimensional Process CLT
The one-time CLT is extended to finite-dimensional convergence under the homogenised time scale, proving that for every finite set of times, (Jα(u1),..., Jα(un)) ⇒ (G(u1),..., G(un)) This is achieved by controlling the cumulant generating function and using the Cramér–Wold theorem
Tagged-Particle Process: Multi-Time Scaling Limit
The final result is that for every finite collection u1,..., un > 0, (Xα(u1),. p. 1.., Xα(un)) ⇒ (G(u1),. p. 1.., G(un)) This is proven by applying the same arguments to each fixed linear combination of currents at times u1/α,. p. 3.., un/α The covariance structure of the limit G has covariance (5. p. 23)
References
[AS26] T. Assiotis and K. p. 5 Suzuki, Collision and non-collision for diffusions on configuration space, preprint, arXiv:2607. p. 522333, 2026 [Bil95] P. p. 0 Billingsley, Probability and Measure, third edition, John Wiley & Sons, New York, 1995 [Bor11] A. p. 1 Borodin, Determinantal point processes, in The Oxford Handbook of Random Matrix Theory, G. p. 0 Akemann, J. Baik, and P. Di Francesco (eds.), Oxford University Press, 2011, 231–249; arXiv:0911. p. 01153 [BD14] J. Breuer and M. Duits, The Nevai condition and a local law of large numbers for orthogonal polynomial ensembles, Advances in Mathematics 265 (2014), 441–484 [Bri91] C. p. 0 Brislawn, Traceable integral kernels on countably generated measure spaces, Pacific Journal of Mathematics 150 (1991), no. p. 4 2, 229–240 [DT17] A. p. 4 Dembo and L. p. 4-C. Tsai, Equilibrium fluctuation of the Atlas model, Ann. p. 4 Probab. 45 (2017), no. 6B, 4529–4560 [DGL85] D. p. 0 D¨urr, S. Goldstein, and J. p. 0 L. Lebowitz, Asymptotics of particle trajectories in infinite one-dimensional systems with collisions, Communications on Pure and Applied Mathematics 38 (1985), no. p. 0 5, 573–597 [FOT11] M. Fukushima, Y. Oshima and M. Takeda, Dirichlet Forms and Symmetric Markov Processes, second revised and extended edition, De Gruyter Studies in Mathematics, vol. 19, Walter de Gruyter, Berlin, 2011 [Har65] T. p. 0 E. Harris, Diffusion with “collisions” between particles, Journal of Applied Probability 2 (1965), no. p. 0 2, 323–338 [KT10] M. p. 1 Katori and H. Tanemura, Non-equilibrium dynamics of Dyson’s model with an infinite number of particles, Communications in Mathematical Physics 293 (2010), 469–497 [KT11] M. p. 2 Katori and H. Tanemura, Markov property of determinantal processes with extended sine, Airy, and Bessel kernels, Markov Processes and Related Fields 17 (2011), no. p. 0 4, 541–580 [Osa04] H. p. 5 Osada, Non-collision and collision properties of Dyson’s model in infinite dimension and other stochastic dynamics whose equilibrium states are determinantal random point fields, in Stochastic Analysis on Large Scale Interacting Systems, T. Funaki and H. Osada (eds. p. 0), Advanced Studies in Pure Mathematics, vol. p. 0 39, Mathematical Society of Japan, Tokyo, 2004, pp. p. 4 325–343 [Osa12] H. p. 1 Osada, Infinite-dimensional stochastic differential equations related to random matrices, Probability Theory and Related Fields 153 (2012), no. p. 1 3–4, 471–509 [Osa16] H. p. 4 Osada, Self-diffusion constants of non-colliding interacting Brownian motions in one spatial dimension, RIMS Kˆokyˆuroku Bessatsu B59 (2016), 253–272 [OT16] H. p. 4 Osada and H. p. 0 p. 0 Tanemura, Strong Markov property of determinantal processes with extended kernels, Stochastic Processes and their Applications 126 (2016), no. p. 1 1, 186–208 [OT20] H. p. 1 Osada and H. Tanemura, Infinite-dimensional stochastic differential equations and tail σ-fields, Probability Theory and Related Fields 177 (2020), no. p. 1 3–4, 1137–1242 [Sim77] B. Simon, Notes on infinite determinants of Hilbert space operators, Advances in Mathematics 24 (1977), no. p. 1 3, 244–273 [Sos00] A. Soshnikov, Determinantal random point fields, Russian Mathematical Surveys 55 (2000), no. p. 0 5, 923–975 [Spi69] F. p. 0 Spitzer, Uniform motion with elastic collision of an infinite particle system, Journal of Mathematics and Mechanics 18 (1969), no. p. 0 10, 973–989
Pruning redundant reasoning tokens
The single most informative sentence about the paper's subject and finding is that the fluctuation is of order √log t, which is subdiffusive and substantially slower than the polynomial fluctuations of classical short-range interacting particle systems
How the training signal is built
The proof relies on several key steps to establish the scaling limit and CLT
- The variance of the tagged-particle displacement grows only logarithmically in time, EX0(t) − X0(0)2 ∼ 1/π squared log t, t → ∞ This is based on the estimate of the current J(t) The proof involves approximating the pathwise current by compactly supported linear statistics and showing that its variance converges to that of J(t) 2
- The CLT for the current J(t) is proven by controlling higher cumulants, showing that κm(Xt) ≤ m! C m log t σm/t = O (log t) 1−m/2 −→ 0 This is shown by bounding the cyclic integrals using the space-time determinantal structure of DBM The method of moments then yields the standard Gaussian limit for J(t) 3
**- The CLT for the tagged particle is transferred from the current to Xt using Slutsky’s theorem, as Var(Xt) = π−2 log t + O(1) This relies on showing that Var(Xt) − Var(J(t)) = O (log t)/t!
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved AI systems can perform a central limit theorem for tagged particles in infinite-dimensional Dyson Brownian motion at inverse temperature beta equals 2, showing that fluctuations are of order square root log t, which is subdiffusive.
-
The system can identify the scaling limit as the singular Gaussian field corresponding to the zero-Hurst parameter, providing a realization of this field as a scaling limit of a tagged particle raised by Dürr, Goldstein and Lebowitz.
-
AI systems can analyze long-range logarithmic repulsion in interacting particle systems to determine that large scale density fluctuations are strongly suppressed, leading to subdiffusive variance growth.
-
The system can derive the covariance formula for the singular Gaussian field limit as E[G(u)G(v)] equals one half plus one half if u equals v, and one half otherwise.
-
AI systems can establish a multi-time scaling limit where the convergence is in finite dimensional distribution and identify it with a singular Gaussian field with Hurst parameter H equals zero.
-
The system can calculate the variance of the current across the origin as Var(J(t)) equals one half pi over two log t plus O one, which is essential for proving the central limit theorem for the current.
-
The AI can prove that for a fixed m greater than or equal to 3, kappa m (Xt) / sqrt log t goes to zero as t approaches infinity, establishing the Gaussian limit through moment-cumulant estimates.
-
The system can establish a central limit theorem for the current J(t) as t approaches infinity, showing that pi J(t) / sqrt log t converges in distribution to N (0, 1).
-
The AI can prove that for every finite collection of times u1 through un, the joint convergence of scaled currents follows the limiting Gaussian field covariance structure defined by G.
-
The system can deduce a central limit theorem for the displacement of a tagged particle X0(t) from Theorem 3.1, showing that pi (X0(t) - X0(0)) / sqrt log t converges in probability to N (0, 1).
Sources
Related papers
- Rank-One HCIZ Integrals and Multiplicative Convolutions
- Repeated averaging on random d-regular graphs
- Precise cover times for branching random walks on Hamming graphs: (iterated) logarithmic corrections
- Markov Chain CLTs: Resolving Open Problems
- Marked stable L'evy processes in Brownian local time
- A Sharp Refinement of Samuels' Inequality via Cap Moves