{
  "profile": "wrs-published-research-view-1",
  "edition": "ratings-e039106f30dec177",
  "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS",
  "entry": {
    "id": "lanczos-minimized-eigenvalue-iterations-1950",
    "title": "Lanczos: minimized iterations for matrix eigenvalues and operator problems",
    "summary": "Lanczos develops progressive polynomial and minimized-iteration procedures for spectral and resolvent problems, including the symmetric three-term orthogonal recurrence and a two-sided nonsymmetric construction, with numerical vibration examples and proposed integral/differential operator extensions.\n\nEigenvalues help describe how a bridge vibrates, how an electrical network behaves or which modes dominate a physical model. Solving a large spectral problem directly can be expensive. Lanczos uses repeated matrix actions to build a smaller polynomial or orthogonal-basis description, allowing important spectral information to emerge progressively. His minimized iterations address the way ordinary powers can drown out weaker directions. The original numerical examples also checked and repaired orthogonality. Starting vectors can miss directions, and a nonzero two-sided residual can still cause breakdown, so reliable later solvers add safeguards and error analysis. These conditions help readers understand when the method is useful rather than treating it as an automatic solution to every operator problem.",
    "boundary": "Exact October 1950 NBS 45(4), 255-282 paper. Finite-matrix exact-arithmetic recurrence is distinct from floating-point behavior and infinite-operator convergence. The source explicitly discusses missing axes and repeated/defective eigenvalues: one starting chain need not recover the entire eigensystem. Symmetric orthogonal recurrence differs from two-sided nonsymmetric biorthogonalization, whose scalar denominators can vanish with nonzero residual vectors; nonzero initial moments alone do not make it breakdown-free. Later look-ahead theory is separate. The reported bar computations monitored and corrected orthogonality, so short recurrence alone is not a universal stability guarantee. Resonance poles require appropriate resolvent interpretation. Source claims of general Fredholm convergence are not independently certified here; operator domains, invertibility, compactness and boundary conditions require separate justification. Truncated spectral stabilization and coarse discretization are not automatic error certificates. Higher-dimensional input-output applications are proposed future work.",
    "area": "Numerical linear algebra",
    "topic": "Mathematics",
    "year": 1950,
    "lane": "HUMAN",
    "result": "ALGORITHM",
    "package_id": "lanczos-minimized-eigenvalue-iterations-1950",
    "target": {
      "record_id": "urn:uuid:9d68a67a-ab4a-410a-a72b-d2d547a7ceb3",
      "version": 1,
      "content_hash": "sha256:bebdc6f446fcc88b36e81b649c37b54d46aba99a34de34fcd6b99d93d940d359"
    },
    "humans": [
      "Cornelius Lanczos"
    ],
    "authors_complete": true,
    "human_role_note": "Source-listed contributors to this scoped work. Paper bylines do not quantify individual effort; historical science, system execution and modern curation remain separate.",
    "agents": [],
    "related": [
      "hestenes-stiefel-conjugate-gradients-1952",
      "saad-schultz-gmres-1986"
    ],
    "subresults": [],
    "scientific_assessment": {
      "source_review": "SOURCE_CURATED",
      "scientific_review": "NOT_REVIEWED_BY_WRS",
      "formal_replay": "NOT_RUN",
      "provider_attestation": "NONE"
    },
    "sources": [
      {
        "id": "wrs24-lanczos1950-original",
        "title": "Lanczos, An iteration method for the solution of the eigenvalue problem of linear differential and integral operators; exact1950 NBS original",
        "url": "https://nvlpubs.nist.gov/nistpubs/jres/045/jresv45n4p255_A1b.pdf"
      },
      {
        "id": "wrs24-lanczos1950-crossref",
        "title": "NIST-deposited exact1950 Lanczos journal identity",
        "url": "https://api.crossref.org/works/10.6028%2Fjres.045.026"
      },
      {
        "id": "wrs24-parletttaylorliu1985-boundary",
        "title": "Parlett, Taylor and Liu: publisher-deposited look-ahead Lanczos identity and abstract",
        "url": "https://api.crossref.org/works/10.1090%2FS0025-5718-1985-0771034-2"
      }
    ],
    "points": "250000",
    "published_grade": {
      "id": "lanczos-minimized-eigenvalue-iterations-1950",
      "entry_id": "lanczos-minimized-eigenvalue-iterations-1950",
      "band": "S",
      "modifier": null,
      "target": {
        "record_id": "urn:uuid:9d68a67a-ab4a-410a-a72b-d2d547a7ceb3",
        "version": 1,
        "content_hash": "sha256:bebdc6f446fcc88b36e81b649c37b54d46aba99a34de34fcd6b99d93d940d359"
      },
      "what_changed": "Lanczos develops progressive polynomial and minimized-iteration procedures for spectral and resolvent problems, including the symmetric three-term orthogonal recurrence and a two-sided nonsymmetric construction, with numerical vibration examples and proposed integral/differential operator extensions.",
      "scope": "Lanczos develops progressive polynomial and minimized-iteration procedures for spectral and resolvent problems, including the symmetric three-term orthogonal recurrence and a two-sided nonsymmetric construction, with numerical vibration examples and proposed integral/differential operator extensions. Exact October 1950 NBS 45(4), 255-282 paper. Finite-matrix exact-arithmetic recurrence is distinct from floating-point behavior and infinite-operator convergence. The source explicitly discusses missing axes and repeated/defective eigenvalues: one starting chain need not recover the entire eigensystem. Symmetric orthogonal recurrence differs from two-sided nonsymmetric biorthogonalization, whose scalar denominators can vanish with nonzero residual vectors; nonzero initial moments alone do not make it breakdown-free. Later look-ahead theory is separate. The reported bar computations monitored and corrected orthogonality, so short recurrence alone is not a universal stability guarantee. Resonance poles require appropriate resolvent interpretation. Source claims of general Fredholm convergence are not independently certified here; operator domains, invertibility, compactness and boundary conditions require separate justification. Truncated spectral stabilization and coarse discretization are not automatic error certificates. Higher-dimensional input-output applications are proposed future work.",
      "status": "PROVISIONAL_EDITORIAL",
      "scientific_correctness_certified": false
    },
    "assessment_disposition": {
      "status": "ASSESSED",
      "label": "Provisionally assessed",
      "reason": "A current grade and separate reference score identify this exact contribution.",
      "target": {
        "record_id": "urn:uuid:9d68a67a-ab4a-410a-a72b-d2d547a7ceb3",
        "version": 1,
        "content_hash": "sha256:bebdc6f446fcc88b36e81b649c37b54d46aba99a34de34fcd6b99d93d940d359"
      },
      "edition": "ratings-e039106f30dec177"
    }
  },
  "version": {
    "kind": "PUBLIC_RECORD_VIEW",
    "previous": null,
    "published_at": "2026-10-03T00:00:00Z",
    "publisher": {
      "principal_id": "urn:uuid:80116931-785b-40a2-a24b-9d59db0c7e33",
      "role": "WRS_EDITORIAL_PUBLICATION"
    },
    "record": {
      "contributors": [
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "historical_person",
          "name": "Cornelius Lanczos",
          "note": "Archive attribution; detailed role assignment remains source-bound",
          "roles": [
            "formulator"
          ],
          "source": {
            "citation": "Lanczos, An iteration method for the solution of the eigenvalue problem of linear differential and integral operators; exact1950 NBS original",
            "url": "https://nvlpubs.nist.gov/nistpubs/jres/045/jresv45n4p255_A1b.pdf"
          }
        }
      ],
      "evidence": [
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-0",
          "kind": "source_reference",
          "source": {
            "citation": "Lanczos, An iteration method for the solution of the eigenvalue problem of linear differential and integral operators; exact1950 NBS original",
            "url": "https://nvlpubs.nist.gov/nistpubs/jres/045/jresv45n4p255_A1b.pdf"
          }
        },
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-1",
          "kind": "source_reference",
          "source": {
            "citation": "NIST-deposited exact1950 Lanczos journal identity",
            "url": "https://api.crossref.org/works/10.6028%2Fjres.045.026"
          }
        },
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-2",
          "kind": "source_reference",
          "source": {
            "citation": "Parlett, Taylor and Liu: publisher-deposited look-ahead Lanczos identity and abstract",
            "url": "https://api.crossref.org/works/10.1090%2FS0025-5718-1985-0771034-2"
          }
        }
      ],
      "origin": "HUMAN_LEGACY",
      "record_profile": "wrs-local-research-notes-0.1",
      "record_type": "research_note",
      "rrp_version": "0.1",
      "scope": {
        "description": "Exact October 1950 NBS 45(4), 255-282 paper. Finite-matrix exact-arithmetic recurrence is distinct from floating-point behavior and infinite-operator convergence. The source explicitly discusses missing axes and repeated/defective eigenvalues: one starting chain need not recover the entire eigensystem. Symmetric orthogonal recurrence differs from two-sided nonsymmetric biorthogonalization, whose scalar denominators can vanish with nonzero residual vectors; nonzero initial moments alone do not make it breakdown-free. Later look-ahead theory is separate. The reported bar computations monitored and corrected orthogonality, so short recurrence alone is not a universal stability guarantee. Resonance poles require appropriate resolvent interpretation. Source claims of general Fredholm convergence are not independently certified here; operator domains, invertibility, compactness and boundary conditions require separate justification. Truncated spectral stabilization and coarse discretization are not automatic error certificates. Higher-dimensional input-output applications are proposed future work.",
        "domain": "Numerical linear algebra"
      },
      "sources": [
        {
          "citation": "Lanczos, An iteration method for the solution of the eigenvalue problem of linear differential and integral operators; exact1950 NBS original",
          "url": "https://nvlpubs.nist.gov/nistpubs/jres/045/jresv45n4p255_A1b.pdf"
        },
        {
          "citation": "NIST-deposited exact1950 Lanczos journal identity",
          "url": "https://api.crossref.org/works/10.6028%2Fjres.045.026"
        },
        {
          "citation": "Parlett, Taylor and Liu: publisher-deposited look-ahead Lanczos identity and abstract",
          "url": "https://api.crossref.org/works/10.1090%2FS0025-5718-1985-0771034-2"
        }
      ],
      "statement": "# Lanczos: minimized iterations for matrix eigenvalues and operator problems\n\nLanczos develops progressive polynomial and minimized-iteration procedures for spectral and resolvent problems, including the symmetric three-term orthogonal recurrence and a two-sided nonsymmetric construction, with numerical vibration examples and proposed integral/differential operator extensions.\n\nEigenvalues help describe how a bridge vibrates, how an electrical network behaves or which modes dominate a physical model. Solving a large spectral problem directly can be expensive. Lanczos uses repeated matrix actions to build a smaller polynomial or orthogonal-basis description, allowing important spectral information to emerge progressively. His minimized iterations address the way ordinary powers can drown out weaker directions. The original numerical examples also checked and repaired orthogonality. Starting vectors can miss directions, and a nonzero two-sided residual can still cause breakdown, so reliable later solvers add safeguards and error analysis. These conditions help readers understand when the method is useful rather than treating it as an automatic solution to every operator problem.\n\n## Scientific scope\n\nExact October 1950 NBS 45(4), 255-282 paper. Finite-matrix exact-arithmetic recurrence is distinct from floating-point behavior and infinite-operator convergence. The source explicitly discusses missing axes and repeated/defective eigenvalues: one starting chain need not recover the entire eigensystem. Symmetric orthogonal recurrence differs from two-sided nonsymmetric biorthogonalization, whose scalar denominators can vanish with nonzero residual vectors; nonzero initial moments alone do not make it breakdown-free. Later look-ahead theory is separate. The reported bar computations monitored and corrected orthogonality, so short recurrence alone is not a universal stability guarantee. Resonance poles require appropriate resolvent interpretation. Source claims of general Fredholm convergence are not independently certified here; operator domains, invertibility, compactness and boundary conditions require separate justification. Truncated spectral stabilization and coarse discretization are not automatic error certificates. Higher-dimensional input-output applications are proposed future work.\n\n## People and systems\n\nCornelius Lanczos\n\n\n\n## Original sources\n\n- [Lanczos, An iteration method for the solution of the eigenvalue problem of linear differential and integral operators; exact1950 NBS original](https://nvlpubs.nist.gov/nistpubs/jres/045/jresv45n4p255_A1b.pdf)\n- [NIST-deposited exact1950 Lanczos journal identity](https://api.crossref.org/works/10.6028%2Fjres.045.026)\n- [Parlett, Taylor and Liu: publisher-deposited look-ahead Lanczos identity and abstract](https://api.crossref.org/works/10.1090%2FS0025-5718-1985-0771034-2)\n",
      "summary": "Lanczos develops progressive polynomial and minimized-iteration procedures for spectral and resolvent problems, including the symmetric three-term orthogonal recurrence and a two-sided nonsymmetric construction, with numerical vibration examples and proposed integral/differential operator extensions.\n\nEigenvalues help describe how a bridge vibrates, how an electrical network behaves or which modes dominate a physical model. Solving a large spectral problem directly can be expensive. Lanczos uses repeated matrix actions to build a smaller polynomial or orthogonal-basis description, allowing important spectral information to emerge progressively. His minimized iterations address the way ordinary powers can drown out weaker directions. The original numerical examples also checked and repaired orthogonality. Starting vectors can miss directions, and a nonzero two-sided residual can still cause breakdown, so reliable later solvers add safeguards and error analysis. These conditions help readers understand when the method is useful rather than treating it as an automatic solution to every operator problem.",
      "title": "Lanczos: minimized iterations for matrix eigenvalues and operator problems"
    },
    "record_id": "urn:uuid:9d68a67a-ab4a-410a-a72b-d2d547a7ceb3",
    "version": 1,
    "source_ref": {
      "record_id": "urn:uuid:9d68a67a-ab4a-410a-a72b-d2d547a7ceb3",
      "version": 1,
      "content_hash": "sha256:bebdc6f446fcc88b36e81b649c37b54d46aba99a34de34fcd6b99d93d940d359"
    },
    "export_view": "wrs-public-editorial-view-1",
    "public_view_hash": "sha256:6357d86d41db29335dd1b10df90fff5331af783a18342603462ce94ab97613f5"
  },
  "target": {
    "record_id": "urn:uuid:9d68a67a-ab4a-410a-a72b-d2d547a7ceb3",
    "version": 1,
    "content_hash": "sha256:bebdc6f446fcc88b36e81b649c37b54d46aba99a34de34fcd6b99d93d940d359"
  },
  "published_grade": {
    "profile": "wrs-published-grade-1",
    "edition": "2026-10-28.1",
    "id": "lanczos-minimized-eigenvalue-iterations-1950",
    "entry_id": "lanczos-minimized-eigenvalue-iterations-1950",
    "revision": 1,
    "title": "Lanczos: minimized iterations for matrix eigenvalues and operator problems",
    "context": "Numerical linear algebra",
    "track": "Numerical linear algebra",
    "band": "S",
    "modifier": null,
    "status": "PROVISIONAL_EDITORIAL",
    "assessed_on": "2026-10-04",
    "evidence_cutoff": "2026-10-04",
    "achievement_state": "SUPPORTED_SCOPED_RESULT",
    "what_changed": "Lanczos develops progressive polynomial and minimized-iteration procedures for spectral and resolvent problems, including the symmetric three-term orthogonal recurrence and a two-sided nonsymmetric construction, with numerical vibration examples and proposed integral/differential operator extensions.",
    "scope": "Lanczos develops progressive polynomial and minimized-iteration procedures for spectral and resolvent problems, including the symmetric three-term orthogonal recurrence and a two-sided nonsymmetric construction, with numerical vibration examples and proposed integral/differential operator extensions. Exact October 1950 NBS 45(4), 255-282 paper. Finite-matrix exact-arithmetic recurrence is distinct from floating-point behavior and infinite-operator convergence. The source explicitly discusses missing axes and repeated/defective eigenvalues: one starting chain need not recover the entire eigensystem. Symmetric orthogonal recurrence differs from two-sided nonsymmetric biorthogonalization, whose scalar denominators can vanish with nonzero residual vectors; nonzero initial moments alone do not make it breakdown-free. Later look-ahead theory is separate. The reported bar computations monitored and corrected orthogonality, so short recurrence alone is not a universal stability guarantee. Resonance poles require appropriate resolvent interpretation. Source claims of general Fredholm convergence are not independently certified here; operator domains, invertibility, compactness and boundary conditions require separate justification. Truncated spectral stabilization and coarse discretization are not automatic error certificates. Higher-dimensional input-output applications are proposed future work.",
    "evidence_summary": "Source-linked, scoped account of the original work. The published scope states the scientific assumptions and remaining verification limits. WRS has not independently certified the full proof, experimental result or system benchmark.",
    "references": [
      {
        "title": "Lanczos, An iteration method for the solution of the eigenvalue problem of linear differential and integral operators; exact 1950 NBS original",
        "url": "https://nvlpubs.nist.gov/nistpubs/jres/045/jresv45n4p255_A1b.pdf",
        "locator": "Original linked work. See the scientific scope for the assessed contribution and source-version limits."
      },
      {
        "title": "NIST-deposited exact 1950 Lanczos journal identity",
        "url": "https://api.crossref.org/works/10.6028%2Fjres.045.026",
        "locator": "Original linked work. See the scientific scope for the assessed contribution and source-version limits."
      },
      {
        "title": "Parlett, Taylor and Liu: publisher-deposited look-ahead Lanczos identity and abstract",
        "url": "https://api.crossref.org/works/10.1090%2FS0025-5718-1985-0771034-2",
        "locator": "Original linked work. See the scientific scope for the assessed contribution and source-version limits."
      }
    ],
    "roles": [
      {
        "who": "Cornelius Lanczos",
        "did": "Original source-listed scientific author of this scoped work; byline does not determine effort fractions.",
        "credit": "SCIENCE"
      }
    ],
    "target": {
      "record_id": "urn:uuid:9d68a67a-ab4a-410a-a72b-d2d547a7ceb3",
      "version": 1,
      "content_hash": "sha256:bebdc6f446fcc88b36e81b649c37b54d46aba99a34de34fcd6b99d93d940d359"
    },
    "scientific_correctness_certified": false,
    "independent_expert_calibration": false,
    "method": "PROPRIETARY",
    "points": "250000"
  },
  "canonical_agent_submission": false,
  "scientific_verification_awarded": false,
  "reader_guide": null,
  "reading_status": "SOURCE_LINKED_NOTE"
}