Source-linked AI summary
Difference equations of average entropies
Youyi Huang, Linfeng Wei, Lu Wei, Peter J. Forrester
TL;DR
The paper addresses the case-by-case and increasingly laborious random-matrix derivation of entropy statistics for random state ensembles. It embeds entropic quantities in Toda tau functions to derive linear difference equations, whose solutions recover known average entropy formulas and motivate a broader cumulant framework. The higher-order cumulant extension remains conjectural and requires more intricate tau functions.
Problem
Random matrix methods require ensemble-dependent cumulant calculations, with substantially greater effort for successive cumulants.
Method
The paper differentiates deformed Toda tau functions to obtain linear difference equations for spectral-moment and entropy averages, then uses moment conversions for constrained ensembles.
Results
The solved difference equations recover known formulas for average von Neumann entropy and average purity across the studied ensembles.
Takeaways & Limitations
The framework provides a unified alternative to ensemble-specific derivations and suggests more efficient higher-order cumulant calculations through integrable hierarchies.
Takeaways & Limitations
The higher-order cumulant conjecture requires a more intricate tau function that relates cumulants of different orders while avoiding increasingly many initial conditions.
Abstract
from arXiv · showhide
Exact cumulants of entanglement entropies of random state ensembles have traditionally been studied within the random matrix framework. In this work, we propose an alternative approach based on the intrinsic connection to integrable systems. The central idea is to embed entropic quantities into tau functions satisfying Toda-type lattice equations, which in turn yield linear difference equations for their averages. Directly solving the difference equations recovers exact entropy formulas in the literature. The integrable systems approach bypasses the case-by-case, ensemble-dependent derivations required by random matrix methods. The approach also suggests a possible route towards unified and more efficient higher-order cumulant calculations by exploring integrable hierarchies.
1 Introduction
The paper replaces case-by-case random-matrix entropy calculations with an integrable-systems framework based on Toda tau functions and linear difference equations. It applies this approach across several random-state ensembles and extends the analysis to mixed matrix-dimension and power-parameter equations.
- Motivation: Random matrix methods require ensemble-specific cumulant calculations and increasingly greater effort for higher-order cumulants.The paper identifies Laguerre-polynomial integrals, two-matrix models, and moment-matrix factorization as examples of this dependence.
- Integrable-systems approach: Differentiating tau-function parameters transforms nonlinear lattice equations into linear difference equations in matrix dimension, whose elementary solutions recover known average entropy formulas.The recovered formulas include average von Neumann entropy and average purity.
- Integrable-systems approach: The proposed framework embeds unconstrained ensemble densities into Toda tau functions whose deformations generate entropic quantities through difference equations.Differentiating the relevant Toda systems produces equations for averages of spectral moments and entropies.
- Entropy construction: The method computes induced von Neumann entropy by differentiating with respect to the power parameter at s = 1, while s = 2 yields induced purity.Moment conversions then produce the desired average entropy formulas for constrained ensembles.
- Ensemble scope: The approach is adapted to Hilbert–Schmidt, Bures–Hall, and Muttalib–Borodin ensembles, with the latter interpolating between Hilbert–Schmidt and Bogoliubov–Kubo–Mori cases.The paper also treats the constrained Bogoliubov–Kubo–Mori ensemble through an additional deformation of the Toda equation.
- Extensions: The paper establishes mixed difference equations in matrix dimensions and power parameters and discusses integrable hierarchies as a route toward higher-order cumulants.Its main spectral-moment difference equations are formulated in matrix dimensions, unlike earlier results focused mainly on power parameters.
2 New difference equations of spectral moments
The paper derives matrix-dimension difference equations for spectral moments and entropies across several random matrix ensembles using integrable tau-function structures. Solving these equations recovers known average entropy and purity formulas, while extending results to real moment parameters and mixed ensemble limits.
- Main results: Difference equations for spectral moments are derived in matrix dimensions for Hilbert–Schmidt, Bures–Hall, and Muttalib–Borodin ensembles.The results are presented as propositions and are identified as potentially useful beyond entropy calculations.
- Integrable construction: Tau functions satisfy Toda, Pfaff Toda, or deformed lattice structures whose differentiated equations produce linear recurrences for average spectral moments.The Muttalib–Borodin case requires an additional u-flow to account for its θ-deformation.
- Entropy observables: The spectral-moment parameter s is real, extending beyond existing results restricted to integer s values.Average purity follows from s = 2, while average von Neumann entropy follows by differentiating with respect to s at s = 1.
- Recovered formulas: Solving the linear recurrences recovers known average von Neumann entropy and purity formulas for the Hilbert–Schmidt and Bures–Hall ensembles.The Hilbert–Schmidt derivation uses moment conversion, while the Bures–Hall results similarly reproduce established formulas.
- Relation to prior derivations: The Hilbert–Schmidt recurrence can also be obtained from a terminating hypergeometric representation and the difference equation for Hahn polynomials.This provides a related non-integrable derivation of the same difference equation.
- Deformed and limiting ensembles: The Muttalib–Borodin recurrence is expressed through covariance and yields the known average entropy formula; the Bogoliubov–Kubo–Mori case follows through the θ → 0 limit.For the Bogoliubov–Kubo–Mori ensemble, the limiting procedure is essential because no direct derivation is available.
3 Relations to known spectral moment recurrences
The paper connects matrix-dimension and power-parameter recurrences for spectral moments through mixed difference equations, using Toda hierarchies and Virasoro constraints. These relations recover known recurrences for both Hilbert–Schmidt and Bures–Hall ensembles.
- Mixed recurrences: Mixed difference equations connect spectral-moment recurrences in matrix dimension m with established recurrences in power parameter s.The connection is derived for the Hilbert–Schmidt and Bures–Hall ensembles.
- Hilbert–Schmidt ensemble: Combining the mixed recurrence with the matrix-dimension recurrence reproduces the known recurrence in s for the Hilbert–Schmidt ensemble.The derivation proceeds by shifting m and s and eliminating neighboring-dimension moments.
- Hilbert–Schmidt ensemble: Virasoro constraints reveal relations among spectral moments with different power parameters s.For the Hilbert–Schmidt ensemble, s is allowed to be real-valued.
- Hilbert–Schmidt ensemble: Higher Toda flows eliminate covariances by supplying additional independent relations among spectral moments and cumulants.For the Hilbert–Schmidt case, a second Toda flow is introduced for this purpose.
- Bures–Hall ensemble: The Bures–Hall ensemble yields an analogous mixed difference equation through Pfaffian representations, skew-moment evolution, and the B-Toda hierarchy.The resulting relation also recovers the known recurrence in s.
4 Discussion
The discussion extends the integrable-systems framework from average entropies toward higher-order cumulants and spectral-moment correlators. It proposes recursive and hierarchical constructions while identifying unresolved requirements for arbitrary cumulant orders.
- Higher-order cumulants: Higher-order cumulants are a natural next target beyond average entropies, with the discussion focusing first on von Neumann entropy in the Hilbert–Schmidt ensemble.The first six cumulants in this setting were previously obtained by random matrix methods.
- Higher-order cumulants: Differentiating a multi-parameter tau function preserves Toda structures and suggests reducing each cumulant order to a second-order difference equation in matrix dimension m.The proposed construction generates joint cumulants for real-valued power parameters.
- Higher-order cumulants: For any l > 2, Conjecture 1 proposes that the l-th cumulant of the induced entropy satisfies an l-th order difference equation.This statement is presented as a conjecture rather than an established result.
- Open requirements: The conjectured polynomial structure requires a more intricate tau function that avoids increasing initial-condition calculations and relates cumulants of different orders.The paper identifies these requirements as unresolved aspects of the proposed higher-order program.
- Further applications: Higher-order cumulant methods may extend to spectral-moment correlators with real-valued power parameters through recurrence relations across power parameters and matrix dimensions.The focus is on recurrence relations rather than summation representations.
Appendix A Trace-scaling identity of joint cumulants
Appendix A presents a cumulant identity for homogeneous random-variable densities. The identity gives a scaling rule for joint cumulants involving traces and spectral moments, supporting simplifications used in the main text.
- Lemma 1: Lemma 1 applies to joint densities whose factor h_m is homogeneous of degree d.The lemma covers the densities used for the Hilbert–Schmidt, Bures–Hall, and other models listed in the paper.
- Role in the paper: Lemma 1 is used to simplify joint cumulants arising from Toda lattice equations and Virasoro constraints.The appendix states that the identity is useful throughout the main text.
- Scaling identity: The lemma scales a joint cumulant involving n traces and l spectral moments with real-valued parameters s_1, . . . , s_l.The displayed scaling relation is the appendix's main identity.
- Proof strategy: The proof constructs a tau function generating the joint cumulant and uses the homogeneity condition to derive the scaling relation.Differentiation with respect to the source parameters is performed before setting them to zero.