Noise effects on the diagnostics of quantum chaos
Tingfei Li · 2025
Read Noise effects on the diagnostics of quantum chaos by Tingfei Li (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-2503-01279.txt.
Introduction
Tingfei Li, writing alone in 2025, examines how noise corrupts the standard tools used to detect quantum chaos. Three diagnostics come under study: the spectral form factor, Krylov complexity, and out-of-time-order correlators. Real quantum devices never run in silence, so knowing when these tools stop working matters for experiments on today's hardware.
Using a closed quantum system with added white noise, Li shows that stronger noise suppresses all three measures exponentially. In the strong-noise limit, none of them can tell chaotic motion apart from regular motion. The spectral form factor suffers most, decaying until its signature pattern vanishes. The results argue that chaos diagnostics need redesign for noisy, open systems.
Catalog record
| CREATOR | Tingfei Li |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 62 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-01279
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Noise effects on the diagnostics of quantum chaos Tingfei Li Abstract This paper investigates the effects of noise on the diagnostics of quantum chaos, focusing on three primary tools: the spectral form factor (SFF), Krylov complexity, and out-of-time correlators (OTOCs). Utilizing a closed quantum system model with white noise, we demonstrate that increasing noise strength leads to an exponential suppression of these diagnostic measures. Specifically, our findings reveal that in the strong noise limit, the SFF, two-point correlation function, and OTOCs become ineffective in distinguishing chaotic behavior. The SFF is particularly impacted, exhibiting a significant decay that obscures its ability to identify quantum chaos. This study highlights the challenges posed by environmental noise in accurately diagnosing quantum chaotic systems and suggests that traditional methods may require adaptation to remain effective in realistic open quantum systems. Our results underscore the need for further research into robust diagnostic techniques that can account for noise-induced effects in quantum chaotic systems. I Introduction Motivation Quantum chaos has emerged as a pivotal area of research, bridging the realms of quantum mechanics and classical chaotic systems. As a quantum analogue to classical chaos, it plays a critical role in understanding complex many-body quantum systems, and it has garnered significant attention due to its profound implications in fields such as black hole physics and quantum gravity. In classical systems, chaos is characterized by the exponential growth of uncertainty stemming from minor variations in initial conditions [ 1 , 2 , 3 ] , which can be easily conceptualized through the dynamics governed by ordinary differential equations (ODEs) [ 4 ] . In contrast, quantum systems evolve according to linear equations that govern the density matrix. This means that irregular motion in quantum mechanics cannot be characterized by extreme sensitivity to small changes in initial conditions [ 5 , 6 ] . This intrinsic property makes it difficult to identify chaotic behavior in quantum contexts, especially in open quantum systems, where certain density matrices may exhibit exponential decay. As a result, the observable differences between quantum states diminish over time, complicating the interpretation of quantum chaos as an emergent feature arising from slight initial perturbations. Due to these challenges, researchers have focused more on the statistical characteristics of quantum systems. Following the well-known conjecture by Bohigas, Giannoni, and Schmit [ 7 ] , certain universal features of spectral fluctuations in classically chaotic systems have been found to align with predictions from random-matrix theory [ 8 , 9 ] . Consequently, the spectral form factor (SFF) [ 10 , 11 , 12 , 13 ] has been proposed as a diagnostic tool for quantum chaos. In a broad sense, a chaotic system acts as an efficient information scrambler [ 14 , 15 , 16 ] . This is indicated by the exponential decay of out-of-time-order correlators (OTOCs) [ 17 , 18 , 19 , 20 , 21 ] , efficient operator spreading [ 22 , 23 ] , small fluctuations in purity [ 24 ] , and information scrambling [ 25 , 26 ] . In this paper, we examine three methods for diagnosing quantum chaos: OTOC, Krylov complexity [ 27 , 28 , 29 , 30 ] , and SFF. The OTOC is a four-point correlator, while Krylov complexity is derived from a two-point function using a thermal state with inverse temperature \beta , emphasizing the importance of the energy spectrum’s statistical behavior. The SFF, on the other hand, relies solely on the energy spectrum of the system. Notably, the SFF remains a valuable indicator of quantum chaos even in open systems affected by energy dephasing [ 31 ] . This study investigates how noise affects these diagnostic measures by using a closed quantum system model exposed to white noise. Our results show that as the noise strength increases, there is an exponential suppression of the SFF, two-point correlation functions, and OTOCs, which ultimately makes them ineffective for distinguishing chaotic behavior in high noise conditions. This research highlights the need for more robust diagnostic techniques that can address the challenges posed by environmental noise in realistic open quantum systems, paving the way for future studies on quantum chaos in noisy environments. The Model We conduct the calculations in the energy basis, ensuring that the Hamiltonian is diagonal. Our focus is primarily on noise without time correlation, specifically white noise. For simplicity, we consider a closed quantum system with a D -dimensional Hilbert space and examine the effects of noise. In this context, the time evolution is governed by the noisy Hamiltonian: \displaystyle H_{ij}=E_{i}\delta_{ij}+\eta_{ij}(t),i,j=1,2,\ldots,D (1) where \eta_{ij}(t) is the white noise with zero mean and a non-vanishing variance \displaystyle\mathbb{E}(\eta_{ij}(t))=0,~\mathbb{E}(\eta_{ij}(t)\eta_{kl}(t^{\prime}))=\lambda_{ijkl}\delta(t-t^{\prime})\ . (2) To distinguish from the state expectation \langle\bullet\rangle=\text{Tr}(\hat{\rho}(t)\bullet) , we use \mathbb{E}(\bullet) to denote the ensemble average of the noise. For general \lambda_{ijkl} , obtaining an analytical solution is challenging. Therefore, in this paper, we focus on two special cases: \lambda_{ijkl}=\lambda_{ij}\delta_{il}\delta_{jk} (GUE noise) and \lambda_{ijkl}=\lambda_{ij}\frac{\delta_{ik}\delta_{jl}+\delta_{il}\delta_{jk}}{2} (GOE noise). We consistently use J to denote the strength of the noise, i . e ., \lambda_{ijkl}\sim J . Consequently, we sometimes denote the noise ensemble average of A as A_{J}\equiv\mathbb{E}(A) . We adopt similar notation as in [ 32 ] . For generic k -replica observables in our model, we ultimately obtain an effective time evolution on 2k -contours: \displaystyle\mathcal{U}_{k}\equiv\mathbb{E}[U_{t}^{\otimes k}\otimes U_{t}^{*\otimes k}]\equiv e^{\mathcal{L}_{k}t} (3) where U_{t}=\mathcal{T}e^{-\text{i}\int_{0}^{t}H(\tau)d\tau} is the noisy time evolution operator. Writing H=H_{0}+\eta , we denote U_{t}^{(0)}=e^{-\text{i}H_{0}t} as the noise-free time evolution operator. For an operator A , we define \displaystyle A_{t}\equiv U_{t}^{\dagger}AU_{t},A_{t}^{(0)}\equiv U_{t}^{(0)\dagger}AU_{t}^{(0)}\ . (4) In this paper, we primarily study the cases k=1,2 , which are needed for the calculation of quantum chaos diagnostics. Summary of results We investigate the effects of noise on three quantum chaos diagnostics: SFF, Krylov complexity, and OTOC. As an example, we consider GUE noise with \lambda_{ij}=\frac{J}{D} . We find that the noise-averaged SFF consists of a spectral-dependent term and a universal noise term (Eq.( 25 ) in main text) \displaystyle K_{J}(t)=e^{-Jt}K_{J=0}(t)+\frac{1}{D^{2}}\left(1-e^{-Jt}\right) (5) where K_{J=0}(t) is the SFF without noise. It is exponentially suppressed by the noise, indicating that the SFF is not an effective diagnostic for quantum chaos when the noise strength is sufficiently large. Meanwhile, the noise-averaged two-point function exhibits similar behavior (Eq.( 28 ) in main text) \displaystyle C_{J}(t)=e^{-Jt}C_{J=0}(t)+{1\over D^{2}}\left(1-e^{-Jt}\right)\text{Tr}O\text{Tr}O^{\dagger}\ . (6) Thus, one can observe that noise induces oscillations in the Lanczos coefficients b_{n} . When J is sufficiently large, the linear growth of b_{n} in the original system is no longer apparent. As for the two-replica dynamics, we find that \displaystyle\left(\mathcal{U}_{2}(t)\right)_{ijkl;i^{\prime}j^{\prime}k^{\prime}l^{\prime}}=e^{-\text{i}E_{ijkl}t}\sum_{i=1}^{8}f_{a}(t)\mathsf{F}_{a} (7) where the factor e^{-\text{i}E_{ijkl}t} encodes information about the original system, while the noise effects are represented by the eight coefficients f_{a}(t) and the corresponding graphs \mathsf{F}_{a} . With the explicit expression of \mathcal{U}_{2} , one can study OTOCs and other two-replica observables. For two traceless Hermitian operators A and B , we find a simple formula for the disorder-averaged OTOC at infinite temperature (Eq.( 94 ) in main text) \displaystyle\text{OTOC}_{J}=e^{-2Jt}\text{OTOC}_{J=0}+O(1/D)\ . (8) Comparing Eq. ( 5 ), Eq. ( 6 ), and Eq. ( 8 ), one finds that the noise introduces an exponential suppression factor, which can obscure the features of the original system for large Jt . This may represent a universal effect of