Quantum Chaos Diagnostics

• Quantum Chaos Diagnostics are a suite of analytical, numerical, and experimental probes that distinguish chaotic, integrable, and intermediate dynamics in many-body systems.
• They integrate methods like spectral statistics, spectral form factors, and Krylov complexity—using both Hermitian and non-Hermitian frameworks—to reveal operator growth and state spreading.
• The combination of bi-Lanczos recursions and modern quantum simulation platforms offers practical insights for diagnosing chaos in both closed and open quantum systems.

Quantum chaos diagnostics are a suite of analytical, numerical, and experimental probes that distinguish between chaotic, integrable, and intermediate dynamical regimes in quantum many-body systems. Modern approaches integrate random-matrix theory, operator growth measures, entanglement diagnostics, and advanced statistical characterizations of non-Hermitian spectra. The rapid development of quantum simulation platforms and open quantum system techniques has driven the evolution of chaos diagnostics from Hermitian, closed systems to non-Hermitian and dissipative contexts. This article systematically reviews the conceptual bases, key methodologies, and operational regimes of quantum chaos diagnostics, with particular emphasis on Krylov complexity and bi-Lanczos approaches that are necessary in open and non-Hermitian quantum systems.

1. Classical and Quantum Chaos: Baseline Diagnostics

Quantum chaos is rooted in the analogy between exponential instability in classical trajectories and specific quantum features. Classic diagnostics include:

• Spectral statistics: Level spacing distributions probe spectral correlations. Integrable systems exhibit Poisson statistics, $P(s)=e^{-s}$, with no level repulsion. Chaotic systems manifest Wigner–Dyson statistics, $P(s)\sim s^\beta e^{-c s^2}$, with repulsion exponent $\beta=$ 1,2,4 (GOE, GUE, GSE). Random matrix theory (RMT) thus plays a central conceptual and computational role (Baggioli et al., 19 Aug 2025).
• Spectral form factor (SFF): $K(\tau)=|\mathrm{Tr}\,e^{-iH\tau}|^2$ provides a time-domain probe featuring “slope–dip–ramp–plateau” structure, where the presence and scaling of the correlation hole and ramp distinguish chaos from integrability (Das et al., 2024).
• OTOC (out-of-time-ordered correlators): Quantify the operator growth/scrambling via $C(t) = -\langle[W(t), V]^2\rangle$. Chaotic systems show early-time exponential growth, $C(t)\sim e^{\lambda_L t}$, with Lyapunov exponent $\lambda_L$ (Alonso et al., 2022, Kudler-Flam et al., 2019, Baggioli et al., 19 Aug 2025).

These diagnostics have been thoroughly validated in both model systems (e.g., quantum kicked top, kicked rotor, SYK model) and natural many-body contexts (spin chains, cold atoms, superconducting qubits).

2. Krylov Complexity and Bi-Lanczos Diagnostics

Krylov complexity provides a real-time, physically intuitive measure of the “distance” traversed by an evolving quantum state or operator in the dynamically constructed Krylov basis.

• Hermitian Krylov complexity: For a Hermitian $H$, the Lanczos algorithm constructs an orthonormal basis $\{|q_n\rangle\}$ via three-term recursions:

$$|r_{n+1}\rangle = (H-a_n)|q_n\rangle - b_n|q_{n-1}\rangle, \qquad \text{with}~a_n,~b_n~\text{from overlaps}.$$

The time-evolved state expands as $P(s)\sim s^\beta e^{-c s^2}$0 and complexity is defined by

$P(s)\sim s^\beta e^{-c s^2}$1

Chaotic systems exhibit near-linear initial growth, a pronounced peak, and eventual saturation, whereas integrable models display oscillatory or sublinear behavior (Baggioli et al., 19 Aug 2025).

• Non-Hermitian bi-Lanczos complexity: In open systems governed by a non-Hermitian effective generator ($P(s)\sim s^\beta e^{-c s^2}$2 or Liouvillian $P(s)\sim s^\beta e^{-c s^2}$3), standard orthogonality fails. The solution is a bi-orthogonal bi-Lanczos construction, producing dual right/left Krylov bases $P(s)\sim s^\beta e^{-c s^2}$4 with $P(s)\sim s^\beta e^{-c s^2}$5. The recursive structure is (for $P(s)\sim s^\beta e^{-c s^2}$6):

$P(s)\sim s^\beta e^{-c s^2}$7

and analogously for left vectors, followed by a spectral normalization step (Baggioli et al., 19 Aug 2025, Zhou et al., 27 Jan 2025). The non-Hermitian Krylov complexity reads

$P(s)\sim s^\beta e^{-c s^2}$8

The dynamics of $P(s)\sim s^\beta e^{-c s^2}$9 serve as a robust diagnostic: rapid growth and saturation signal chaos, while suppressed and early-saturating $\beta=$0 indicates integrability or strong dissipation.

• Physical insight: In the bi-Lanczos scheme, the effective tridiagonal matrix (with entries $\beta=$1, $\beta=$2, $\beta=$3) admits interpretation as a non-reciprocal tight-binding model in Krylov space. Reciprocity breaking, tracked via $\beta=$4, can indicate transitions between chaos and non-chaos as disorder or dissipation is increased (Zhou et al., 27 Jan 2025).
• Advantages: Unlike SVD-based approaches, the bi-Lanczos method naturally incorporates left/right dynamic asymmetry and is numerically stabilized by Gram–Schmidt bi-orthogonalization. It produces clear, interpretable indicators for both open and closed systems (Baggioli et al., 19 Aug 2025).

3. Comparative Structure and Universality of Chaos Diagnostics

Several distinct diagnostics are operationally and statistically linked:

Diagnostic	Mechanism	Chaotic Signature
Level spacings	RMT spectral statistics	Wigner–Dyson/Ginibre universality, level repulsion
Spectral form factor	Two-point correlations	Slope–dip–ramp–plateau (“correlation hole”)
OTOC	Operator growth	Early-time exponential (Lyapunov) and saturation
Krylov complexity	Operator/state spreading	Rapid (linear) growth, pronounced peak, saturation
Complex spacing ratios	Ginibre vs Poisson class	Anisotropy/isotropy, cubic vs. linear repulsion

• In Hermitian models, the spectral SFF, OTOC growth, and Krylov complexity all reproducibly diagnose quantum chaos, with extensive agreement between RMT predictions and numerics (Das et al., 2024, Alonso et al., 2022).
• In non-Hermitian and open systems, additional measures such as the complex spectral gap, complex spacing distributions, and complex spacing ratios (CSR) [$\beta=$5] reinforce the identification of chaotic, integrable, and localized regimes (Baggioli et al., 19 Aug 2025, Zhou et al., 27 Jan 2025).
• In specific cases (e.g., non-Hermitian SYK or random matrix ensembles), the alignment of $\beta=$6 peak, Ginibre-unitary ensemble spectral statistics, and CSR anisotropy constitutes a universal diagnostic triad (Baggioli et al., 19 Aug 2025).

4. Case Studies: Open System Chaos and Phase Diagrams

Recent developments allow for precise phase identification in open quantum systems:

• Non-Hermitian SYK and random matrices: For $\beta=$7 SYK, Krylov complexity peaks, GinUE $\beta=$8, and anisotropic CSR jointly confirm chaotic dynamics; for $\beta=$9 (integrable), these signatures disappear (Baggioli et al., 19 Aug 2025).
• Disordered non-Hermitian spin chains: The transition from chaos to non-chaos is diagnosed via the suppression of the mid-time linear ramp in $K(\tau)=|\mathrm{Tr}\,e^{-iH\tau}|^2$0 and a “reciprocity-breaking” parameter [$K(\tau)=|\mathrm{Tr}\,e^{-iH\tau}|^2$1 flips from $K(\tau)=|\mathrm{Tr}\,e^{-iH\tau}|^2$2 (chaos) to $K(\tau)=|\mathrm{Tr}\,e^{-iH\tau}|^2$3 (non-chaos)] as disorder is increased. A second, distinct “Krylov-localization” transition at lower disorder reflects weak ergodicity breaking (Zhou et al., 27 Jan 2025).
• Complex spacing ratios: In PT-symmetric and general non-Hermitian systems, the universal values of $K(\tau)=|\mathrm{Tr}\,e^{-iH\tau}|^2$4 (e.g., 0.74 for GinOE) precisely delineate integrable, chaotic, and PT-broken chaotic phases (Sharma et al., 2024).

These methodologies are robust to finite-size effects and numerically tractable for a broad range of system sizes and disorder strengths.

5. Experimental Platforms and Implementation Regimes

Quantum chaos diagnostics have practical import for experimental platforms including:

• Quantum simulators: Engineered dissipation in cold atoms, photonic lattices, and superconducting qubits allows implementation of non-Hermitian Hamiltonians and Liouvillian dynamics, enabling direct access to bi-Lanczos diagnostics (Baggioli et al., 19 Aug 2025).
• Complexity growth and operator spreading: The advantages of Krylov-based measures include their sensitivity to both left/right dynamics and compatibility with experimental state preparation and measurement, especially in open-system contexts where singular value decomposition and unitary OTOCs are not directly applicable.
• Noise robustness: White-noise and decoherence suppress traditional chaos signatures (SFF, OTOC, Krylov growth) exponentially in the noise rate. Modified diagnostics—such as SFF plateaus in Lindbladian spectra or renormalized bi-Lanczos metrics—are required to reliably differentiate chaos in realistic open systems (Li, 3 Mar 2025).
• Other proposed extensions: Further generalizations include applications to non-Hermitian symmetry classes (AI$K(\tau)=|\mathrm{Tr}\,e^{-iH\tau}|^2$5, AII$K(\tau)=|\mathrm{Tr}\,e^{-iH\tau}|^2$6), large system-size scaling, and the construction of Krylov-space entropies and metrics as information-theoretical probes of complexity and ergodicity (Baggioli et al., 19 Aug 2025).

6. Outlook and Open Directions

Open questions in the field involve deeper connections between operator growth, information scrambling, entanglement diagnostics, and non-Hermitian spectral theory:

• Universality and scaling: The observed universality of bi-Lanczos Krylov complexity as a marker of chaos, supported by both numerical and random-matrix analytic results, prompts investigation into its limits across other symmetry classes and models of engineered dissipation (Baggioli et al., 19 Aug 2025).
• Operator-space measures: Extensions to higher Krylov-space moments (e.g., entropy, participation ratio), operator entanglement, and entropic diagnostics are expected to yield finer distinctions among chaotic, weakly ergodic, and many-body localized phases—including systems on the ergodic–scarring border (Giasemis, 20 Mar 2026).
• Experimental feasibility: Advances in quantum hardware bolster the prospects for direct measurement of Krylov complexity, correlation holes, and related metrics, especially as sophisticated state preparation and readout schemes become standard.
• Theoretical integration: Bi-Lanczos Krylov diagnostics provide a geometric and algebraic framework unifying older measures (SFF, OTOC, level statistics) with new operator-based diagnostics, promoting a holistic understanding of quantum dynamical complexity, ergodicity breaking, and universality in open quantum systems.

In sum, Krylov complexity and bi-Lanczos recursions stand at the forefront of quantum chaos diagnostics, offering a mathematically rigorous, physically interpretable, and operationally robust approach for both closed and open quantum systems (Baggioli et al., 19 Aug 2025, Zhou et al., 27 Jan 2025). Their universality, sensitivity, and experimental compatibility make them essential tools in the continuing effort to chart the landscape of quantum chaotic dynamics.