Integrability and Chaos via fractal analysis of Spectral Form Factors: Gaussian approximations and exact results
Lorenzo Campos Venuti, Jovan Odavić, Alioscia Hamma · 2025
Read Integrability and Chaos via fractal analysis of Spectral Form Factors: Gaussian approximations and exact results by Lorenzo Campos Venuti, Jovan Odavić, Alioscia Hamma (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-05199.txt.
Introduction
Lorenzo Campos Venuti and co-authors propose, in a 2025 paper, reading a quantum system's chaos from the fractal shape of a random walk hidden in its energy spectrum. The spectral form factor, a standard chaos probe, can be drawn as a wandering path in the plane. The authors suggest measuring the roughness of that path's frontier. For chaotic systems they conjecture the roughness approaches a universal value tied to ordinary Brownian wandering.
Simulations bear the idea out: chaotic models land near the universal value, while tidy integrable ones sit at the smooth extreme. A surprise case, walkers from an exactly solvable method, groups with the chaotic side. The team also proves when the walk turns truly Brownian and computes exact moments for uneven-step walks, settling a classical side problem. Temperature limits where the tidy approximation breaks are mapped too.
Catalog record
| CREATOR | Lorenzo Campos Venuti, Jovan Odavić, Alioscia Hamma |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 83 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-05199
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Integrability and Chaos via fractal analysis of Spectral Form Factors: Gaussian approximations and exact results Lorenzo Campos Venuti, Jovan Odavić, Alioscia Hamma Abstract We establish the mathematical equivalence between the spectral form factor, a quantity used to identify the onset of quantum chaos and scrambling in quantum many-body systems, and the classical problem of statistical characterization of planar random walks. We thus associate to any quantum Hamiltonian a random process on the plane. We set down rigorously the conditions under which such random process becomes a Wiener process in the thermodynamic limit and the associated distribution of the distance from the origin becomes Gaussian. This leads to the well known Gaussian behavior of the spectral form factor for quantum chaotic (non-integrable) models, which we show to be violated at low temperature. For systems with quasi-free spectrum (integrable), instead, the distribution of the SFF is Log-Normal. We compute all the moments of the spectral form factor exactly without resorting to the Gaussian approximation. Assuming degeneracies in the quantum chaotic spectrum we solve the classical problem of random walker taking steps of unequal lengths. Furthermore, we demonstrate that the Hausdorff dimension of the frontier of the random walk, defined as the boundary of the unbounded component of the complement, approaches 1 for the ‘integrable’ Brownian motion, while the non-integrable walk approaches that obtained by the Schramm-Loewner Evolution (SLE) with the fractal dimension 4 / 3 4/3 . Additionally, we numerically show that ‘Bethe Ansatz’ walkers fall into a category similar to the non-integrable walkers. I Introduction The spectral form factor (SFF) S(t) has the form S(t)=\left|\chi(t)\right|^{2},\,\,\chi(t)=\tr\left(\rho e^{-itH}\right)\,, (1) where H is the Hamiltonian under consideration and \rho is a possibly not normalized quantum state. The number of interesting physical problems that can be written, often exactly, in this form, either quantum or classical, is frankly quite amazing. Lagrange showed that the study of the motion of perihelia is connected to the argument of an expression like \chi(t) with \rho={\mathrm{1\mkern-4.8muI}} ( Lagrange 1781 ) . For a pure quantum state \rho=|\psi\rangle\langle\psi| , S(t) is known as Loschmidt echo or survival probability and is connected to the theory of Fermi-edge singularity in metals ( Schotte and Schotte 1969 ) , the statistics of the work done after a quench ( Silva 2008 ) , the approach to thermodynamic equilibrium ( Campos Venuti and Zanardi 2010 ; Campos Venuti et al. 2013 ; Campos Venuti and Zanardi 2014 ) and (especially for \rho=1 ) quantum chaos ( Berry 1985 ; Brézin and Hikami 1997 ; Prosen 1998 ; Prosen 1999 ; Kunz 1999 ; Cotler et al. 2017a ; Roy and Prosen 2020 ; Šuntajs et al. 2021 ; Roy et al. 2022 ; Altland et al. 2025 ) . More recently, the SFF has found applications in quantum gravity ( Cotler et al. 2017b ; Cáceres et al. 2022 ; Choi et al. 2023 ) . Using the SYK as a model for a black hole, the late-time behavior of the SFF probes the discreteness of the black hole spectrum. Three regimes have been identified for the SFF. A short-time regime (the “slope”) where S(t) is observed to be self-averaging, a long time “plateau” where S(t) is approximately constant but subject to large fluctuations, and an intermediate regime connecting the two (the “ramp”). It was recognized already in ( Campos Venuti and Zanardi 2010 ; Liu 2018 ) that a cumulant expansion implies S(t)\simeq e^{-\left(\left\langle H^{2}\right\rangle-\left\langle H\right\rangle^{2}\right)t^{2}} for short times, which, due to the self-averaging property of the SYK moments \left\langle H^{2p}\right\rangle , implies the observed universal short-time regime. For larger time the cumulant expansion breaks down and S(t) start oscillating erratically. It is exactly this challenging regime we shall put more emphasis here. It has been known for quite some time that the value of the plateau is related to a particular Gaussian approximation Prange 1997 ; Šuntajs et al. 2021 (we will come back to this point below). Yet another way to look at Eq. ( 1 ) is as the position of a random walker on the complex plane. This analogy was first utilized by Rayleigh ( Rayleigh 1905 ; Pearson 1905 ) and the celebrated Rayleigh distribution is precisely the distribution of the distance of the walker from the origin \left|\chi(t)\right| , under the above mentioned Gaussian approximation. In this paper, we further exploit the random walk analogy. The entire path of the walker is seen as a random process of discrete time, of which \left|\chi(t)\right| represents the modulus of the last position. We set down rigorously the conditions under which the Gaussian approximation for \left|\chi(t)\right| is legitimate. We show that these hypotheses are generally satisfied in the high-temperature phase, but fail at low temperature and for \rho=|\psi\rangle\langle\psi| for a small quench at a critical point. Under similar hypotheses, the entire path becomes a 2D Wiener process in a precise scaling limit. The properties of the random walk, seen as a fractal, reflect the integrability or lack thereof of the Hamiltonian H . In particular, our numerics for non-integrable models, show that the dimension of the frontier of the random walk is close to 4/3 , which is the value expected for the Wiener process, while it is close to 1 for quasi-free systems. While, in the generic non-integrable case, \left|\chi(t)\right| is a sum of independent variables, for quasi-free models, \left|\chi(t)\right| is rather a product of independent variables and its distribution is log-normal. Finally, we also give the exact expression for the moments of the SFF beyond the Gaussian approximation as a closed-form recursion and in full generality accounting for possible degeneracies in the spectrum. These exact results extend, in the plateau region, the beyond-Gaussian approximation obtained in the ramp-plateau regime of Altland et al. 2025 (see also Haake 2001 ; Flack et al. 2020 ; Šuntajs et al. 2020 ; Chan et al. 2021 ; Winer and Swingle 2022 ; Legramandi and Talwar 2024 ; Kumar et al. 2025 ), and can serve as a check for further approximations on the time dependence of the SFF. II The Fractal Geometry of Spectral Statistics Figure 1: Fractals and their frontiers, in black, corresponding to physical models: non-integrable (left) vs integrable (right). Color from blue to pink corresponds to increasing time-steps of the random walks. The frontier is essentially the boundary of the fractal without the inner islands, see text for details. The non-integrable Hamiltonian is the XXZ model with next nearest neighbor interactions (see Eq. ( 21 )), with (\Delta,\alpha)=(0.4,0.5) . The integrable Hamiltonian is the XY model with parameters (h,\gamma)=(0.2,0.3) ; see Appendix H.6 . For each “system size” L , we assume the Hamiltonian H lives in a Hilbert space of dimension D e.g. for qubits D=2^{L} . We allow the Hamiltonian to have degenerate eigenvalues. Indeed, general physical models, even non-integrable, chaotic ones, possess several symmetries and degeneracies. Using the spectral resolution of the Hamiltonian, H=\sum_{j=1}^{N_{B}}E_{j}\Pi_{j} ( E_{j} eigenenergies, \Pi_{j} eigenprojectors, N_{B} the number of blocks) we have, with d_{j}=\tr\left(\rho\Pi_{j}\right)>0 , \chi(t)=\sum_{j=1}^{N_{B}}d_{j}e^{-itE_{j}}. (2) We now show how to associate to any pair \left(H,\rho\right) a random walk on the plane. In this way, we encode information about the pair \left(H,\rho\right) into a single function, more precisely into a