{"id":"e74a64d9-49e1-44d8-8fda-4af58c115d17","arxiv_id":"2506.03447","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The new Observable Equilibration Complexity Measure, entropy times L1 distance to the dephased equilibrium distribution, tracks equilibration and separates high- from low-effective-dimension dynamics.","lead":"This paper defines a complexity measure for how a quantum observable's statistics approach equilibrium and derives bounds on its time average. It tests the measure on an Ising spin chain and argues it distinguishes true equilibration from quasi-periodic coherence-preserving dynamics.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Figure 4a shows the quasi-periodic Up state has the highest time-averaged OECM, contradicting the introduction's claim that OECM 'displays a comparatively faster decay' for Up; the distinguishing power must be reassigned to C(⟨p_t⟩_T).","rationale":"I read the paper as proposing OECM as a diagnostic for equilibration complexity. What would have to be true is that the state with simpler quasi-periodic dynamics is flagged as less complex by the proposed measure. The reported numerical evidence goes against this for the measure defined in Def. 1: the low-effective-dimension Up state is assigned the highest time-averaged complexity in Fig. 4a, while the introduction claims it shows a faster decay and a less complex trajectory. The paper's later, more nuanced discussion presents ⟨C(p_t)⟩_T and C(⟨p_t⟩_T) as complementary witnesses, and that two-measure framework is plausible. But the abstract and introduction present a single measure that 'effectively distinguishes' complex from non-complex trajectories, so the central claim as stated is not yet established. This inconsistency is more load-bearing than the Theorem 2 proof gap identified by the reader, because Theorem 2 is a probabilistic side-result; even if that proof were repaired, the ambiguity about which complexity functional actually supports the headline would remain. The proposed computational check can settle the contradiction directly by comparing the long-time ordering and decay behavior of the two functionals. Because the concern points to a fixable presentation and interpretation issue rather than a fatal flaw, I keep the reader's CONDITIONAL verdict.","tokens_in":15779,"tokens_out":11070,"duration_ms":129694,"concrete_test":"Independently reproduce the N=10 Ising simulation with g=(5+√5)/8, h=(1+√5)/4, J=1, computing for the Up, Dw, and Pm initial states the long-time values and decay rates of both ⟨C(p_t)⟩_T and C(⟨p_t⟩_T) over a common time window. If ⟨C(p_t)⟩_T for Up remains the largest at large T, the introductory 'faster decay' claim for Eq. (16) is false and the central diagnostic claim must be narrowed to C(⟨p_t⟩_T). If instead the Up curve crosses below the others after a transient, the discrepancy is a labeling/definition issue and the central claim survives.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim in the abstract and Section 3 is that the measure defined in Eq. (16), C(p_t)=H_O(p_t)||p_t-p∞||_1, 'effectively distinguishes between complex, irregular dynamics and simpler, quasi-periodic behaviors.' The numerical section reports the opposite ordering for this very measure: Fig. 4a is described as showing that the Up state 'exhibits the highest average observable equilibration complexity, remaining persistently, i.e. pointwise, far from equilibrium,' while Pm and Dw are rapidly suppressed. Yet the introduction asserts that for the same low-effective-dimension Up state 'the Observable Equilibration Complexity Measure displays a comparatively faster decay, indicative of a less complex trajectory.' The two statements cannot both refer to Eq. (16). The paper later moves the distinction to a second functional, C(⟨p_t⟩_T) (Def. 2, Fig. 4b), where Up is indeed low. Thus the headline claim is internally inconsistent unless it is restated to apply only to the time-averaged distribution. A peripheral but real proof gap also exists in Theorem 2: the statement conditions on fixed d_eff, while the proof samples Haar-random states in which d_eff fluctuates; this does not invalidate Def. 1 but should be repaired before the theorem is used.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a statistical complexity measure for observable equilibration, C(p_t) = H_O(p_t) ||p_t - p_∞||_1 (Eq. 16), together with a time-averaged variant C(⟨p_t⟩_T) (Eq. 24), intended to track equilibration and to distinguish irregular, complex dynamics from simpler quasi-periodic behavior. It proves an upper bound on the time-averaged complexity in terms of the effective dimension (Theorem 1), a probabilistic deviation bound (Theorem 2), and supports the analysis with exact-diagonalization simulations of a non-integrable N = 10 Ising spin chain for three initial states (Up, Dw, and Pm). The paper claims that the measure is sensitive to non-complex, coherence-preserving quasi-periodic dynamics.","tokens_in":16053,"tokens_out":7543,"duration_ms":86099,"significance":"If the presentation can be made internally consistent, the proposed two-function diagnostic is a useful contribution: the instantaneous time-average ⟨C(p_t)⟩_T and the complexity of the time-averaged distribution C(⟨p_t⟩_T) are complementary witnesses of temporal non-equilibrium and structural memory. The numerical study explicitly acknowledges that the effective dimension alone does not order the complexities, and Appendix A is candid that the bound in Eq. (23) is generally not tight. The proof of Theorem 1 is a clean Cauchy-Schwarz step conditional on the quoted squared-L1 equilibration bound. However, as written the manuscript's central claim is contradicted by its own Figure 4, and the proof of Theorem 2 has an ensemble mismatch; these issues must be repaired before the contribution can be fully assessed.","major_comments":[{"comment":"The introduction states that for the Up state 'the Observable Equilibration Complexity Measure displays a comparatively faster decay, indicative of a less complex trajectory,' and the abstract and Section 3 make the same separation claim for Eq. (16). Yet Section 4, Figure 4a, reports that 'the Up state exhibits the highest average observable equilibration complexity, remaining persistently, i.e. pointwise, far from equilibrium.' Since Figure 4a plots ⟨C(p_t)⟩_T for the same measure defined in Eq. (16), these statements are mutually contradictory. The numerical results locate the claimed separation in C(⟨p_t⟩_T) (Definition 2, Fig. 4b), where the Up state is indeed low. The abstract, introduction, and Section 3 must be restated so that 'low complexity' of quasi-periodic dynamics refers to the time-averaged distribution, not to the instantaneous OECM of Eq. (16).","section":"Section 1 and Section 4 (Fig. 4a)"},{"comment":"The theorem states that ρ0 is drawn from an ensemble with effective dimension d_eff, but the proof samples |ψ(0)⟩ according to the Haar measure. Under the Haar measure, d_eff is a random variable, so the step E_{ρ0}[f(ε,T)/d_eff] = f(ε,T)/d_eff in Eq. (29) is not justified. If the intended ensemble is Haar-random states conditioned on a fixed d_eff, that ensemble and the corresponding expectation must be defined explicitly; alternatively, the theorem should be formulated for Haar states with d_eff replaced by its typical or averaged value. As written, the proof of Theorem 2 is incomplete.","section":"Section 3, Theorem 2"},{"comment":"The proof of Theorem 1 relies on the squared L1 equilibration bound ⟨||p_t - p_∞||_1^2⟩_T ≤ (r/4d_eff) f(ε,T), attributed without derivation to Appendix 1 of Ref. [14]. This is load-bearing because the advertised 1/√d_eff suppression does not follow from the first-moment bound in Eq. (13) alone; the latter gives ⟨X⟩_T ≤ ..., not ⟨X^2⟩_T ≤ .... Please quote the exact theorem and hypotheses from Ref. [14], or provide a self-contained derivation of Eq. (22), so that Theorem 1 is not conditional on an unstated result.","section":"Section 3, Theorem 1 and Eq. (22)"}],"minor_comments":[{"comment":"The sentence beginning 'However, it is essential to note that even when the observable exhibits substantial oscillations...' is grammatically incomplete and obscures the intended distinction between instantaneous and time-averaged complexity; it should be rewritten.","section":"Section 3, after Definition 1"},{"comment":"The proof invokes a 'Schur-convex function f(t)' that plays no role in the statement; the inequality is Jensen's inequality for the temporal average, and removing the extraneous terminology would improve clarity.","section":"Section 3, Lemma 1"},{"comment":"The sentence 'which is zero for p_t that are pure probability vectors or when they approach equilibrium' should refer to the time-averaged vector ⟨p_t⟩_T, since C(⟨p_t⟩_T) vanishes when ⟨p_t⟩_T is pure or equals p_∞, not when the instantaneous p_t is pure.","section":"Section 3, Definition 2"},{"comment":"Reference [2] is malformed ('J. L. L.') and should be completed, and the nonstandard name 'Riemann's bound' in the proof of Theorem 2 should be replaced by a citation to the specific equilibration result being used.","section":"Section 3 and References"},{"comment":"The claim that the bound for C(⟨p_t⟩_T) follows from 'Theorem 3 of Ref. [49]' is made without stating that theorem; please include the precise statement or derive the bound directly.","section":"Appendix A"},{"comment":"The statement that Definition 2 is 'a particular instance of Definition 1' is potentially misleading if meant to transfer Theorem 1's time-average bound directly; the bound for C(⟨p_t⟩_T) needs its own one-line argument using ||⟨p_t⟩_T - p_∞||_1 ≤ ⟨||p_t - p_∞||_1⟩_T.","section":"Section 3, Eq. (24)"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and the central idea is worth pursuing, but the internal contradiction between the abstract/introduction and Figure 4, together with the Theorem 2 proof gap, require substantive revision. I do not see grounds for rejection if these issues are addressed properly."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my read. The new thing is the OECM, C(p)=H_O(p)||p-p∞||_1, and its time-averaged companion. Redefining López-Ruiz complexity around the dephased equilibrium instead of the uniform distribution is a simple move that nobody seems to have made, and the two-function split (⟨C(p_t)⟩_T versus C(⟨p⟩_T)) gives a genuinely complementary picture: one tracks persistent temporal fluctuations, the other tracks structural memory in the averaged distribution. The numerics on the Ising chain are clear and support that distinction. The paper deserves credit for that.\n\nThe math is thinner. Theorem 1 is a Cauchy-Schwarz step away from the squared-L1 equilibration bound quoted from Ref. [14]; assuming that bound holds, the result is fine. Appendix A is honest that the bound is not tight. Theorem 2 is the real soft spot: the theorem assumes a fixed d_eff for the ensemble, but the proof samples Haar-random states, for which d_eff fluctuates. That's a genuine gap, though not one that sinks Definition 1; it can probably be fixed by conditioning on a fixed d_eff subspace or by rewriting the statement. I'd want that repaired before using Theorem 2.\n\nThe stress-test note is right about the intro. Figure 4a says Up has the highest ⟨C(p_t)⟩_T, but the introduction claims Up's OECM decays faster. Those can't both refer to Eq. (16). The later discussion shifts the claim to C(⟨p⟩_T), where Up is indeed low. So the paper's central 'distinguishes complex from quasi-periodic' claim only works if restated as a property of the time-averaged distribution. That is an internal inconsistency in the presentation, not a fatal mathematical flaw, but it needs correcting.\n\nThe measure is definitionally zero at equilibrium, so 'complexity vanishes at equilibrium' is partly by construction; that is a limitation, not a circularity that destroys the diagnostic. The more interesting content is the separation between the two complexity quantifiers, which is not forced by the definition.\n\nNo code or data is provided; for a numerical paper that's a minor but real deficit.\n\nWho should read it: people working on quantum equilibration diagnostics and quantum thermodynamics. It's not a breakthrough, but the OECM pair is a reasonable addition to the toolbox. I'd send it to peer review, with a request for revision: fix Theorem 2, rewrite the introduction's Up-state claim, clarify which measure does the distinguishing, and deposit the simulation code. After that it would be acceptable.","headline":"A modest but real new diagnostic for observable equilibration, with one proof gap and an internal inconsistency between the introduction and Figure 4a that need fixing before publication.","tokens_in":16601,"tokens_out":2195,"would_cite":false,"duration_ms":23663,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A product of entropy and disequilibrium tracks how closed quantum systems equilibrate.","keywords":["observable equilibration","statistical complexity","effective dimension","quantum thermalization","observable entropy","equilibration bounds","closed quantum systems","spin chain dynamics"],"falsifier":"Compute the time-averaged squared $\\ell^1$ deviation in Eq. (22) directly for the Up state ($d_{\\mathrm{eff}}\\approx 2.95$) on the $N=10$ spin chain at times satisfying $T \\gg 8\\log_2(n)/\\epsilon$; if it ever exceeds $\\frac{r}{4d_{\\mathrm{eff}}} f(\\epsilon,T)$, the quoted bound on which Theorem 1 rests is violated. A second test targets Theorem 2: sample Haar-random initial states, group them by fixed effective dimension, and compare the empirical frequency of $\\lVert \\langle \\mathbf{p}(t)\\rangle_T - \\mathbf{p}_\\infty \\rVert_2 \\ge \\epsilon$ with the promised $\\frac{r}{d_{\\mathrm{eff}}\\epsilon^2} f(\\epsilon,T)$; the proof samples Haar states without conditioning on $d_{\\mathrm{eff}}$, so this test probes exactly that gap.","tokens_in":15575,"feed_emoji":"⚛️","tokens_out":10823,"duration_ms":114032,"temperature":0.7,"pith_summary":"This paper tries to establish that a statistical complexity measure, built from the entropy of an observable's measurement outcomes and their distance from equilibrium, can serve as a quantitative diagnostic of equilibration in closed quantum systems. The measure, called the Observable Equilibration Complexity Measure (OECM), is defined as $C(\\mathbf{p}(t)) = H_O(\\mathbf{p}(t))\\,\\lVert \\mathbf{p}(t)-\\mathbf{p}_\\infty \\rVert_1$, and it is designed to be large only in the intermediate regime where the outcome distribution has both uncertainty and residual memory of the initial state. The paper proves a time-averaged upper bound in terms of effective dimension and a spectral factor, and it argues numerically that the measure separates quasi-periodic, low-effective-dimension behavior from genuinely equilibrating high-effective-dimension dynamics. A sympathetic reader would care because this offers a way to watch equilibrium emerge from unitary dynamics using only observable statistics, without needing the full quantum state.","feed_headline":"Entropy-times-distance tracks quantum equilibration","feed_subtitle":"Two complementary readings of one measure separate quasi-periodic states from true equilibrating dynamics.","key_machinery":"The central object is the product $C(\\mathbf{p}(t)) = H_O(\\mathbf{p}(t))\\,\\lVert \\mathbf{p}(t)-\\mathbf{p}_\\infty \\rVert_1$, which adapts the classical statistical complexity formula 'entropy times disequilibrium' by replacing the uniform reference distribution with the dephased equilibrium distribution $\\mathbf{p}_\\infty$. The argument is carried by three ingredients: the entropy cap $H_O \\le \\log r$, the squared $\\ell^1$ equilibration bound $\\langle \\lVert \\mathbf{p}(t)-\\mathbf{p}_\\infty \\rVert_1^2\\rangle_T \\le \\frac{r}{4 d_{\\mathrm{eff}}} f(\\epsilon,T)$ quoted from [14], and the Cauchy-Schwarz inequality, which together yield the time-averaged bound in Theorem 1. The effective dimension $d_{\\mathrm{eff}}$ is the control parameter: it enters both the bound and the physical interpretation, since low effective dimension means the state explores only a small region of Hilbert space and retains coherence.","core_discovery":"The paper's central claim is that the statistical complexity of an observable's measurement statistics, rather than the global quantum state, can serve as a quantitative diagnostic of equilibration in a closed system. It defines the Observable Equilibration Complexity Measure as $C(\\mathbf{p}(t)) = H_O(\\mathbf{p}(t))\\,\\lVert \\mathbf{p}(t)-\\mathbf{p}_\\infty \\rVert_1$, where $H_O$ is the Shannon entropy of the outcome probabilities and $\\mathbf{p}_\\infty$ is the dephased, infinite-time distribution. The paper proves that the time average of this measure is bounded by $\\frac{\\log r}{2}\\sqrt{\\frac{r}{d_{\\mathrm{eff}}}}\\,f(\\epsilon,T)$, and it shows numerically, on a non-integrable spin chain with $N=10$ spins, that the bound is obeyed while the two time-averaging variants of the measure respond differently to low- and high-effective-dimension initial states. The Up state, with $d_{\\mathrm{eff}}\\approx 2.95$, remains quasi-periodic and coherent: its instantaneous-average complexity stays high while its time-averaged complexity is low. The Dw state, with $d_{\\mathrm{eff}}\\approx 93.74$, equilibrates in distribution yet retains sizable structural complexity because dephasing is slow. The discovery is the complementarity: $\\langle C(\\mathbf{p}(t))\\rangle_T$ records temporal non-equilibrium, while $C(\\langle \\mathbf{p}(t)\\rangle_T)$ records structural memory.","pith_inferences":["Editorial inference: because the measure uses only the probability distribution of measurement outcomes, it could be estimated from repeated projective measurements on a quantum simulator, making it a tomography-free experimental witness of equilibration; the paper does not discuss this route.","Editorial inference: the same construction with the equilibrium reference $\\mathbf{p}_\\infty$ replaced by a generalized Gibbs ensemble would test whether complexity tracks equilibration to non-thermal stationary states; the paper only gestures at extensions to other settings.","Editorial inference: the observed complementarity between the two time-averaging variants hints at a general two-time diagnostic of memory in quantum dynamics, one that might also apply to open systems, although the paper restricts its claims to closed unitary evolution."],"forward_implications":["If the bound in Theorem 1 holds, then for any observable of rank $r$ and any initial state with effective dimension $d_{\\mathrm{eff}}$, the time-averaged OECM must decay at least as $\\frac{1}{2}\\log r\\,\\sqrt{r/d_{\\mathrm{eff}}}$ in the long-time regime where the spectral factor $f(\\epsilon,T)$ approaches 1.","The two variants of the measure can be read together: $\\langle C(\\mathbf{p}(t))\\rangle_T$ is a witness of persistent temporal fluctuations away from equilibrium, while $C(\\langle \\mathbf{p}(t)\\rangle_T)$ is a witness of structural memory retained in the time-averaged distribution; this separates quasi-periodic from genuinely equilibrating dynamics.","Since the measure vanishes both for perfectly ordered (pure) outcome distributions and for fully equilibrated distributions, a nonzero value flags the transient co-existence of uncertainty and order, giving a finite-time signature of the approach to equilibrium.","On the non-integrable spin chain, high-effective-dimension initial states (Dw and Pm) show OECM suppression toward zero in the time-averaged sense, while the low-effective-dimension Up state keeps a non-vanishing instantaneous complexity, demonstrating the diagnostic power on a concrete model."],"supporting_citations":[{"why":"Supplies the squared $\\ell^1$ equilibration bound $\\langle \\lVert \\mathbf{p}(t)-\\mathbf{p}_\\infty \\rVert_1^2\\rangle_T \\le \\frac{r}{4d_{\\mathrm{eff}}} f(\\epsilon,T)$ that Theorem 1 invokes, and motivates the observable-entropy framework.","marker":"[14]"},{"why":"Provides the underlying time-averaged observable equilibration bound and the spectral function $f(\\epsilon,T)=N(\\epsilon)(1+8\\log_2(n)/(\\epsilon T))$ used throughout.","marker":"[22]"},{"why":"Defines the classical statistical complexity measure whose entropy-times-disequilibrium product the OECM adapts.","marker":"[20]"},{"why":"Introduces the quantum statistical complexity measure and notes it vanishes for pure states under unitary evolution, motivating the observable-based classical version.","marker":"[21]"},{"why":"Establishes the dephased equilibrium state $\\omega = \\sum_i \\Pi_i \\rho_0 \\Pi_i$ and the finite-time equilibration tools used in the definitions.","marker":"[9]"},{"why":"Gives the asymptotic time-scale condition under which $f(\\epsilon,T)\\approx 1$, yielding the simplified bound $\\lesssim \\frac{\\log r}{2}\\sqrt{r/d_{\\mathrm{eff}}}$ used in Appendix A.","marker":"[49]"}],"fun_headline_variants":["Complementary complexity averages reveal equilibration's two modes","Entropy–distance measure separates quasi-periodic from equilibrating","Observable complexity tracks emergence of classicality","Two views of complexity tell when quantum states equilibrate","Quantum complexity bound predicts when observables equilibrate"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the squared $\\ell^1$ equilibration bound quoted from [14], $\\langle \\lVert \\mathbf{p}(t)-\\mathbf{p}_\\infty \\rVert_1^2\\rangle_T \\le \\frac{r}{4 d_{\\mathrm{eff}}} f(\\epsilon,T)$, remains valid in the low-effective-dimension regime where Theorem 1 is applied numerically; the paper takes this bound on faith rather than re-deriving it.","fun_headline_variants_meta":{"raw":{"variants":["Complementary complexity averages reveal equilibration's two modes","Entropy–distance measure separates quasi-periodic from equilibrating","Observable complexity tracks emergence of classicality","Two views of complexity tell when quantum states equilibrate","Quantum complexity bound predicts when observables equilibrate"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001556,"raw_usage":{"total_tokens":6270,"prompt_tokens":1048,"completion_tokens":5222,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":664,"completion_tokens_details":{"reasoning_tokens":5144}},"tokens_in":664,"tokens_out":5222,"duration_ms":44847,"temperature":1.0,"reasoning_tokens":5144,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T11:03:04.136396+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the time-averaged squared $\\ell^1$ deviation in Eq. (22) directly for the Up state ($d_{\\mathrm{eff}}\\approx 2.95$) on the $N=10$ spin chain at times satisfying $T \\gg 8\\log_2(n)/\\epsilon$; if it ever exceeds $\\frac{r}{4d_{\\mathrm{eff}}} f(\\epsilon,T)$, the quoted bound on which Theorem 1 rests is violated. A second test targets Theorem 2: sample Haar-random initial states, group them by fixed effective dimension, and compare the empirical frequency of $\\lVert \\langle \\mathbf{p}(t)\\rangle_T - \\mathbf{p}_\\infty \\rVert_2 \\ge \\epsilon$ with the promised $\\frac{r}{d_{\\mathrm{eff}}\\epsilon^2} f(\\epsilon,T)$; the proof samples Haar states without conditioning on $d_{\\mathrm{eff}}$, so this test probes exactly that gap.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the squared $\\ell^1$ equilibration bound $\\langle \\lVert \\mathbf{p}(t)-\\mathbf{p}_\\infty \\rVert_1^2\\rangle_T \\le \\frac{r}{4d_{\\mathrm{eff}}} f(\\epsilon,T)$ that Theorem 1 invokes, and motivates the observable-entropy framework."},{"cited_title":"Lostaglio, D","cited_arxiv_id":null,"evidence_quote":"Provides the underlying time-averaged observable equilibration bound and the spectral function $f(\\epsilon,T)=N(\\epsilon)(1+8\\log_2(n)/(\\epsilon T))$ used throughout."},{"cited_title":"Anz` a and V","cited_arxiv_id":null,"evidence_quote":"Introduces the quantum statistical complexity measure and notes it vanishes for pure states under unitary evolution, motivating the observable-based classical version."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the dephased equilibrium state $\\omega = \\sum_i \\Pi_i \\rho_0 \\Pi_i$ and the finite-time equilibration tools used in the definitions."},{"cited_title":"L´ opez-Ruiz, J","cited_arxiv_id":null,"evidence_quote":"Gives the asymptotic time-scale condition under which $f(\\epsilon,T)\\approx 1$, yielding the simplified bound $\\lesssim \\frac{\\log r}{2}\\sqrt{r/d_{\\mathrm{eff}}}$ used in Appendix A."}],"review_version":1}