{"id":"106cbd95-0cc8-4678-a9c2-782263af56d0","arxiv_id":"2411.08303","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A coupling inequality bounds the discrepancy between min-max statistics of two independent random matrices under finite third moments, extending the Gaussian Gordon-type inequalities.","lead":"This paper proves a tail inequality that controls how different the min-max statistics of two independent random matrices can be, assuming only finite third moments for the entries. It extends earlier bounds that required Gaussian entries, so it applies to broader random matrix models.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The proof's final step misuses Strassen's lemma: distributional closeness of min-max statistics does not imply a small tail probability for the independent pair X,X' stated in Theorem 1.2.","rationale":"Reader's rejection is justified, but the most load-bearing point is not the derivative-sum inheritance for h in Theorem 2.8. That concern is real but readily repairable: for h(x)=integral_0^1 (1/(2t)) E[f(sqrt(t)x+sqrt(1-t)Y)-f(Y)] dt, one has d^2 h = integral_0^1 (1/2) E[d^2 f(...)] dt and d^3 h = integral_0^1 (sqrt(t)/2) E[d^3 f(...)] dt, so the sup-norm bounds of Lemma 2.5 pass to h with unchanged constants. The missing (1-epsilon)^{-1} factor in the final chain is also a genuine algebraic error, but it only weakens the numerical bound. The decisive problem is that Theorem 1.2's event is defined on the independent pair X,X', while the proof establishes at most a distributional comparison of the two laws. Strassen's lemma cannot convert that comparison into a tail bound for the independent coupling; the Bernoulli counterexample shows the inference is false in general. Thus the main theorem is not proven as stated. The verdict remains REJECT; a corrected statement (either as a true coupling inequality or with an additional argument controlling the independent coupling) and a repaired proof would be needed.","tokens_in":10082,"tokens_out":18075,"duration_ms":185785,"concrete_test":"Take n=m=1 with X and X' i.i.d. taking values +/-1 with probability 1/2. The distributional inequality mu(A)<=nu(A_{0.5}) holds for all A with epsilon=0, but P(|X-X'|>0.5)=1/2. Run this counterexample through the final paragraph of the proof of Theorem 1.2: Lemma 2.1 produces a coupling with small discrepancy only when one is allowed to choose W jointly with V; for the fixed independent pair it is silent. This settles that the proof's last inference is invalid, independent of the values of B1, B2, and B3.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central theorem as stated compares independent matrices X and X', yet the proof's final step ('By Lemma 2.1, we complete our proof') concludes this from a distributional inequality for every Borel set A. Lemma 2.1 (Strassen) is an equivalence with the existence of some coupling (V,W) with P(|V-W|>delta)<=epsilon; it does not control the product coupling that corresponds to the independent X and X'. Consequently the inequality P(|minmax X - minmax X'|>...) <= ... is not established. The gap is not a constant-factor issue: for mu=nu=Bernoulli(1/2), the distributional condition mu(A)<=nu(A_delta) holds with epsilon=0 for delta=0.5, while independent copies differ with probability 1/2. Thus the proof's inference is invalid. A separate algebraic slip in the chain after (2.3) drops a (1-epsilon)^{-1} factor, but repairing it would still not bridge the Strassen gap. If the intended claim is instead a coupling inequality, Theorem 1.2 must be restated; as written, the theorem is unsupported.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims a quantitative comparison of min-max statistics of two independent n x m random matrices with mean zero and finite absolute third moments. Theorem 1.2 asserts a tail bound on |min_i max_j X_ij - min_i max_j X'_ij| in terms of the moment quantities B1, B1', B2, B2', B3 and a constant epsilon defined in (1.1). The proof smooths the non-smooth min-max functional through the function F_{beta,delta}, approximates indicators by a smooth function via Lemma 2.4, compares expectations E[f(X)] and E[f(X')] using exchangeable pairs and Gaussian interpolation in Theorem 2.8, and concludes by invoking Lemma 2.1 (Strassen's theorem).","tokens_in":10294,"tokens_out":6638,"duration_ms":66910,"significance":"If valid, the result would be a substantial non-Gaussian extension of the quantitative Gordon/Sudakov-Fernique inequalities obtained by Peccati and Turchi, with explicit dependence on third-moment quantities. The smoothing construction and the decomposition into B1, B2, B3 are natural and the constants are explicit rather than fitted. However, the proof as written has a load-bearing gap in the final Strassen step and an unjustified application of Lemma 2.5 to the Stein helper function h, so the main theorem is not established.","major_comments":[{"comment":"The final inference 'By Lemma 2.1, we complete our proof' is invalid. Lemma 2.1 (Strassen) states that the distributional inequality mu(A) <= nu(A_delta) + epsilon is equivalent to the existence of some coupling (V,W) on a common probability space with P(|V-W| > delta) <= epsilon. The chain preceding it establishes, at most, such a distributional inequality for the laws of minmax(X) and minmax(X'). It does not control P(|minmax(X) - minmax(X')| > ...) for the independent pair X,X' stated in Theorem 1.2. As a concrete obstruction, if mu = nu = Bernoulli(1/2) and delta = 1/2, the distributional condition holds with epsilon = 0, while two independent copies differ by 1 with probability 1/2. Thus the theorem as stated is not a consequence of the preceding argument; at best the proof supports a different coupling statement.","section":"Section 2, proof of Theorem 1.2, final sentence"},{"comment":"There is an algebraic slip in the chain following Theorem 2.8. From (2.3), P(minmax(X) in A) <= (1-epsilon)^{-1} E[g(F(X))]. Combining this with (2.10) gives (1-epsilon)^{-1} E[g(F(X'))] + C beta (1+delta) tau^{-1}(...)/(1-epsilon), not E[g(F(X'))] + C beta (1+delta) tau^{-1}(...)/(1-epsilon) as written. Applying the upper bound for g from Lemma 2.4 then yields an epsilon/(1-epsilon) term rather than epsilon, and the claimed constants in Theorem 1.2 are not supported as stated.","section":"Equations (2.3) and the chain after (2.10)"},{"comment":"Lemma 2.5 gives derivative-sum bounds for h0 = g composed with F_{beta,delta}, namely (2.4) and (2.5) in terms of ||g'||, ||g''||, and ||g'''||. In the proof of Theorem 2.8, these bounds are instead applied to the function h defined in Lemma 2.6, the Stein/Poisson-solver helper for f = g composed with F_{beta,delta}. The proof does not show that h inherits the needed bounds on the sums of its second and third derivatives from f; in particular, the estimates 'by (2.4)' and 'by (2.5)' in the bounds for the B1 and B2 terms are unjustified. This is a substantive technical gap, not a typo, because h is defined via an integral involving f(sqrt(t)x + sqrt(1-t)Y) and its derivatives are not directly the same as those of f.","section":"Proof of Theorem 2.8, paragraphs after Lemma 2.6"}],"minor_comments":[{"comment":"There are several typos: 'Chaterjee' should be 'Chatterjee', 'whcih' should be 'which', 'borel' should be 'Borel', and 'Elizabeth' should be 'Meckes'.","section":"Throughout"},{"comment":"The line 'P(g composed with F(X') in A_{lambda+3tau}) + epsilon + ...' is not the correct intermediate step: Lemma 2.4 gives E[g(F)] <= epsilon + (1-epsilon) P(F in A_{lambda+3tau}), so the intermediate probability should involve F(X') in A_{lambda+3tau}, not g(F(X')) in that set. The subsequent final bound on minmax(X') is consistent with the corrected version, but the displayed chain should be repaired.","section":"Proof of Theorem 1.2, displayed chain"},{"comment":"The bound on the second-derivative sum in the Gaussian interpolation step is attributed to '(2.5)', but the relevant second-derivative bound is (2.4). This is a citation typo that should be corrected.","section":"Equation (2.9)"},{"comment":"The proof of Lemma 2.2 introduces beta' and then identifies beta' = beta delta at the end; the intermediate notation is confusing and the displayed inequalities would benefit from a clearer separation of the two approximation steps.","section":"Lemma 2.2 proof"}],"recommendation":"reject","confidential_remarks":"I agree with the reader's assessment. The main theorem as stated is not supported by the proof: the Strassen step is a fundamental logical gap, and the missing derivative bounds for h affect the core estimate in Theorem 2.8. These are not local presentation issues. If the authors restate the main result as an existence-of-coupling statement and supply the missing derivative estimates, a substantially revised version might be reconsidered, but the current manuscript does not establish its central claim."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper has a genuine new idea: a quantitative min-max comparison for matrices with only finite third moments, obtained by combining exchangeable-pair Stein methods with a smooth free-energy approximation. The derivative computations for F are careful, and the borrowing from Chernozhukov-Chetverikov-Kato and Meckes is sensible. If the goal were a coupling inequality, the strategy would be close to right.\n\nBut the theorem as stated is not proved, and in fact it cannot be true. The proof ends by showing, for every Borel set A, that P(minmax X in A) is bounded by P(minmax X' in A^{2lambda+3tau}) plus an error term. Strassen's lemma (Lemma 2.1) turns that distributional inequality into the existence of some coupling (V,W) with P(|V-W|>delta) small. It says nothing about the original independent pair X,X'. For Bernoulli(1/2) marginals, the distributional condition holds with delta=0.5 and epsilon=0, while two independent copies differ with probability 1/2. So the final 'By Lemma 2.1' step is invalid. The theorem must be restated as a coupling or distributional comparison, or the tail probability must refer to the coupled pair, not the independent matrices.\n\nSeparately, after (2.3) a factor (1-epsilon)^{-1} is dropped: the chain should give (1-epsilon)^{-1} E[g(F(X'))], and then the epsilon from Lemma 2.4 should appear as epsilon/(1-epsilon). That is a minor algebraic slip, but it needs fixing. The application of Lemma 2.5 to the Poisson-solver function h is also not justified: h is an integral of f, and one has to check that the derivative-sum bounds survive the integration. This is likely repairable by differentiating under the integral and integrating the t-powers, but it is not in the paper. Finally, the Gaussian corollary in Theorem 1.2 drops the B1 and B2 terms, which do not vanish for Gaussian matrices; that statement as written is also false.\n\nSpecialists in high-dimensional probability who care about quantitative Gordon-type inequalities would find the intended result useful, but the main theorem is unsupported as written. I would not accept it. If the authors restate the theorem, fix the algebra, and prove the h bounds, the paper could become publishable. It deserves a careful referee's time in the sense that a referee could guide that repair, but it should not be published in this form.","headline":"A promising non-Gaussian min-max comparison idea, but the final Strassen step proves a coupling, not the independent-pair tail the theorem claims.","tokens_in":10830,"tokens_out":7430,"would_cite":false,"duration_ms":78102,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60F05","60G15","60G70","60H05"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves a quantitative Gordon-type comparison inequality for min-max statistics of two random matrices, requiring only finite absolute third moments.","keywords":["min-max statistics","random matrices","Stein's method","exchangeable pairs","Gaussian approximation","Gaussian vectors","coupling","quantitative Gordon inequality"],"falsifier":"For a concrete non-Gaussian ensemble such as i.i.d. Rademacher entries, compute by simulation the probability that the min-max statistics of two independent copies differ by more than the threshold, and compare it to the bound for several $\\beta$, $\\delta$, $\\tau$; a single violation for valid parameters would refute the theorem.","tokens_in":9830,"feed_emoji":"🎲","tokens_out":12124,"duration_ms":98420,"temperature":0.7,"pith_summary":"This paper proves a quantitative comparison inequality for the min-max statistics of two independent random matrices. The only conditions on the entries are zero mean and finite absolute third moments, replacing the Gaussian assumption required by earlier Gordon-type bounds. The theorem gives an explicit upper bound on the probability that the row-minimum/column-maximum statistics of the two matrices differ by more than a threshold set by the parameters $\\beta$, $\\delta$, and $\\tau$. The bound is expressed through the fluctuation of entry products, the maximum third moment of entries, and the difference of the covariance structures. This matters because min-max statistics appear throughout random matrix theory, optimization, and empirical process theory, where Gaussian assumptions are often too restrictive.","feed_headline":"Min-max gap bound extends to finite third moments","feed_subtitle":"A quantitative comparison for min-max statistics of non-Gaussian random matrices, requiring only finite third moments.","key_machinery":"The key machinery is the smooth 'soft min-max' function $F_{\\beta\\delta}(x) = -\\frac{1}{\\beta\\delta} \\log \\sum_{i=1}^n \\left(\\sum_{j=1}^m \\exp(\\beta x_{ij})\\right)^{-\\delta}$, which approximates $\\min_i \\max_j x_{ij}$ within a window of width $\\frac{1}{\\beta}\\log m + \\frac{1}{\\beta\\delta}\\log n$. The proof's core is Lemma 2.5, which shows that for any thrice-smooth $g$, the sums of the second and third derivatives of $g \\circ F_{\\beta\\delta}$ over all index pairs and triples are bounded by $\\beta$, $\\delta$, and the derivative norms of $g$. These bounds let the expectation of $g \\circ F_{\\beta\\delta}(X)$ be compared to that of $g \\circ F_{\\beta\\delta}(X')$ through two Stein-type steps: an exchangeable-pair identity that handles the non-Gaussian difference from a matched Gaussian $Y$, and a Gaussian interpolation formula that bounds the difference between the two Gaussian bridges $Y$ and $Y'$.","core_discovery":"The central discovery is that two independent $n \\times m$ random matrices $X$ and $X'$ with mean-zero entries and finite absolute third moments have min-max statistics that are close in distribution in a quantitatively controlled way. Specifically, for every $\\beta>0$, $\\delta>0$, and $\\tau>1/(\\beta(1+\\delta))$, the probability that $|\\min_i \\max_j X_{ij} - \\min_i \\max_j X'_{ij}|$ exceeds $2(\\log n/(\\beta\\delta) \\vee \\log m/\\beta)+3\\tau$ is at most $\\varepsilon + C\\beta(1+\\delta)\\tau^{-1}(B_1+B_1'+B_3+\\beta(1+\\delta)(B_2+B_2'))/(1-\\varepsilon)$, where $\\varepsilon=\\sqrt{e^{-\\alpha}(1+\\alpha)}<1$ with $\\alpha=\\beta^2(1+\\delta)^2\\tau^2-1$, and $B_1,B_2,B_3$ are explicit data-dependent quantities. In the Gaussian case the bound simplifies to $\\varepsilon+C\\beta(1+\\delta)\\tau^{-1}B_3/(1-\\varepsilon)$. The result is obtained by smoothing the non-smooth min-max map with a log-sum-exp free-energy function and then comparing expectations through Stein's method with exchangeable pairs and Gaussian interpolation.","pith_inferences":["The paper does not explore, but the same smoothing-plus-Stein strategy could in principle apply to other non-smooth functionals of random matrices, such as the spectral norm or the largest eigenvalue, whenever a soft approximation with controlled derivative sums exists.","The choice of $\\beta$ and $\\delta$ is left open; an optimal tuning might yield sharper effective bounds for specific entry distributions, but the paper does not investigate this.","The argument suggests that the third-moment quantities $B_1$, $B_2$, $B_3$ are the natural measures of discrepancy; one could test empirically whether they are also necessary by constructing distributions with finite second but infinite third moments where the bound should fail."],"forward_implications":["The min-max statistics of any two mean-zero random matrices with finite third moments are close in distribution, with a probability tail that decays at least like $\\tau^{-1}$.","Choosing $\\beta$, $\\delta$, $\\tau$ appropriately recovers the Gaussian quantitative Gordon inequality as the Gaussian special case.","The result holds even when entries within a matrix are dependent, since only moment conditions on the entries are assumed.","The bound depends only on $B_1$, $B_2$, $B_3$, which are in principle computable or estimable from the entry distributions."],"supporting_citations":[{"why":"Supplies the Strassen-type coupling lemma and the smooth approximation of indicator functions that turn the min-max event into a smooth expectation.","marker":"[3]"},{"why":"The original Gordon min-max inequality for Gaussian processes that this paper generalizes to finite third moments.","marker":"[5]"},{"why":"Provides the exchangeable-pair Stein identity used to bound the expectation difference between X and its matched Gaussian Y.","marker":"[8]"},{"why":"Identifies the smoothing function F_{βδ} as the spin-glass free energy, whose derivative sums are controlled in the proof.","marker":"[9]"},{"why":"The Gaussian quantitative Gordon bound that serves as the motivation and as the Gaussian case of the new theorem.","marker":"[10]"},{"why":"Supplies the Gaussian interpolation formula used with [17] to compare the two Gaussian bridges Y and Y'.","marker":"[15]"},{"why":"Co-supplies the Gaussian interpolation formula used in bounding the Gaussian discrepancy term B3.","marker":"[17]"}],"fun_headline_variants":["Min-max discrepancy tamed for non-Gaussian matrices","Finite third moments bound min-max gap","Non-Gaussian min-max: explicit probability bound","Stein's method tightens min-max for heavy tails","Min-max closeness proven with only third moments"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes that the Stein helper function $h$ built from the smoothed min-max $f$ inherits the same derivative-sum bounds as $f$, a step that is asserted rather than proved in the proof of Theorem 2.8.","fun_headline_variants_meta":{"raw":{"variants":["Min-max discrepancy tamed for non-Gaussian matrices","Finite third moments bound min-max gap","Non-Gaussian min-max: explicit probability bound","Stein's method tightens min-max for heavy tails","Min-max closeness proven with only third moments"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000192,"raw_usage":{"total_tokens":1320,"prompt_tokens":894,"completion_tokens":426,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":510,"completion_tokens_details":{"reasoning_tokens":354}},"tokens_in":510,"tokens_out":426,"duration_ms":4288,"temperature":1.0,"reasoning_tokens":354,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T21:45:52.596820+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a concrete non-Gaussian ensemble such as i.i.d. Rademacher entries, compute by simulation the probability that the min-max statistics of two independent copies differ by more than the threshold, and compare it to the bound for several $\\beta$, $\\delta$, $\\tau$; a single violation for valid parameters would refute the theorem.","supporting_citations":[{"cited_title":"Chetverikov, D","cited_arxiv_id":null,"evidence_quote":"Supplies the Strassen-type coupling lemma and the smooth approximation of indicator functions that turn the min-max event into a smooth expectation."},{"cited_title":"Some inequalities for Gaussian processes and applications","cited_arxiv_id":null,"evidence_quote":"The original Gordon min-max inequality for Gaussian processes that this paper generalizes to finite third moments."},{"cited_title":"On Stein’s method for multivariate normal app roximation","cited_arxiv_id":null,"evidence_quote":"Provides the exchangeable-pair Stein identity used to bound the expectation difference between X and its matched Gaussian Y."},{"cited_title":"The sherrington-kirkpatrick model","cited_arxiv_id":null,"evidence_quote":"Identifies the smoothing function F_{βδ} as the spin-glass free energy, whose derivative sums are controlled in the proof."},{"cited_title":"The discrepancy between min– max statistics of Gaussian and Gaussian-subordinated matrices","cited_arxiv_id":null,"evidence_quote":"The Gaussian quantitative Gordon bound that serves as the motivation and as the Gaussian case of the new theorem."},{"cited_title":"Probability in high dimension","cited_arxiv_id":null,"evidence_quote":"Supplies the Gaussian interpolation formula used with [17] to compare the two Gaussian bridges Y and Y'."},{"cited_title":"High-dimensional probability","cited_arxiv_id":null,"evidence_quote":"Co-supplies the Gaussian interpolation formula used in bounding the Gaussian discrepancy term B3."}],"review_version":1}