Symmetry-driven thermalization via finite de Finetti theorems

Nicolas J. Cerf; Uttam Singh

arxiv: 2604.09410 · v1 · submitted 2026-04-10 · 🪐 quant-ph · math-ph· math.MP

Symmetry-driven thermalization via finite de Finetti theorems

Uttam Singh , Nicolas J. Cerf This is my paper

Pith reviewed 2026-05-10 17:00 UTC · model grok-4.3

classification 🪐 quant-ph math-phmath.MP

keywords quantum thermalizationde Finetti theoremenergy-preserving unitariessymmetry in quantum mechanicsLindblad dynamicsfinite-size quantum systemsreduced density operators

0 comments

The pith

Invariance under energy-preserving unitaries forces the reduced states of large quantum systems to approach thermal mixtures without statistical assumptions.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that a pure symmetry condition on a closed quantum system is enough to produce thermal behavior in its subsystems. Any state of many qudits that remains exactly the same under all unitaries conserving total energy has single-qudit marginals that sit close to convex combinations of thermal product states. Explicit bounds on the trace distance and relative entropy show the approximation improves as the total number of qudits grows. The authors also exhibit a concrete Lindblad evolution that drives any initial state toward this symmetry class, so the thermal appearance arises dynamically from the symmetry alone.

Core claim

A finite de Finetti-type theorem holds for quantum states invariant under the full group of energy-preserving unitaries: the reduced marginals of any such N-qudit state are close in trace distance and relative entropy to convex mixtures of thermal product states, with error bounds that vanish as N tends to infinity. An energy-conserving Lindblad dynamics is constructed whose long-time limit belongs to the same invariant class.

What carries the argument

Finite de Finetti theorem for states invariant under energy-preserving unitaries, which supplies explicit bounds turning symmetry into thermal marginals.

If this is right

Thermal appearance of subsystems can arise deterministically from symmetry alone, without chaos or ergodicity.
The approach supplies quantitative error bounds that remain valid for finite but large systems.
A concrete Lindblad dynamics realizes the required symmetry class at long times.
The same symmetry principle can be checked in any system whose Hamiltonian commutes with the total energy operator.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The result may extend to other conserved quantities besides energy, such as particle number or angular momentum.
It offers a possible explanation for thermal-like behavior in integrable systems that lack classical chaos.
Small-scale quantum simulators could test the finite-N bounds by preparing approximately symmetric states and measuring marginals.

Load-bearing premise

The full N-qudit state must be exactly invariant under every energy-preserving unitary, not merely approximately or under a subgroup.

What would settle it

An explicit N-qudit state that is invariant under all energy-preserving unitaries yet whose single-qudit reduced density operator lies a fixed distance away from every possible mixture of thermal product states, even as N grows large.

Figures

Figures reproduced from arXiv: 2604.09410 by Nicolas J. Cerf, Uttam Singh.

**Figure 2.** Figure 2: Convergence of reduced marginals to thermal mixtures. The single-qutrit Hamiltonian is taken to be [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗

read the original abstract

Thermal behavior in subsystems of closed quantum systems is commonly attributed to dynamical chaos, quantum ergodicity, canonical typicality, or the eigenstate thermalization hypothesis, suggesting a fundamentally statistical origin of thermalization. Here, we propose a potential alternative mechanism in which thermal structures emerge deterministically from symmetry considerations alone, without recourse to statistical arguments. We prove a finite de Finetti-type theorem for quantum states invariant under energy-preserving unitaries, establishing that the reduced marginals of any such invariant $N$-qudit state are close (both in trace distance and relative entropy) to convex mixtures of thermal product states, with explicit error bounds vanishing as $N \to \infty$. We further present an example of energy-conserving Lindblad dynamics whose long-time limit is invariant under energy-preserving unitaries, providing a dynamical realization of the desired symmetry class. These results imply that invariance under energy-preserving unitaries suffices as a sole fundamental, deterministic principle to enforce thermal structures.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Exact invariance under energy-preserving unitaries forces reduced states close to thermal mixtures via a finite de Finetti theorem, but the Lindblad dynamics only hits exact invariance at infinite time.

read the letter

The core result is that any state exactly invariant under the full group of energy-preserving unitaries on N qudits has single-qudit reductions within explicit trace-distance and relative-entropy bounds of convex mixtures of product thermal states, with the distance vanishing as N grows. They also give a Lindblad master equation whose long-time limit sits in that invariant set, so the symmetry can be reached dynamically while conserving energy. That combination is what stands out: a deterministic symmetry principle replacing statistical assumptions, with concrete finite-N estimates rather than asymptotic hand-waving. The de Finetti extension to this symmetry class looks like the genuinely new piece, and the bounds are the part that could actually be used for finite-size checks. The example is straightforward and shows the symmetry class is dynamically accessible. The weak link is the gap between the theorem and the dynamics. The theorem requires exact invariance under the entire group; the Lindblad flow only reaches that at infinite time. At any finite t the state is only approximately invariant, so the total error is the sum of the dynamical distance to the invariant manifold plus the de Finetti distance. Without a rate on how fast the dynamics approaches invariance, it is not obvious that the combined error still goes to zero for large but finite N and physically relevant t. The assumption of exact full-group invariance is also strong; most real systems will have only approximate or subgroup symmetries. This is for people working on foundations of quantum thermodynamics who are looking for symmetry-based alternatives to ETH or typicality. The math is clean enough and the idea distinct enough that it deserves a serious referee, though the dynamical application will need tighter error control before the claim lands fully.

Referee Report

1 major / 0 minor

Summary. The paper claims that thermal structures can emerge deterministically from symmetry under energy-preserving unitaries alone, without statistical arguments. It proves a finite de Finetti-type theorem establishing that the reduced marginals of any N-qudit state exactly invariant under the full group of energy-preserving unitaries are close (in trace distance and relative entropy) to convex mixtures of thermal product states, with explicit bounds vanishing as N→∞. It further gives an energy-conserving Lindblad dynamics whose long-time limit lies in this invariant set, as a dynamical realization.

Significance. If the central theorem holds with the stated bounds, the work provides a concrete alternative deterministic mechanism for thermalization grounded in symmetry, which could be significant for quantum thermodynamics and foundations of statistical mechanics. The explicit finite-N error bounds (both trace distance and relative entropy) and the machine-checkable nature of the de Finetti derivation are strengths that allow falsifiable predictions. The dynamical example adds value by linking the symmetry class to open-system evolution, though its quantitative connection to the theorem remains to be tightened.

major comments (1)

[Lindblad dynamics example] The finite de Finetti theorem (as stated in the abstract and proved in the main text) applies only to states that are exactly invariant under the entire group of energy-preserving unitaries. The Lindblad example is constructed so that its long-time limit is exactly invariant, but any finite-time state is only approximately invariant. No quantitative bound is supplied on the rate at which the dynamical state approaches the invariant manifold, so it is unclear whether the total error (dynamical distance to invariance plus de Finetti distance) still vanishes as N→∞ for finite times and physically relevant system sizes.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We appreciate the referee's detailed review and the identification of this subtlety in connecting the dynamical example to the main theorem. We respond to the major comment as follows.

read point-by-point responses

Referee: [Lindblad dynamics example] The finite de Finetti theorem (as stated in the abstract and proved in the main text) applies only to states that are exactly invariant under the entire group of energy-preserving unitaries. The Lindblad example is constructed so that its long-time limit is exactly invariant, but any finite-time state is only approximately invariant. No quantitative bound is supplied on the rate at which the dynamical state approaches the invariant manifold, so it is unclear whether the total error (dynamical distance to invariance plus de Finetti distance) still vanishes as N→∞ for finite times and physically relevant system sizes.

Authors: We agree with the referee that the theorem applies strictly to exactly invariant states, whereas the Lindblad evolution only reaches exact invariance in the long-time limit. The manuscript presents the Lindblad dynamics as an illustrative example of an energy-conserving process whose attractor is the desired invariant set, thereby providing a dynamical means to realize the symmetry class considered in the theorem. However, we acknowledge that a quantitative analysis of the convergence rate is absent, which limits the ability to bound the total error for finite times. In the revised version, we will add a paragraph in the section discussing the Lindblad dynamics to explicitly state this limitation and clarify the scope of the example. Specifically, we will note that while the long-time limit satisfies the conditions of the theorem, applications to finite-time dynamics would require additional estimates on the distance to invariance, which may depend on the system size N. This revision will prevent potential misinterpretation of the dynamical example as providing finite-time thermalization bounds. revision: yes

Circularity Check

0 steps flagged

No circularity: theorem derives thermal closeness directly from invariance assumption via explicit bounds.

full rationale

The paper states and proves a finite de Finetti-type theorem establishing that reduced marginals of exactly invariant N-qudit states are close to convex mixtures of thermal product states, with error bounds vanishing as N→∞. This is a direct mathematical derivation from the group-invariance premise; the Lindblad dynamics example only shows that the long-time limit lies in the invariant set and is not used to derive or fit the theorem itself. No parameters are fitted and then relabeled as predictions, no self-citations are invoked as load-bearing uniqueness results, and the central claim does not reduce to a renaming or self-definition of its inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The proof rests on standard axioms of quantum mechanics and representation theory for the unitary group; no free parameters or invented entities are introduced in the abstract.

axioms (2)

standard math Quantum states are density operators on finite-dimensional Hilbert spaces.
Invoked implicitly for N-qudit states and reduced marginals.
domain assumption Energy-preserving unitaries form a group that leaves the total Hamiltonian invariant.
Central to the invariance class used in the theorem.

pith-pipeline@v0.9.0 · 5461 in / 1191 out tokens · 37575 ms · 2026-05-10T17:00:25.702813+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

43 extracted references · 43 canonical work pages

[1]

Choos- ingt≥Γ −1 log(1/ϵ) implies that etL(ρ0)− T EPU 1 ≤ ϵ

Then, we show in Appendix E that etL(ρ0)− T EPU 1 ≤e −tΓ ||∆ρ0||1 ≤2e −tΓ,(18) where Γ := min E{γE, λE}and||∆ρ 0||1 ≤2. Choos- ingt≥Γ −1 log(1/ϵ) implies that etL(ρ0)− T EPU 1 ≤ ϵ. This means that any state approaches its EPU- twirl under the above Lindbladian dynamics on a timescaleO Γ−1 log(1/ϵ) . Therefore, as a consequence of Theorem 1, its marginals ...

work page 2024
[2]

Kinoshita, T

T. Kinoshita, T. Wenger, and D. S. Weiss, A quantum Newton’s cradle, Nature440, 900 (2006)

work page 2006
[3]

Langen, S

T. Langen, S. Erne, R. Geiger, B. Rauer, T. Schweigler, M. Kuhnert, W. Rohringer, I. E. Mazets, T. Gasenzer, and J. Schmiedmayer, Experimental observation of a gen- eralized Gibbs ensemble, Science348, 207 (2015)

work page 2015
[4]

Langen, R

T. Langen, R. Geiger, and J. Schmiedmayer, Ultracold atoms out of equilibrium, Annual Review of Condensed Matter Physics6, 201 (2015)

work page 2015
[5]

A. M. Kaufman, M. E. Tai, A. Lukin, M. Rispoli, R. Schittko, P. M. Preiss, and M. Greiner, Quantum ther- malization through entanglement in an isolated many- body system, Science353, 794 (2016)

work page 2016
[6]

E. T. Jaynes, Information theory and statistical mechan- ics, Phys. Rev.106, 620 (1957)

work page 1957
[7]

E. T. Jaynes, Information theory and statistical mechan- ics. ii, Phys. Rev.108, 171 (1957)

work page 1957
[8]

Pusz and S

W. Pusz and S. L. Woronowicz, Passive states and KMS states for general quantum systems, Communications in Mathematical Physics58, 273 (1978)

work page 1978
[9]

Lenard, Thermodynamical proof of the Gibbs formula for elementary quantum systems, Journal of Statistical Physics19, 575 (1978)

A. Lenard, Thermodynamical proof of the Gibbs formula for elementary quantum systems, Journal of Statistical Physics19, 575 (1978)

work page 1978
[10]

Popescu, A

S. Popescu, A. J. Short, and A. Winter, Entanglement and the foundations of statistical mechanics, Nature Physics2, 754 (2006)

work page 2006
[11]

Goldstein, J

S. Goldstein, J. L. Lebowitz, R. Tumulka, and N. Zangh` ı, Canonical typicality, Phys. Rev. Lett.96, 050403 (2006)

work page 2006
[12]

J. M. Deutsch, Quantum statistical mechanics in a closed system, Phys. Rev. A43, 2046 (1991)

work page 2046
[13]

Srednicki, Chaos and quantum thermalization, Phys

M. Srednicki, Chaos and quantum thermalization, Phys. Rev. E50, 888 (1994)

work page 1994
[14]

Rigol, V

M. Rigol, V. Dunjko, and M. Olshanii, Thermalization and its mechanism for generic isolated quantum systems, Nature452, 854 (2008)

work page 2008
[15]

Reimann, Foundation of statistical mechanics under experimentally realistic conditions, Phys

P. Reimann, Foundation of statistical mechanics under experimentally realistic conditions, Phys. Rev. Lett.101, 190403 (2008)

work page 2008
[16]

Rigol, Breakdown of thermalization in finite one- dimensional systems, Phys

M. Rigol, Breakdown of thermalization in finite one- dimensional systems, Phys. Rev. Lett.103, 100403 (2009)

work page 2009
[17]

Linden, S

N. Linden, S. Popescu, A. J. Short, and A. Winter, Quan- tum mechanical evolution towards thermal equilibrium, Phys. Rev. E79, 061103 (2009)

work page 2009
[18]

A. J. Short and T. C. Farrelly, Quantum equilibration in finite time, New Journal of Physics14, 013063 (2012)

work page 2012
[19]

D’Alessio, Y

L. D’Alessio, Y. Kafri, A. Polkovnikov, and M. Rigol, From quantum chaos and eigenstate thermalization to statistical mechanics and thermodynamics, Advances in Physics65, 239 (2016)

work page 2016
[20]

D. A. Abanin, E. Altman, I. Bloch, and M. Serbyn, Col- loquium: Many-body localization, thermalization, and entanglement, Rev. Mod. Phys.91, 021001 (2019)

work page 2019
[21]

T. Mori, T. N. Ikeda, E. Kaminishi, and M. Ueda, Thermalization and prethermalization in isolated quan- tum systems: a theoretical overview, Journal of Physics B: Atomic, Molecular and Optical Physics51, 112001 (2018)

work page 2018
[22]

Leverrier and N

A. Leverrier and N. J. Cerf, Quantum de Finetti theorem in phase-space representation, Phys. Rev. A80, 010102 (2009)

work page 2009
[23]

T. M. Cover and J. A. Thomas,Elements of Information Theory, 2nd ed. (Wiley-Interscience, 2006)

work page 2006
[24]

Csisz´ ar and J

I. Csisz´ ar and J. K¨ orner,Information Theory: Cod- ing Theorems for Discrete Memoryless Systems, 2nd ed. (Cambridge University Press, 2011)

work page 2011
[25]

Diaconis and D

P. Diaconis and D. Freedman, Finite exchangeable se- quences, The Annals of Probability8, 745 (1980)

work page 1980
[26]

The complexity of thermalization in finite quantum systems

D. Devulapalli, T. C. Mooney, and J. D. Watson, The complexity of thermalization in finite quantum systems 6 (2025), arXiv:2507.00405 [quant-ph]

work page arXiv 2025
[27]

M. L. LaBorde, S. Rethinasamy, and M. M. Wilde, Test- ing symmetry on quantum computers, Quantum7, 1120 (2023)

work page 2023
[28]

Rethinasamy, M

S. Rethinasamy, M. L. LaBorde, and M. M. Wilde, Quan- tum computational complexity and symmetry, Canadian Journal of Physics103, 215 (2025)

work page 2025
[29]

Robbins, A remark on Stirling’s formula, The Ameri- can Mathematical Monthly62, 26 (1955)

H. Robbins, A remark on Stirling’s formula, The Ameri- can Mathematical Monthly62, 26 (1955)

work page 1955
[30]

T. M. Lapi´ nski, Approximations of the sum of states by Laplace’s method for a system of particles with a finite number of energy levels and application to limit theo- rems, Mathematical Physics, Analysis and Geometry23, 9 (2020)

work page 2020
[31]

Johnson, L

O. Johnson, L. Gavalakis, and I. Kontoyiannis, Relative entropy bounds for sampling with and without replace- ment (2024), arXiv:2404.06632 [math.PR]

work page arXiv 2024
[32]

A. J. Stam, Distance between sampling with and without replacement, Statistica Neerlandica32, 81 (1978)

work page 1978
[33]

Borderi,De Finetti methods in quantum information, Ph.D

F. Borderi,De Finetti methods in quantum information, Ph.D. thesis, Imperial College London (2022)

work page 2022
[34]

Gavalakis, O

L. Gavalakis, O. Johnson, and I. Kontoyiannis, Fi- nite de Finetti bounds in relative entropy (2024), arXiv:2407.12921 [math.PR]

work page arXiv 2024
[35]

Marvian and R

I. Marvian and R. W. Spekkens, Extending Noether’s theorem by quantifying the asymmetry of quantum states, Nature Communications5, 3821 (2014)

work page 2014
[36]

Lostaglio, D

M. Lostaglio, D. Jennings, and T. Rudolph, Description of quantum coherence in thermodynamic processes re- quires constraints beyond free energy, Nature Communi- cations6, 6383 (2015)

work page 2015
[37]

Gour and R

G. Gour and R. W. Spekkens, The resource theory of quantum reference frames: manipulations and mono- tones, New Journal of Physics10, 033023 (2008)

work page 2008
[38]

T. M. Cover and J. A. Thomas,Elements of Information Theory, 2nd ed. (Wiley-Interscience, Hoboken, NJ, USA, 2006). The Appendix is organized as follows. In Appendix A, we introduce a simple model ofNqutrits and numer- ically demonstrate the convergence of marginals of EPU–invariant states toward thermal states. In Appendix B, we prove that the set of EPU-...

work page 2006
[39]

(C24) and (C25), we have ProjK =MatK=I−K −1C T (CK −1C T )−1C.(C27) Thus,K −1C T (CK −1C T )−1C=I−Proj K is a projector and 111T K −1C T (CK −1C T )−1C111≥0

Thus 111T Matv= 111T111−111T K −1C T (CK −1C T )−1C111.(C26) From Eqs. (C24) and (C25), we have ProjK =MatK=I−K −1C T (CK −1C T )−1C.(C27) Thus,K −1C T (CK −1C T )−1C=I−Proj K is a projector and 111T K −1C T (CK −1C T )−1C111≥0. Then using the fact that 111T111 =d, we have 111T Matv≤d. Thus d−1X i,j=0 Pβ(j)−1bT i eK −1bj ≤d. Appendix D: de Finetti theorem...

work page
[40]

We first comment on the time translation symmetry which is strictly weaker than the EPU invariance

Approximate symmetry Here we show how the asymmetry of a state, quantifying the departure of the state from being invariant under all energy-preserving unitaries, controls the closeness of a state to a mixture of thermal states via a de Finetti approximation. We first comment on the time translation symmetry which is strictly weaker than the EPU invarianc...

work page
[41]

tr( ˆPE′′ρ(t) ˆPE′′) g(E ′′) ˆPE′′ − ˆPE′′ ρ(t) ˆPE′′ # ˆPE + ˆPE X E′′ λE h ˆPE′′ρ(t) ˆPE′′ − 1 2 { ˆPE′′, ρ(t)} i ˆPE =γ

EPU-invariance from Lindblad dynamics We consider a physically motivated class of open-system dynamics that converges to anexactlyEPU-invariant state, thereby enabling a direct application of our de Finetti theorem. Let a system ofNqudits have Hamiltonian H= PN j=1 h(j), whereh (j) =Pd−1 x=0 Ex|x⟩⟨x|.{|x⟩} d−1 x=0 is the local energy eigenbasis and theE x...

work page
[42]

We show thatLis diagonalizable and construct an explicit orthonormal eigenbasis together with all spectral projectors

The generatorLacts on the operator spaceB(H), a vector space of dimension dimB(H) =D 2, whereD=d N. We show thatLis diagonalizable and construct an explicit orthonormal eigenbasis together with all spectral projectors. Block decomposition of operator space.Recall that ˆPE denotes the orthogonal projector onto the total-energy-E subspace and letg(E) = tr( ...

work page
[43]

Letλbe an eigenvalue ofL

+ P E̸=E ′ g(E)g(E ′) =D 2. Letλbe an eigenvalue ofL. Then etL = X λ eλtPλ,(E17) whereP λ is a projector onto eigenspaceλofL. Therefore, for any operatorX, etL(X) 1 = X λ eλtPλ(X) 1 ≤e tmax λ{λ} ||X||1 , where we used the fact that P λ Pλ =I sup−op. Further, max λ{λ}=−min E{γE, λE}. Using Γ = min E{γE, λE}, we have etL(X) 1 ≤e −tΓ ||X||1 . etL(ρ0)− T EPU ...

work page

[1] [1]

Choos- ingt≥Γ −1 log(1/ϵ) implies that etL(ρ0)− T EPU 1 ≤ ϵ

Then, we show in Appendix E that etL(ρ0)− T EPU 1 ≤e −tΓ ||∆ρ0||1 ≤2e −tΓ,(18) where Γ := min E{γE, λE}and||∆ρ 0||1 ≤2. Choos- ingt≥Γ −1 log(1/ϵ) implies that etL(ρ0)− T EPU 1 ≤ ϵ. This means that any state approaches its EPU- twirl under the above Lindbladian dynamics on a timescaleO Γ−1 log(1/ϵ) . Therefore, as a consequence of Theorem 1, its marginals ...

work page 2024

[2] [2]

Kinoshita, T

T. Kinoshita, T. Wenger, and D. S. Weiss, A quantum Newton’s cradle, Nature440, 900 (2006)

work page 2006

[3] [3]

Langen, S

T. Langen, S. Erne, R. Geiger, B. Rauer, T. Schweigler, M. Kuhnert, W. Rohringer, I. E. Mazets, T. Gasenzer, and J. Schmiedmayer, Experimental observation of a gen- eralized Gibbs ensemble, Science348, 207 (2015)

work page 2015

[4] [4]

Langen, R

T. Langen, R. Geiger, and J. Schmiedmayer, Ultracold atoms out of equilibrium, Annual Review of Condensed Matter Physics6, 201 (2015)

work page 2015

[5] [5]

A. M. Kaufman, M. E. Tai, A. Lukin, M. Rispoli, R. Schittko, P. M. Preiss, and M. Greiner, Quantum ther- malization through entanglement in an isolated many- body system, Science353, 794 (2016)

work page 2016

[6] [6]

E. T. Jaynes, Information theory and statistical mechan- ics, Phys. Rev.106, 620 (1957)

work page 1957

[7] [7]

E. T. Jaynes, Information theory and statistical mechan- ics. ii, Phys. Rev.108, 171 (1957)

work page 1957

[8] [8]

Pusz and S

W. Pusz and S. L. Woronowicz, Passive states and KMS states for general quantum systems, Communications in Mathematical Physics58, 273 (1978)

work page 1978

[9] [9]

Lenard, Thermodynamical proof of the Gibbs formula for elementary quantum systems, Journal of Statistical Physics19, 575 (1978)

A. Lenard, Thermodynamical proof of the Gibbs formula for elementary quantum systems, Journal of Statistical Physics19, 575 (1978)

work page 1978

[10] [10]

Popescu, A

S. Popescu, A. J. Short, and A. Winter, Entanglement and the foundations of statistical mechanics, Nature Physics2, 754 (2006)

work page 2006

[11] [11]

Goldstein, J

S. Goldstein, J. L. Lebowitz, R. Tumulka, and N. Zangh` ı, Canonical typicality, Phys. Rev. Lett.96, 050403 (2006)

work page 2006

[12] [12]

J. M. Deutsch, Quantum statistical mechanics in a closed system, Phys. Rev. A43, 2046 (1991)

work page 2046

[13] [13]

Srednicki, Chaos and quantum thermalization, Phys

M. Srednicki, Chaos and quantum thermalization, Phys. Rev. E50, 888 (1994)

work page 1994

[14] [14]

Rigol, V

M. Rigol, V. Dunjko, and M. Olshanii, Thermalization and its mechanism for generic isolated quantum systems, Nature452, 854 (2008)

work page 2008

[15] [15]

Reimann, Foundation of statistical mechanics under experimentally realistic conditions, Phys

P. Reimann, Foundation of statistical mechanics under experimentally realistic conditions, Phys. Rev. Lett.101, 190403 (2008)

work page 2008

[16] [16]

Rigol, Breakdown of thermalization in finite one- dimensional systems, Phys

M. Rigol, Breakdown of thermalization in finite one- dimensional systems, Phys. Rev. Lett.103, 100403 (2009)

work page 2009

[17] [17]

Linden, S

N. Linden, S. Popescu, A. J. Short, and A. Winter, Quan- tum mechanical evolution towards thermal equilibrium, Phys. Rev. E79, 061103 (2009)

work page 2009

[18] [18]

A. J. Short and T. C. Farrelly, Quantum equilibration in finite time, New Journal of Physics14, 013063 (2012)

work page 2012

[19] [19]

D’Alessio, Y

L. D’Alessio, Y. Kafri, A. Polkovnikov, and M. Rigol, From quantum chaos and eigenstate thermalization to statistical mechanics and thermodynamics, Advances in Physics65, 239 (2016)

work page 2016

[20] [20]

D. A. Abanin, E. Altman, I. Bloch, and M. Serbyn, Col- loquium: Many-body localization, thermalization, and entanglement, Rev. Mod. Phys.91, 021001 (2019)

work page 2019

[21] [21]

T. Mori, T. N. Ikeda, E. Kaminishi, and M. Ueda, Thermalization and prethermalization in isolated quan- tum systems: a theoretical overview, Journal of Physics B: Atomic, Molecular and Optical Physics51, 112001 (2018)

work page 2018

[22] [22]

Leverrier and N

A. Leverrier and N. J. Cerf, Quantum de Finetti theorem in phase-space representation, Phys. Rev. A80, 010102 (2009)

work page 2009

[23] [23]

T. M. Cover and J. A. Thomas,Elements of Information Theory, 2nd ed. (Wiley-Interscience, 2006)

work page 2006

[24] [24]

Csisz´ ar and J

I. Csisz´ ar and J. K¨ orner,Information Theory: Cod- ing Theorems for Discrete Memoryless Systems, 2nd ed. (Cambridge University Press, 2011)

work page 2011

[25] [25]

Diaconis and D

P. Diaconis and D. Freedman, Finite exchangeable se- quences, The Annals of Probability8, 745 (1980)

work page 1980

[26] [26]

The complexity of thermalization in finite quantum systems

D. Devulapalli, T. C. Mooney, and J. D. Watson, The complexity of thermalization in finite quantum systems 6 (2025), arXiv:2507.00405 [quant-ph]

work page arXiv 2025

[27] [27]

M. L. LaBorde, S. Rethinasamy, and M. M. Wilde, Test- ing symmetry on quantum computers, Quantum7, 1120 (2023)

work page 2023

[28] [28]

Rethinasamy, M

S. Rethinasamy, M. L. LaBorde, and M. M. Wilde, Quan- tum computational complexity and symmetry, Canadian Journal of Physics103, 215 (2025)

work page 2025

[29] [29]

Robbins, A remark on Stirling’s formula, The Ameri- can Mathematical Monthly62, 26 (1955)

H. Robbins, A remark on Stirling’s formula, The Ameri- can Mathematical Monthly62, 26 (1955)

work page 1955

[30] [30]

T. M. Lapi´ nski, Approximations of the sum of states by Laplace’s method for a system of particles with a finite number of energy levels and application to limit theo- rems, Mathematical Physics, Analysis and Geometry23, 9 (2020)

work page 2020

[31] [31]

Johnson, L

O. Johnson, L. Gavalakis, and I. Kontoyiannis, Relative entropy bounds for sampling with and without replace- ment (2024), arXiv:2404.06632 [math.PR]

work page arXiv 2024

[32] [32]

A. J. Stam, Distance between sampling with and without replacement, Statistica Neerlandica32, 81 (1978)

work page 1978

[33] [33]

Borderi,De Finetti methods in quantum information, Ph.D

F. Borderi,De Finetti methods in quantum information, Ph.D. thesis, Imperial College London (2022)

work page 2022

[34] [34]

Gavalakis, O

L. Gavalakis, O. Johnson, and I. Kontoyiannis, Fi- nite de Finetti bounds in relative entropy (2024), arXiv:2407.12921 [math.PR]

work page arXiv 2024

[35] [35]

Marvian and R

I. Marvian and R. W. Spekkens, Extending Noether’s theorem by quantifying the asymmetry of quantum states, Nature Communications5, 3821 (2014)

work page 2014

[36] [36]

Lostaglio, D

M. Lostaglio, D. Jennings, and T. Rudolph, Description of quantum coherence in thermodynamic processes re- quires constraints beyond free energy, Nature Communi- cations6, 6383 (2015)

work page 2015

[37] [37]

Gour and R

G. Gour and R. W. Spekkens, The resource theory of quantum reference frames: manipulations and mono- tones, New Journal of Physics10, 033023 (2008)

work page 2008

[38] [38]

T. M. Cover and J. A. Thomas,Elements of Information Theory, 2nd ed. (Wiley-Interscience, Hoboken, NJ, USA, 2006). The Appendix is organized as follows. In Appendix A, we introduce a simple model ofNqutrits and numer- ically demonstrate the convergence of marginals of EPU–invariant states toward thermal states. In Appendix B, we prove that the set of EPU-...

work page 2006

[39] [39]

(C24) and (C25), we have ProjK =MatK=I−K −1C T (CK −1C T )−1C.(C27) Thus,K −1C T (CK −1C T )−1C=I−Proj K is a projector and 111T K −1C T (CK −1C T )−1C111≥0

Thus 111T Matv= 111T111−111T K −1C T (CK −1C T )−1C111.(C26) From Eqs. (C24) and (C25), we have ProjK =MatK=I−K −1C T (CK −1C T )−1C.(C27) Thus,K −1C T (CK −1C T )−1C=I−Proj K is a projector and 111T K −1C T (CK −1C T )−1C111≥0. Then using the fact that 111T111 =d, we have 111T Matv≤d. Thus d−1X i,j=0 Pβ(j)−1bT i eK −1bj ≤d. Appendix D: de Finetti theorem...

work page

[40] [40]

We first comment on the time translation symmetry which is strictly weaker than the EPU invariance

Approximate symmetry Here we show how the asymmetry of a state, quantifying the departure of the state from being invariant under all energy-preserving unitaries, controls the closeness of a state to a mixture of thermal states via a de Finetti approximation. We first comment on the time translation symmetry which is strictly weaker than the EPU invarianc...

work page

[41] [41]

tr( ˆPE′′ρ(t) ˆPE′′) g(E ′′) ˆPE′′ − ˆPE′′ ρ(t) ˆPE′′ # ˆPE + ˆPE X E′′ λE h ˆPE′′ρ(t) ˆPE′′ − 1 2 { ˆPE′′, ρ(t)} i ˆPE =γ

EPU-invariance from Lindblad dynamics We consider a physically motivated class of open-system dynamics that converges to anexactlyEPU-invariant state, thereby enabling a direct application of our de Finetti theorem. Let a system ofNqudits have Hamiltonian H= PN j=1 h(j), whereh (j) =Pd−1 x=0 Ex|x⟩⟨x|.{|x⟩} d−1 x=0 is the local energy eigenbasis and theE x...

work page

[42] [42]

We show thatLis diagonalizable and construct an explicit orthonormal eigenbasis together with all spectral projectors

The generatorLacts on the operator spaceB(H), a vector space of dimension dimB(H) =D 2, whereD=d N. We show thatLis diagonalizable and construct an explicit orthonormal eigenbasis together with all spectral projectors. Block decomposition of operator space.Recall that ˆPE denotes the orthogonal projector onto the total-energy-E subspace and letg(E) = tr( ...

work page

[43] [43]

Letλbe an eigenvalue ofL

+ P E̸=E ′ g(E)g(E ′) =D 2. Letλbe an eigenvalue ofL. Then etL = X λ eλtPλ,(E17) whereP λ is a projector onto eigenspaceλofL. Therefore, for any operatorX, etL(X) 1 = X λ eλtPλ(X) 1 ≤e tmax λ{λ} ||X||1 , where we used the fact that P λ Pλ =I sup−op. Further, max λ{λ}=−min E{γE, λE}. Using Γ = min E{γE, λE}, we have etL(X) 1 ≤e −tΓ ||X||1 . etL(ρ0)− T EPU ...

work page