{"id":"b743fc6f-03b3-47f6-aaeb-d897d9257773","arxiv_id":"2607.12286","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The unique affine-invariant Kronecker projection minimizes the Hessian-relative condition number iff extreme spectral states share identical tensor marginals; a 2×3 example is the minimal strict separation.","lead":"This paper finds when the nearest Kronecker-structured positive-definite matrix under the affine-invariant metric is also the best structured preconditioner for the Hessian-relative condition number. A smart generalist may care because it supplies certified residuals, error bounds, and a dimension-minimal counterexample that separates two classical matrix goals.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified","rationale":"The single load-bearing geometric hypothesis invoked for uniqueness of the affine-invariant projection holds by an elementary calculation that uses only the mixed-product identity and the standard expression for affine-invariant geodesics; consequently that hypothesis does not constitute a soft spot. No other inconsistency appears among the stated normal equations, residual descent property, spectral-level forcing of optimality, or residual-calibrated bounds. The paper therefore remains correctly UNVERDICTED solely because the full proofs, constructions and interval enclosures are unavailable for audit—the same reason already given by the reader. No verdict adjustment is warranted.","tokens_in":2075,"tokens_out":505,"duration_ms":25480,"concrete_test":"Draw two random SPD factors of sizes 2 and 3, form the corresponding Kronecker matrices P and Q, evaluate the affine-invariant geodesic at t=1/2 by the closed-form formula above, and check that the midpoint equals some A'⊗B' to machine precision (e.g., Frobenius residual <1e-12 after optimal scaling). Separately recompute the condition numbers of the claimed 2×3 separation example; if both checks pass, the geometric foundation and the dimension-minimal strict separation stand.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's flagged weakest assumption (closed geodesic convexity of the Kronecker-SPD family under the affine-invariant metric) is in fact correct and can be verified directly from the abstract's setting. For P = A⊗B and Q = C⊗D the intermediate geodesic point is P^{1/2}(P^{-1/2}QP^{-1/2})^t P^{1/2}; the mixed-product property yields P^{-1/2}QP^{-1/2} = Ã⊗B̃ and therefore (Ã⊗B̃)^t = Ã^t⊗B̃^t, so the whole geodesic remains Kronecker. Closedness follows by continuous recovery of the factors (up to scaling) via partial traces. With uniqueness of the projection thereby secured, the abstract's residual certificates, the iff linking projection optimality to identical tensor marginals of extreme spectral states, the automatic 2×2 optimality, and the explicit 2×3 separation form a coherent first-order optimality package with no internal gap visible from the given statements.","agreement_with_reader":"disagree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript determines the exact relation between nearest structured approximation and best structured preconditioning for Kronecker positive-definite matrices under the affine-invariant Riemannian metric. It asserts that the Kronecker-SPD family is closed and geodesically convex, so every full SPD matrix has a unique affine-invariant projection whose logarithmic residual satisfies partial-trace normal equations and yields Armijo certificates. The central claim is that this projection minimizes the Hessian-relative condition number if and only if the extreme spectral states admit identical tensor marginals; a computable marginal-mismatch residual either vanishes at a condition-optimal point or produces a strict descent direction. Automatic optimality holds for two relative spectral levels and for every 2\times2 Kronecker projection, while an explicit 2\times3 construction is presented as a dimension-minimal strict separation. Residual-calibrated bounds, classical Loewner sandwiches, and interval-safe numerical checks complete the package.","tokens_in":2249,"tokens_out":877,"duration_ms":17206,"significance":"If the claimed equivalence, residual certificates, and dimension-minimal separation hold, the paper supplies a precise geometric dictionary between two standard structured-matrix problems and a practical first-order optimality package for Kronecker preconditioning. The residual-calibrated bounds on attainable condition numbers and the placement of classical diagonal/block Loewner sandwiches in a common certificate language would be useful reference points. The emphasis on validated numerical enclosures, outward-rounded comparisons, and a multistart generic log-factor oracle independent of the partial-trace solver is a methodological strength that supports reproducibility.","major_comments":[{"comment":"The central iff theorem (unique affine-invariant projection is condition-optimal precisely when extreme spectral states have identical tensor marginals) is load-bearing for the paper’s main claim. Only the abstract is available, so the derivation of the partial-trace normal equations, the construction of the marginal-mismatch residual, and the proof that a nonzero residual yields a strict descent direction cannot be inspected. The logical outline is coherent, but soundness of the equivalence cannot be confirmed without the full argument.","section":"Abstract (central result)"},{"comment":"Uniqueness of the projection rests on closed geodesic convexity of the Kronecker-SPD family under the affine-invariant metric. The mixed-product property makes geodesic convexity plausible, and continuous recovery of factors via partial traces makes closedness plausible, yet both statements require formal verification in the manuscript; they underwrite every subsequent certificate and the 2\times3 separation.","section":"Abstract (geodesic convexity claim)"},{"comment":"The explicit 2\times3 construction is asserted to be a dimension-minimal strict separation. Without the concrete matrices, the spectral-state marginals, and the numerical or symbolic verification that the projection is not condition-optimal, the minimality claim and the separation itself remain unchecked.","section":"Abstract (2×3 separation)"}],"minor_comments":[{"comment":"The abstract is unusually dense; a clearer separation of the main theorem statement from the supporting residual bounds, Loewner sandwiches, and numerical protocol would improve readability for a first-pass audience.","section":"Abstract"},{"comment":"Terminology such as “Hessian-relative condition number,” “extreme spectral states,” and “marginal-mismatch residual” is introduced without brief parenthetical definitions; a one-line clarification of each would help non-specialists.","section":"Abstract"}],"recommendation":"uncertain","confidential_remarks":"Only the abstract was supplied for this review; a full-text assessment is required before any definitive accept/revise/reject decision. The abstract itself is carefully written, the geometric outline is internally consistent, and the accompanying stress-test correctly notes that geodesic convexity of the Kronecker family is standard rather than ad-hoc. I see no circularity or free-parameter fitting on the face of the claims. If the full manuscript delivers the promised proofs, residual certificates, and interval-safe checks, the work would be a solid contribution; until then the recommendation must remain uncertain."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"Punchline: from the abstract alone this looks like a clean structural result in math.OC—unique affine-invariant projection onto Kronecker-SPD equals a minimizer of the Hessian-relative condition number iff the extreme spectral states share tensor marginals—plus residual certificates and a dimension-minimal 2×3 separation. Useful inside SPD Riemannian optimization and structured preconditioning if the proofs check out.\n\nWhat is actually new and solid: they treat nearest structured approximation and best structured preconditioning as distinct problems and pin their exact relation under the affine-invariant metric. Closed geodesic convexity of the Kronecker family (which holds by the mixed-product property and partial-trace recovery of factors) gives uniqueness of the projection; the log residual satisfies partial-trace normal equations and supplies Armijo certificates. The marginal-mismatch residual either vanishes at a condition-optimal point or produces strict descent. Two relative spectral levels force optimality; every 2×2 Kronecker projection is condition-optimal; the explicit 2×3 example is therefore minimal. Residual-calibrated bounds bracket the best attainable Kronecker condition number, and classical Loewner sandwiches plus fixed-basis dual obstructions sit in the same certificate language. Deterministic small-matrix checks (including a multistart log-factor oracle independent of the partial-trace solver) and an interval-safe corollary are the right kind of evidence for this area.\n\nSoft spots, in proportion: we have only the abstract, so the central iff, the 2×3 construction, and the numerical enclosures cannot be audited. That is a limitation of this read, not an internal gap. The stress-test confirms the convexity claim that underpins uniqueness; no free parameters or circular fitting appear. Soundness is therefore provisional until the full proofs and enclosures are visible.\n\nWho it is for: people who already care about Kronecker preconditioners or affine-invariant geometry on SPD matrices. It is not a broad audience paper, but it is a serious one. Send it to peer review; a competent referee in Riemannian optimization or structured numerical linear algebra can settle the remaining verification questions quickly.","headline":"Abstract-only Kronecker/affine-invariant paper with a clean iff linking projection to condition-optimal preconditioning; coherent package, worth a referee if the proofs hold.","tokens_in":2860,"tokens_out":524,"would_cite":false,"duration_ms":11867,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["90C25","15A69","65F08"],"pacs":[],"model":"grok-4.5","headline":"The unique affine-invariant projection onto Kronecker positive-definite matrices minimizes the Hessian-relative condition number if and only if extreme spectral states share identical tensor marginals.","keywords":["affine-invariant metric","Kronecker product","structured preconditioning","geodesic convexity","condition number","projection certificates","tensor marginals","partial trace"],"falsifier":"Take the explicit 2-by-3 construction whose extreme spectral states have mismatched tensor marginals; compute its affine-invariant Kronecker projection and a nearby Kronecker matrix obtained by residual-guided descent; if the projected matrix already attains the minimal Hessian-relative condition number, the claimed strict separation fails.","tokens_in":2930,"feed_emoji":"🧮","tokens_out":1018,"duration_ms":22698,"temperature":0.7,"pith_summary":"Nearest structured approximation and best structured preconditioning are usually distinct matrix optimization problems. This paper determines their exact relation for Kronecker positive-definite matrices under the affine-invariant Riemannian metric. Because the Kronecker family is closed and geodesically convex, every full positive-definite matrix possesses a unique affine-invariant projection whose logarithmic residual obeys partial-trace normal equations and supplies certified point and objective errors. The central claim is that this projection is simultaneously a minimizer of the Hessian-relative condition number precisely when the extreme spectral states admit identical tensor marginals. A computable marginal-mismatch residual either vanishes at optimality or yields a strict descent direction; residual-calibrated bounds then sandwich both the best attainable Kronecker condition number and the suboptimality of the pure projection. Every 2-by-2 Kronecker projection is automatically condition-optimal, while an explicit 2-by-3 construction is a dimension-minimal strict separation of the two problems.","feed_headline":"Kronecker projection is best preconditioner only if marginals match","feed_subtitle":"A 2\times3 example separates the two problems; residual certificates decide which case holds and bound the gap.","key_machinery":"The affine-invariant projection onto the Kronecker positive-definite family, characterized by partial-trace normal equations on its logarithmic residual; the same residual both certifies projection error via Armijo-safe bounds and, through its marginal-mismatch component, decides condition-number optimality.","core_discovery":"Under the affine-invariant Riemannian metric the unique projection of a positive-definite matrix onto the Kronecker positive-definite family is a minimizer of the Hessian-relative condition number if and only if the extreme spectral states admit identical tensor marginals. A computable marginal-mismatch residual either vanishes at a condition-optimal projection or produces a strict descent direction; two relative spectral levels always force optimality, every 2-by-2 Kronecker projection is optimal, and a 2-by-3 construction is a dimension-minimal strict separation.","pith_inferences":["The marginal-matching test can be run as a cheap post-processing check after any Kronecker projection routine to decide whether a second, condition-number stage is required.","The same geodesic-convexity-plus-residual pattern may extend to other structured families (certain multilevel or sparsity-pattern classes) that remain geodesically convex under the affine-invariant metric.","Because the normal equations involve only partial traces, distributed or low-rank solvers that never assemble the full matrix become feasible for large-scale Kronecker preconditioning.","The dimension-minimal 2-by-3 separation suggests that tensor-product preconditioners arising in higher-dimensional discretizations will frequently need an explicit condition-number refinement beyond pure projection."],"forward_implications":["Projection and condition-optimal Kronecker preconditioning coincide exactly when the extreme spectral states share identical tensor marginals.","A vanishing marginal-mismatch residual certifies that the projection already solves the condition-number problem; a positive residual supplies an explicit descent direction.","Residual-calibrated bounds rigorously bracket both the best attainable Kronecker condition number and the suboptimality gap of the pure projection.","Every 2-by-2 Kronecker projection is automatically condition-optimal, so separation can first appear only in 2-by-3 dimensions.","The same residual-certificate language recovers classical diagonal and block Loewner sandwiches and fixed-basis primal-dual obstructions."],"fun_headline_variants":["Kronecker projection is condition-optimal only if spectral marginals match","Projection minimizes Hessian condition number iff extreme states share tensor marginals","Marginal-mismatch residual certifies when Kronecker projection is not optimal","Every 2x2 Kronecker projection is condition-optimal; 2x3 is minimal separation","Affine-invariant Kronecker projection equals best preconditioner only on matching marginal"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The Kronecker positive-definite family is closed and geodesically convex under the affine-invariant Riemannian metric, guaranteeing a unique projection and underwriting the optimality equivalence.","fun_headline_variants_meta":{"raw":{"variants":["Kronecker projection is condition-optimal only if spectral marginals match","Projection minimizes Hessian condition number iff extreme states share tensor marginals","Marginal-mismatch residual certifies when Kronecker projection is not optimal","Every 2x2 Kronecker projection is condition-optimal; 2x3 is minimal separation","Affine-invariant Kronecker projection equals best preconditioner only on matching marginals"]},"model":"grok-4.5","effort":"low","cost_usd":0.00627,"raw_usage":{"total_tokens":1651,"prompt_tokens":861,"num_sources_used":0,"completion_tokens":103,"cost_in_usd_ticks":62700000,"prompt_tokens_details":{"text_tokens":861,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":687,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":861,"tokens_out":103,"duration_ms":5032,"temperature":1.0,"reasoning_tokens":687,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-15T00:25:38.932249+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Take the explicit 2-by-3 construction whose extreme spectral states have mismatched tensor marginals; compute its affine-invariant Kronecker projection and a nearby Kronecker matrix obtained by residual-guided descent; if the projected matrix already attains the minimal Hessian-relative condition number, the claimed strict separation fails.","supporting_citations":[],"review_version":1}