{
  "source_sha256": "6fc0bf1d87b217c2b7660d0f1d05ab5cf773bce8a2870456c4949c6aca13c298",
  "review_policy": {
    "schema": 1,
    "cutoff_date": "2019-07-20",
    "basis": "received_date",
    "evidence": "Annals journal record https://annals.math.princeton.edu/2019/190-3/p06 lists Received 20 July 2019; arxiv v1 July 1 and reviewed v2 August 29 2019. The received-date cutoff is separate from the later reviewed edition.",
    "status": "recorded",
    "rule": "external-results-available-by-cutoff"
  },
  "nodes": {
    "lem_eig": {
      "outcome": "supported",
      "origin": "curated_source_check",
      "summary": "The signed recursion and both eigenvalue multiplicities check.",
      "source_quote": "\\begin{lemma}\\label{lem_eig}\r\nWe define a sequence of symmetric square matrices iteratively as follows,\r\n$$A_1=\\begin{bmatrix} 0 & 1 \\\\\r\n1 & 0 \\end{bmatrix},~~A_{n}=\\begin{bmatrix}",
      "statement_status": "locally_supported",
      "statement_argument": "Block multiplication gives A_n^2=nI. Symmetry gives a real orthogonal eigendecomposition; trace zero and dimension 2^n force equal multiplicities of +sqrt(n) and -sqrt(n).",
      "explanation_url": "report.pdf",
      "scope": "Check the recursive construction, A_n squared, symmetry, trace and exact multiplicities for every n>=1. Audit this lemma only.",
      "checked_by": "Sol/Fable independent reviews, Codex source checks, and Astra on selected objections; not human grades or formal certificates."
    },
    "lem_key": {
      "outcome": "supported",
      "origin": "curated_source_check",
      "summary": "The maximum-coordinate argument is valid with the standard loopless-graph convention.",
      "source_quote": "\\begin{lemma}\\label{lem_key}\r\nSuppose $H$ is an $m$-vertex undirected graph, and $A$ is a symmetric matrix whose entries are in $\\{-1, 0, 1\\}$ and whose rows and columns are indexe",
      "statement_status": "locally_supported",
      "statement_argument": "For a nonzero eigenvector choose a coordinate of maximum modulus. Nonedge zeros, including the diagonal, and entries bounded by 1 give |lambda| times that modulus <= Delta times the same positive modulus.",
      "explanation_url": "report.pdf",
      "scope": "Check the stated graph and matrix hypotheses, coordinate maximum argument, diagonal convention and conclusion. Audit the supplied proof, not an alternative theorem.",
      "checked_by": "Sol/Fable independent reviews, Codex source checks, and Astra on selected objections; not human grades or formal certificates."
    },
    "thm_main": {
      "outcome": "repair_verified",
      "origin": "curated_source_check",
      "summary": "The lower bound is supported after the trivial n=1 completion. A separate square-sharpness construction is checked; the original citation remains unauthenticated.",
      "source_quote": "\\begin{theorem}\\label{thm_main}\r\nFor every integer $n \\ge 1$, let $H$ be an arbitrary $(2^{n-1}+1)$-vertex induced subgraph of $Q^n$, then \r\n$$\\Delta(H) \\ge \\sqrt{n}.$$\r\nMoreover t",
      "statement_status": "locally_supported",
      "statement_argument": "At n=1, H=K_2. More generally the +sqrt(n) eigenspace and the coordinate subspace of H have dimensions N/2 and N/2+1, hence intersect nontrivially. Restriction yields an eigenvector of A_H with eigenvalue sqrt(n), so the degree lemma applies. For n=k^2, the signed parity product of AND of k ORs on k variables has mean of magnitude 2^(1-n); its majority set has N/2+1 vertices and maximum degree k. Astra checked this separately attributed construction.",
      "explanation_url": "report.pdf",
      "scope": "Check the graph/matrix identification and interlacing deduction including n=1. Separate the cited tightness construction from the proof supplied here.",
      "checked_by": "Sol/Fable independent reviews, Codex source checks, and Astra on selected objections; not human grades or formal certificates.",
      "repair_verified": true,
      "repair_explanation": "At n=1, H=K_2. More generally the +sqrt(n) eigenspace and the coordinate subspace of H have dimensions N/2 and N/2+1, hence intersect nontrivially. Restriction yields an eigenvector of A_H with eigenvalue sqrt(n), so the degree lemma applies. For n=k^2, the signed parity product of AND of k ORs on k variables has mean of magnitude 2^(1-n); its majority set has N/2+1 vertices and maximum degree k. Astra checked this separately attributed construction.",
      "original_hypotheses_preserved": true
    },
    "spectral_sharpness": {
      "outcome": "supported",
      "origin": "curated_source_check",
      "summary": "The spectral bound and its star equality example check for every n>=1.",
      "source_quote": "From the proof, one actually has $\\lambda_1(H) \\ge \\lambda_1(A_H) \\ge \\sqrt{n}$. Since $\\Delta(H) \\ge \\lambda_1(H)$, this result strengthens Theorem \\ref{thm_main}. More interestin",
      "statement_status": "locally_supported",
      "statement_argument": "The absolute-value Rayleigh comparison gives lambda_max(H)>=lambda_max(A_H)>=sqrt(n). All even vertices and one odd vertex induce K_(1,n) plus 2^(n-1)-n isolates, giving largest eigenvalue sqrt(n). This does not give maximum-degree sharpness for n>1.",
      "explanation_url": "report.pdf",
      "scope": "Check the spectral comparison in the remark and the star-plus-isolates equality example for every n. Distinguish this from maximum-degree sharpness.",
      "checked_by": "Sol/Fable independent reviews, Codex source checks, and Astra on selected objections; not human grades or formal certificates."
    },
    "theorem:line:176:10531": {
      "outcome": "supported",
      "origin": "curated_source_check",
      "summary": "The printed conditional deduction and an independent parity/restriction argument check. The full external equivalence is not audited.",
      "source_quote": "\\begin{theorem}\r\nFor every boolean function $f$,\r\n$$s(f) \\ge \\sqrt{\\deg(f)}.$$\r\n\\end{theorem}",
      "statement_status": "locally_supported",
      "statement_argument": "For a degree-d nonconstant function restrict outside a maximal monomial, retaining degree d and not increasing sensitivity. Multiplying signed output by full parity yields an unbalanced partition whose internal edges are sensitive. Apply the cube bound to a majority subset. Constants are immediate. The original cited proof path retains its external-source qualification.",
      "explanation_url": "report.pdf",
      "scope": "Check the deduction of Theorem 1.4 from Theorem 1.1 and the explicitly imported Gotsman-Linial equivalence. State the external-text boundary and handle constant functions. Assess the claimed tightness separately from the supplied deduction.",
      "checked_by": "Sol/Fable independent reviews, Codex source checks, and Astra on selected objections; not human grades or formal certificates."
    },
    "thm_sensitivity": {
      "outcome": "supported",
      "origin": "curated_source_check",
      "summary": "The quartic deduction checks; the exact imported Tal statement was sourced, while its complete proof was not audited.",
      "source_quote": "\\begin{theorem}\\label{thm_sensitivity}\r\nFor every boolean function $f$, \r\n$$bs(f) \\le s(f)^4.$$\r\n\\end{theorem}",
      "statement_status": "locally_supported",
      "statement_argument": "Nonnegativity gives deg(f)<=s(f)^2; Tal gives bs(f)<=deg(f)^2; composing gives bs(f)<=s(f)^4, including constants. This is a verified conditional inference, not a full audit of Tal and its imports.",
      "explanation_url": "report.pdf",
      "scope": "Check the precise deduction of bs(f)<=s(f)^4, the cited degree/block-sensitivity inequalities and all hypotheses. Do not treat absent external proofs as established by this review or as automatic flaws in the paper.",
      "checked_by": "Sol/Fable independent reviews, Codex source checks, and Astra on selected objections; not human grades or formal certificates."
    },
    "lem_cauchy": {
      "outcome": "supported",
      "origin": "curated_source_check",
      "summary": "Standard interlacing statement checked by the elementary dimension/Rayleigh argument.",
      "source_quote": "\\begin{lemma}\\label{lem_cauchy} (Cauchy's Interlace Theorem)\r\nLet $A$ be a symmetric $n \\times n$ matrix, and $B$ be a $m \\times m$ principal submatrix of $A$, for some $m<n$. If t",
      "statement_status": "locally_supported",
      "statement_argument": "Elementary verification from symmetric eigenspaces: the span of the top i eigenvectors of the m-dimensional compression intersects the span of ambient eigenvectors i through n, giving mu_i<=lambda_i. Its bottom m-i+1 eigenvectors intersect the ambient top i+n-m eigenspace, giving mu_i>=lambda_(i+n-m). In both cases the subspace dimensions sum to n+1. This checks the stated interlacing inequalities; it does not authenticate Fisk.",
      "explanation_url": "report.pdf",
      "scope": "Interlacing as stated; external historical attribution is not audited.",
      "checked_by": "Codex direct linear-algebra check."
    },
    "external:gotsman-linial": {
      "outcome": "unresolved",
      "origin": "curated_source_check",
      "summary": "Full cited equivalence and historical version not audited; the needed implication has a separately checked direct argument.",
      "source_quote": "\\bibitem{gotsman-linial}\r\nC.~Gotsman, N.~Linial,\r\n\\newblock{The equivalence of two problems on the cube.}\r\n\\newblock{\\em J. Combin. Theory Ser. A}, {\\bf 61} (1) (1992), 142--146.\r\n\r\n",
      "statement_status": "unresolved",
      "statement_argument": "",
      "explanation_url": "report.pdf",
      "scope": "External-source proof coverage, separate from the checked local deductions.",
      "checked_by": "Codex source audit and explicit panel limits."
    },
    "external:tal": {
      "outcome": "unresolved",
      "origin": "curated_source_check",
      "summary": "The exact Tal statement and pre-cutoff date were sourced; a full external proof audit is outside scope.",
      "source_quote": "\\bibitem{tal}\r\nA.~Tal,\r\n\\newblock{Properties and applications of boolean function composition},\r\n\\newblock{Proceedings of the 4th conference on Innovations in Theoretical Computer Science, ITCS '13}, 441--454.\r\n\r\n",
      "statement_status": "unresolved",
      "statement_argument": "",
      "explanation_url": "report.pdf",
      "scope": "External-source proof coverage, separate from the checked local deductions.",
      "checked_by": "Codex source audit and explicit panel limits."
    }
  }
}
