{"id":"b9f0dc43-9f57-4f90-9194-7b190e93e4ea","arxiv_id":"2506.22717","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Random Lindbladians built from Student's t-distributed interactions with infinite variance show near-zero spectral gaps, a low-dissipation bulk, and order-of-magnitude longer coherence times with enhanced perturbation sensitivity compared with Gaussian ensembles.","lead":"This paper studies random open quantum systems whose interactions with the environment follow heavy-tailed statistics instead of the usual Gaussian distribution. It finds that extremely heavy tails can make quantum states both much longer-lived and more sensitive to small perturbations, a combination that is useful for quantum sensing.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central gapless/long-lived claim rests on N=50 extrapolation; without finite-size scaling, the asymptotic Δλ → 0 and the two-orders-of-magnitude enhancement could be finite-size effects.","rationale":"The strongest claim of the paper is that extremely heavy-tailed system-environment couplings (ν ≤ 2) produce a nearly gapless Liouvillian spectrum, low-dissipation bulk states, and quasi-degeneracies, yielding coherence lifetimes and sensitivities enhanced by roughly two orders of magnitude. The reader's verdict is CONDITIONAL, and the reader's leading weakness is that the N = 50 simulations are used to extrapolate Δλ → 0 without finite-size scaling. I agree that this is the most load-bearing concern: the central asymptotic statement is not directly measured, and the mechanism proposed (single-big-jump collapse of the Kossakowski matrix) is inherently a thermodynamic-limit statement. The concrete scaling test I propose would settle whether the gap truly vanishes and whether the reported enhancement persists. I do not think this concern warrants moving the verdict to REJECT or UNVERDICTED, because the paper is internally consistent, the numerics are extensive at N = 50, and the missing scaling analysis is a well-defined, addressable check. The Methods basis-universality proof is also questionable for ν > 2 because the single-big-jump principle requires infinite variance, but this does not undermine the EHT-regime headline, so it is secondary to the scaling issue. The reader's CONDITIONAL verdict remains appropriate, hence UNCHANGED.","tokens_in":15490,"tokens_out":19953,"duration_ms":307058,"concrete_test":"Run the purely dissipative ensemble at ν = 1 and ν = 2 for system sizes N = 16, 32, 64, 128 (at least 64 realizations each) and plot the median spectral gap Δλ and the normalized off-rank-one weight of K (e.g., 1 − λmax(K)/Tr[K]) against N on a log-log scale. If log Δλ decreases linearly in log N with a nonzero slope, the asymptotic gapless claim is supported; if Δλ saturates to a positive value, the headline enhancement is a finite-size effect.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's load-bearing asymptotic statement — that the Liouvillian gap 'ultimately vanishes (Δλ → 0) near ν = 1' (Fig. 2e, Fig. 3a) — is inferred entirely from N = 50 ensembles with 64 realizations. No finite-size scaling is shown, so the thermodynamic-limit behavior is not established. This matters because the proposed mechanism is itself a large-N phenomenon: for ν ≤ 2, the entries of X have infinite variance, and by the single-big-jump principle the sum Tr[XX†] is dominated by the largest entry, driving K = NXX†/Tr[XX†] toward a rank-one projector. The spectral gap at finite N is then controlled by the small off-rank-one remainder of K, and whether that remainder vanishes fast enough with N is exactly what determines whether the system is asymptotically gapless. If the gap instead saturates to a positive constant or decays slower than the reported enhancement suggests, the 'long-lived and ultrasensitive' coherence would be a finite-size artifact rather than a genuine CLT-violation effect. A secondary but related issue is that the universality proof in Methods invokes the single-big-jump principle for all ν; that principle is only valid in the infinite-variance regime ν ≤ 2. For ν > 2, entries of UX are sums of finite-variance variables and become Gaussian under the CLT, so the claimed basis independence is not established for the HT/MHT regimes. The main headline, however, concerns the EHT regime, so the decisive missing piece is the scaling check.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies Markovian open quantum systems whose system-environment interactions are modeled by a Student-t distributed random matrix X, replacing the Gaussian Ginibre ensemble. The Kossakowski matrix is constructed as K = NXX†/Tr[XX†], and the tail index ν controls a continuous crossover from light-tailed to extremely heavy-tailed statistics. For ν ≤ 2, the 'Extremely Heavy Tail' regime, the authors report a nearly gapless Liouvillian spectrum, a low-dissipation bulk accompanied by high-dissipation outliers, reduced Petermann factors, and an increased density of near-degenerate eigenvalues. Direct Lindblad simulations with GUE Hamiltonians and Haar-random initial states show coherence times and perturbation-induced coherence changes enhanced by roughly two orders of magnitude relative to GinUE predictions. The paper interprets these results as a breakdown of the usual stability-sensitivity tradeoff in open quantum systems.","tokens_in":15811,"tokens_out":5366,"duration_ms":63910,"significance":"If the gapless behavior is confirmed in the thermodynamic limit, the paper identifies a genuinely new mechanism for long-lived coherence in open quantum systems, one that does not rely on decoherence-free subspaces, symmetries, correlated environments, or dynamical decoupling. The numerical work is extensive and transparent: 64 independent realizations of N = 50 systems, 160,000 eigenvalues per ensemble, direct solution of the Lindblad equation, and explicit code and data availability. The qualitative trends are robust across purely dissipative and GUE-Hamiltonian cases, and no target quantity is used to fit parameters, so circularity is not a concern. The main risk is the extrapolation from N = 50 to the claimed gapless limit, which is load-bearing for the headline claims.","major_comments":[{"comment":"The central claim that the spectral gap 'ultimately vanishes (Δλ → 0) near ν = 1' is inferred from ensembles of N = 50 with 64 realizations, with no finite-size scaling presented. This matters because the proposed mechanism is inherently a large-N effect: for ν ≤ 2, the entries of X have infinite variance, and by the single-big-jump principle the sum Tr[XX†] is dominated by the largest entry, driving K toward a rank-one projector. The finite-N gap is then controlled by the small off-rank-one remainder, and whether that remainder vanishes fast enough with N is exactly what determines whether the system is asymptotically gapless. Without a scaling analysis of Δλ versus N (or an analytic bound), the 'two orders of magnitude' enhancement in T2 and the claim of long-lived coherence could be finite-size artifacts rather than genuine CLT-violation effects. Please provide Δλ (and ideally the low-dissipation CCDF and T2) as a function of N for ν = 1 and ν = 2, with the GinUE case as a control.","section":"Tail dependence, Figs. 2e and 3a"},{"comment":"The basis-independence argument as written is compressed and mixes regimes. For ν > 2, the entries of X have finite variance, and under unitary transformations the new entries are sums of many terms; the text invokes the single-big-jump principle to claim the tail index is preserved, but that principle is commonly stated for infinite-variance sums. The EHT regime (ν ≤ 2) is the load-bearing case for the paper's main claims, and the argument there is plausible, but the general statement that 'our modelling... allows the examination of universal tail behaviours independent from the chosen basis' needs either an explicit appeal to the regular-variation convolution theorem, valid for all ν > 0, or a clear restriction of the universality claim to the EHT regime. This is a repair, not a rejection, but it should be addressed for the paper to support its stated universality.","section":"Methods, 'Universality in tail behaviours'"}],"minor_comments":[{"comment":"The definition of ΔCE is garbled in the typeset text; the integral and the prefactor are not readable. Please restore the correct formula.","section":"Eq. (5)"},{"comment":"The phrase 'MTH' appears to be a typo for 'MHT' (Moderately Heavy Tail); please correct it.","section":"Text near Fig. 3"},{"comment":"The caption states that panels (c-e) use the same plot range as panels (a,b), but panel (f) is discussed as showing the full range at ν = 1; please clarify the plotting ranges for all panels, including panel (f).","section":"Caption of Fig. 2"},{"comment":"The phrase 'the first standard form' of the GKSL equation is unusual; consider rewording to 'the standard Lindblad form' or similar.","section":"Introduction and Eq. (1)"},{"comment":"The code is identified as 'Q-ROS' and provided as Supplementary Code S1; please confirm that the code archive is complete and that the exact random-generation procedures (including the Student-t sampling algorithm) are documented so the results can be reproduced.","section":"Data and code availability"}],"recommendation":"major_revision","confidential_remarks":"The paper is a strong candidate if the finite-size scaling confirms the extrapolation from N = 50 to the gapless limit. The missing scaling analysis is the primary obstacle; the universality argument is secondary and repairable. The self-citations are appropriate as motivation, and the numerical effort is substantial and clearly described. If the authors provide the requested scaling data, I would expect the paper to be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The short version: this is a genuinely new numerical result. Using a Student's t distribution to interpolate between Gaussian (GinUE) and heavy-tailed randomness in the Kossakowski matrix, the authors show that for ν ≤ 2 (infinite variance) the Liouvillian spectrum becomes nearly gapless, dissipation concentrates in a few outlier states, and the remaining bulk states are both long-lived and highly sensitive to perturbations. That combination directly challenges the usual stability–sensitivity tradeoff. The numerics are substantial: N=50, 64 realizations, 160,000 eigenvalues, plus checks with GUE Hamiltonians and Haar initial states.\n\nThe genuinely new piece is the continuous ν interpolation and the identification of ν=2 as the threshold where CLT violation kicks in. Previous work on random Lindbladians used Gaussian/Wishart ensembles or separate power-law deformations; the joint long-lived + ultrasensitive behavior is not in those papers. The paper also gives a plausible mechanism: as ν decreases, the Gram matrix K = NXX†/Tr[XX†] becomes dominated by the largest entry, driving a low-dissipation bulk with near-degenerate eigenvalues. The orthogonality recovery (decreasing Petermann factors) is a nice bonus.\n\nThe soft spots are real but addressable. The biggest one: the gapless claim (Δλ→0 near ν=1) is extrapolated from N=50. No finite-size scaling is shown. Since the mechanism is fundamentally a large-N effect, a scaling collapse across N=50, 100, 200 would be the decisive check; without it, the two-orders-of-magnitude enhancement could be a finite-size artifact. A second technical gap: the basis-invariance argument in Methods uses the single-big-jump principle, which is strictly valid for infinite variance (ν≤2). For finite-variance heavy tails (2<ν≤4), the transformed entries are sums of finite-variance variables and are not guaranteed to keep the same tail index. This affects the HT/MHT universality claims, though not the EHT headline. Minor: the coherence time is fit to a single exponential, which is a rough approximation for a multi-exponential decay, but the order-of-magnitude differences are so large that this likely doesn't change the conclusion. And the code is only referenced as Supplementary Code S1, not publicly committed; that should be fixed.\n\nOverall: this deserves a serious referee. A good referee should ask for finite-size scaling, a more careful universality statement, and public code. If the scaling confirms the gap narrowing as N grows, it's a strong paper. If the gap saturates, the headline collapses. My recommendation: send it out, with a major-revision expectation.","headline":"A plausible and well-executed numerical study of heavy-tailed random Lindbladians; the main missing piece is finite-size scaling for the gapless claim.","tokens_in":16349,"tokens_out":3793,"would_cite":true,"duration_ms":42322,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper argues that extremely heavy-tailed system-environment couplings can make open quantum systems simultaneously long-lived and ultrasensitive, breaking the usual stability-sensitivity tradeoff.","keywords":["heavy-tailed distributions","open quantum systems","Lindblad master equation","Liouvillian spectral gap","quantum coherence","quantum sensing","Petermann factor","central limit theorem"],"falsifier":"Compute the spectral gap $\\Delta\\lambda$ at $\\nu=1$ for increasing system sizes ($N=50,100,200,400$) with the same Student's t construction; if $\\Delta\\lambda(N)$ saturates to a positive constant instead of shrinking toward zero, the gapless-spectrum claim fails. A second decisive check is to replace Student's t entries with a different infinite-variance distribution with the same tail exponent, such as symmetric $\\alpha$-stable noise, and verify that the gap narrowing and coherence enhancement persist; if the effect depends on the distribution's interior, the claimed universality is not established.","tokens_in":15275,"feed_emoji":"⚛️","tokens_out":10016,"duration_ms":104429,"temperature":0.7,"pith_summary":"This paper tries to show that the usual way of modeling randomness in open quantum systems—drawing system-environment couplings from Gaussian distributions, as the central limit theorem would suggest—misses a regime with qualitatively different behavior. Replacing those couplings with heavy-tailed Student's t distributions that have infinite variance makes the Liouvillian's spectrum nearly gapless, so most quantum states decay slowly while a few outlier states absorb most of the dissipation. The paper argues that, as a result, quantum coherence can be simultaneously long-lived and highly sensitive to perturbations, with both the coherence time and the response to a small Hamiltonian perturbation enhanced by about two orders of magnitude relative to Gaussian-random systems. If true, this breaks the usual stability-versus-sensitivity tradeoff and points to disordered, symmetry-free platforms for quantum sensing.","feed_headline":"Heavy-tailed randomness gives long-lived, ultrasensitive coherence","feed_subtitle":"Student-t heavy-tailed interactions push coherence time and perturbation response up by about two orders of magnitude.","key_machinery":"The load-bearing object is the Kossakowski matrix $K = NXX^\\dagger/\\mathrm{Tr}[XX^\\dagger]$, the positive semidefinite matrix of interaction coefficients in the Lindblad master equation; the paper constructs it from a Ginibre random matrix $X$ whose entries are Student's t distributed with tail parameter $\\nu$ instead of Gaussian. The argument is carried by the single-big-jump principle: because Student's t tails are regularly varying, unitary changes of the Lindblad operator basis preserve the power-law tail $\\mathbb{P}(|X_{ij}|>x)\\sim x^{-\\nu}$, so the spectral features are claimed to be basis-independent. In the extremely heavy-tail regime ($\\nu\\le2$) the infinite variance of the couplings produces the single-big-jump domination that creates high-dissipation outliers and a low-dissipation bulk, a nearly gapless Liouvillian spectrum, and quasi-degenerate low-lying states.","core_discovery":"On the paper's own terms, the central claim is that when the system-environment interaction matrix entering the Lindblad master equation is built from Student's t-distributed entries with $\\nu \\le 2$, the Liouvillian spectrum reorganizes: a few highly dissipative outlier eigenstates appear, a large low-dissipation bulk forms, the spectral gap $\\Delta\\lambda$ narrows toward zero as $\\nu$ approaches 1, Petermann factors drop (eigenstates become more orthogonal), and near-degenerate eigenvalues become abundant. The paper then shows, in numerical Lindblad dynamics with GUE Hamiltonians and Haar-random initial states, that coherence times grow by one to three orders of magnitude and both coherence-time perturbation and time-averaged coherence perturbation rise by roughly two orders of magnitude compared with Gaussian-random (central-limit-theorem) systems. It concludes that long-lived and ultrasensitive coherence coexist in the same heavy-tailed open quantum system, with quasi-degeneracy supplying sensitivity and gap narrowing supplying longevity.","pith_inferences":["A direct experimental route, not pursued in the paper, would be to engineer reservoirs with tunable heavy-tailed couplings (for example, disorder-controlled photonic or atomic systems) and measure the $\\nu$-dependence of coherence time; a sharp change near $\\nu=2$ would confirm the central-limit-theorem boundary.","The same single-big-jump logic suggests that other infinite-variance distributions with the same tail exponent, not just Student's t, should reproduce the gap narrowing and coherence enhancement, which is a testable universality prediction beyond the paper's explicit numerics.","Because the sensitivity mechanism is quasi-degeneracy rather than exceptional points, heavy-tailed open systems may provide a generic route to high-order perturbative sensitivity in disordered non-Hermitian platforms, an implication the paper raises only for sensing rather than for a broader class of non-Hermitian physics."],"forward_implications":["Coherence times in the $\\nu\\approx1$ regime are predicted to exceed Gaussian-ensemble coherence times by one to three orders of magnitude, with the slowest relaxation set by a spectral gap $\\Delta\\lambda$ that tends to zero.","The same states should respond strongly to weak Hamiltonian perturbations, because quasi-degenerate eigenvalues (nearest-neighbor spacings down to $\\sim10^{-5}$) make higher-order perturbation theory dominate.","At $\\nu=1$ the eigenstates should obey a 20:80 Pareto split: roughly 80% of states sit in the low-dissipation bulk ($\\mathrm{Re}\\,\\lambda\\ge-1$) while a minority of outliers carry most dissipation.","Long-lived coherence should be achievable with static, random, symmetry-free system-environment interactions, without suppressing interaction strength, adding symmetries or correlations, or using measurement-based protection."],"supporting_citations":[{"why":"It supplies the central-limit-theorem baseline: the Wishart-Laguerre Kossakowski matrix and lemon-shaped Liouvillian spectrum with the spectral gap $\\Delta\\lambda \\approx 1 - 2/N$ that the paper's heavy-tail results are compared against.","marker":"[12]"},{"why":"It defines the Ginibre unitary ensemble whose Gaussian entries are replaced by Student's t entries to build the heavy-tailed Kossakowski matrix.","marker":"[18]"},{"why":"It provides the Student's t distribution and its moments property (only moments of order $r<\\nu$ are finite), which sets the EHT, HT, and MHT classification used throughout the paper.","marker":"[27]"},{"why":"It supplies the single-big-jump principle used to argue that power-law tails of $X$ are preserved under unitary basis changes of the Lindblad operators.","marker":"[29]"},{"why":"It establishes the GKSL/Lindblad master equation and the Kossakowski matrix structure that the whole model is built on.","marker":"[24,25]"},{"why":"It defines the Petermann factor used to measure nonorthogonality and first-order sensitivity of Liouvillian eigenstates.","marker":"[33]"},{"why":"It connects the Petermann factor and exceptional-point-enhanced sensitivity, the framework the paper extends to quasi-degenerate heavy-tailed systems.","marker":"[34]"}],"fun_headline_variants":["Heavy tails give quantum coherence a 100x boost in both life and sensitivity","Student-t noise yields long-lived, ultra-sensitive quantum coherence","Beyond Gaussian: heavy tails amplify coherence and sensing twofold","Heavy-tailed interactions stretch coherence and sharpen quantum response"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central conclusion assumes that the nearly gapless spectrum and the two-order-of-magnitude coherence enhancement seen in the $N=50$ simulations persist in the thermodynamic limit, and that the power-law tail exponent survives arbitrary unitary basis changes of the Lindblad operators; neither is proven by finite-size scaling.","fun_headline_variants_meta":{"raw":{"variants":["Heavy tails give quantum coherence a 100x boost in both life and sensitivity","Student-t noise yields long-lived, ultra-sensitive quantum coherence","Beyond Gaussian: heavy tails amplify coherence and sensing twofold","Heavy-tailed interactions stretch coherence and sharpen quantum response"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000174,"raw_usage":{"total_tokens":1269,"prompt_tokens":916,"completion_tokens":353,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":532,"completion_tokens_details":{"reasoning_tokens":283}},"tokens_in":532,"tokens_out":353,"duration_ms":5066,"temperature":1.0,"reasoning_tokens":283,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T22:00:07.228303+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the spectral gap $\\Delta\\lambda$ at $\\nu=1$ for increasing system sizes ($N=50,100,200,400$) with the same Student's t construction; if $\\Delta\\lambda(N)$ saturates to a positive constant instead of shrinking toward zero, the gapless-spectrum claim fails. A second decisive check is to replace Student's t entries with a different infinite-variance distribution with the same tail exponent, such as symmetric $\\alpha$-stable noise, and verify that the gap narrowing and coherence enhancement persist; if the effect depends on the distribution's interior, the claimed universality is not established.","supporting_citations":[{"cited_title":"Statistical ensembles of complex, quaternion, and real matrices","cited_arxiv_id":null,"evidence_quote":"It defines the Ginibre unitary ensemble whose Gaussian entries are replaced by Student's t entries to build the heavy-tailed Kossakowski matrix."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It provides the Student's t distribution and its moments property (only moments of order $r<\\nu$ are finite), which sets the EHT, HT, and MHT classification used throughout the paper."},{"cited_title":"& Burioni, R","cited_arxiv_id":null,"evidence_quote":"It supplies the single-big-jump principle used to argue that power-law tails of $X$ are preserved under unitary basis changes of the Lindblad operators."},{"cited_title":"Calculated spontaneous emission factor for double-heterostructure injection lasers with gain-induced waveguiding","cited_arxiv_id":null,"evidence_quote":"It defines the Petermann factor used to measure nonorthogonality and first-order sensitivity of Liouvillian eigenstates."},{"cited_title":"& Vahala, K","cited_arxiv_id":null,"evidence_quote":"It connects the Petermann factor and exceptional-point-enhanced sensitivity, the framework the paper extends to quasi-degenerate heavy-tailed systems."}],"review_version":1}