OAI/Computability/MaxCut/PCP/TableIteration.lean
63 declarations, 0 contain sorry, 0 declared as axioms
Iteration of the concrete stored-table round, with the original formula's actual initial graph size controlling the number of rounds.
- def
OAI.MaxCutGames.Foundations.PCP.TableIteration.runline 15 - def
OAI.MaxCutGames.Foundations.PCP.TableIteration.runTablesline 19Source documentation
The same data computation on all validated tables; positivity is an invariant, and is erased from the runtime path.
def runTables (H : BaseTable) : Nat → GraphTables.Table → GraphTables.Table := …
- theorem
OAI.MaxCutGames.Foundations.PCP.TableIteration.run_valline 22 - def
OAI.MaxCutGames.Foundations.PCP.TableIteration.iterationCountline 30 - def
OAI.MaxCutGames.Foundations.PCP.TableIteration.outputline 32 - def
OAI.MaxCutGames.Foundations.PCP.TableIteration.outputTableline 34 - theorem
OAI.MaxCutGames.Foundations.PCP.TableIteration.output_valline 37 - def
OAI.MaxCutGames.Foundations.PCP.TableIteration.gapMapline 40 - opaque
OAI.MaxCutGames.Foundations.PCP.TableIteration.degreeDataline 44Source documentation
A checked initialized degree avoids evaluating enormous fixed logarithms while applying polynomial identities. Its bound is part of the stored data.
opaque degreeData : {n : Nat // sizeFactor ≤ 2 ^ n} := … - def
OAI.MaxCutGames.Foundations.PCP.TableIteration.polynomialDegreeline 47 - theorem
OAI.MaxCutGames.Foundations.PCP.TableIteration.sizeFactor_le_powerline 49 - theorem
OAI.MaxCutGames.Foundations.PCP.TableIteration.run_completenessline 52 - theorem
OAI.MaxCutGames.Foundations.PCP.TableIteration.output_completenessline 56 - theorem
OAI.MaxCutGames.Foundations.PCP.TableIteration.gapMap_completenessline 60 - theorem
OAI.MaxCutGames.Foundations.PCP.TableIteration.output_darts_positiveline 66 - theorem
OAI.MaxCutGames.Foundations.PCP.TableIteration.gapMap_nonemptyline 70 - theorem
OAI.MaxCutGames.Foundations.PCP.TableIteration.run_sizeline 73 - theorem
OAI.MaxCutGames.Foundations.PCP.TableIteration.output_sizeline 77 - theorem
OAI.MaxCutGames.Foundations.PCP.TableIteration.initial_size_leline 82 - theorem
OAI.MaxCutGames.Foundations.PCP.TableIteration.output_size_le_inputline 85 - def
OAI.MaxCutGames.Foundations.PCP.TableIteration.tableSizePolynomialline 93 - theorem
OAI.MaxCutGames.Foundations.PCP.TableIteration.tableSizePolynomial_evalline 96 - def
OAI.MaxCutGames.Foundations.PCP.TableIteration.tableBitsPolynomialline 101 - theorem
OAI.MaxCutGames.Foundations.PCP.TableIteration.output_bits_leline 104 - def
OAI.MaxCutGames.Foundations.PCP.TableIteration.formulaBitsPolynomialline 109 - theorem
OAI.MaxCutGames.Foundations.PCP.TableIteration.gapMap_bits_leline 112 - def
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.wordEquivline 131 - def
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.walkEquivline 137 - def
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.addressEquivline 140 - def
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.labelEquivline 147Source documentation
Pull target raw labels back to old port coordinates; this is a bijection.
def labelEquiv (ports : D ≃ E) (t : Nat) : PaddedLabel E t A ≃ PaddedLabel D t A where …
- theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.wordEnd_mapline 159 - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.endpoint_mapline 167 - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.reach_iffline 171 - def
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.ballEquivline 182 - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.wordToBall_mapline 188 - def
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.pullSelectorline 195Source documentation
Use the target selector, including its concrete enumeration choices.
def pullSelector {t : Nat} {v : V} (S : AddressSelector H t v) : AddressSelector G t v where … - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.decode_pullSelectorline 203 - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.decode_transportline 209 - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.opinionAt_pullSelectorline 219 - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.advanceTail_mapline 232 - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.edgeAt_mapline 240 - def
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.pullAcceptsline 255 - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.pathAccepts_mapline 258 - def
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.dartEquivline 284 - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.edgeSatisfied_mapline 287 - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.rejectionCount_mapline 300 - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.satisfiable_iffline 321 - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.pullReverseline 338 - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.base_edgeSatisfied_mapline 344 - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.base_rejectionCount_mapline 354 - theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.uniform_count_gapline 383Source documentation
The generic lazy-graph gap survives a genuine rotation-preserving port change and the target's chosen selectors, for every raw target labeling.
theorem uniform_count_gap [Fintype V] [Fintype D] [Fintype E] [Nonempty V] [Nonempty D] [Fintype A] [Nonempty A] (G : PortGraph V D) (H : PortGraph V E) (ports : (Bool × D) ≃ E) (rotation : ∀ v d, H.rot (v, ports d) = (((lazyGraph G).rot (v,d)).1, ports ((lazyGraph G).rot (v,d)).2)) (lambda : ℝ) (certificate : SpectralCertificate (lazyGraph G) lambda) (accepts : Edge V E → A → A → Bool) (reverse_accepts : ∀ e a b, accepts (H.rot e) b a = accepts e a b) (M : Nat) (hM : 1 ≤ M) (selectors : ∀v, AddressSelector H (2*center (Fintype.card A) M+1) v) (epsilon : ℝ) (hepsilon : 0 ≤ epsilon) …Mentions: OAI.MaxCutGames.Foundations.PCP.accepts
- theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.lazy_table_uniform_count_gapline 423Source documentation
Gap bound for the actual numbered power table built from a lazified port table. The spectral certificate is on the genuine Boolean-flag lazy graph.
theorem lazy_table_uniform_count_gap {vertices d : Nat} (input : PortTables.Table vertices d) (hv : 0 < vertices) (hd : 0 < d) (lambda : ℝ) (certificate : SpectralCertificate (lazyGraph (PortTables.portGraph input)) lambda) (M : Nat) (hM : 1 ≤ M) (epsilon : ℝ) (hepsilon : 0 ≤ epsilon) (hgap : ∀ assignment : Fin vertices → Fin 64, epsilon * (Fintype.card (Fin vertices × Fin (2*d)) : ℝ) ≤ ((PortTables.baseGraph (PreprocessingOverlayTables.lazy input)).rejectionCount assignment : ℝ)) : ∀ labels : PoweringTableSemantics.EncodedLabeling vertices (2*d) (2*center 64 M), (gain 64 M lambda * min epsilon (1 / ((2*center 64 M+1 : Nat) : ℝ))) * ((PoweringTables.table (PreprocessingOverlayTables.lazy input) (2*center 64 M)).darts : ℝ) ≤ ((GenericGraphTables.semantics …Mentions: OAI.MaxCutGames.Foundations.PCP.PortTables.Table, OAI.MaxCutGames.Foundations.PCP.PortTables.portGraph, OAI.MaxCutGames.Foundations.PCP.PortTables.baseGraph, OAI.MaxCutGames.Foundations.PCP.PreprocessingOverlayTables.lazy, OAI.MaxCutGames.Foundations.PCP.PoweringTableSemantics.EncodedLabeling, OAI.MaxCutGames.Foundations.PCP.PoweringTables.table, OAI.MaxCutGames.Foundations.PCP.GenericGraphTables.semantics
- theorem
OAI.MaxCutGames.Foundations.PCP.PoweringPortReindex.lazy_table_uniform_count_gap_atline 463Source documentation
The same exact table bound with a separately stored walk parameter. Its proved equality may be used without evaluating a huge closed alphabet size.
theorem lazy_table_uniform_count_gap_at {vertices d : Nat} (input : PortTables.Table vertices d) (hv : 0 < vertices) (hd : 0 < d) (lambda : ℝ) (certificate : SpectralCertificate (lazyGraph (PortTables.portGraph input)) lambda) (M : Nat) (hM : 1 ≤ M) (n : Nat) (hn : n = 2*center 64 M) (epsilon : ℝ) (hepsilon : 0 ≤ epsilon) (hgap : ∀ assignment : Fin vertices → Fin 64, epsilon * (Fintype.card (Fin vertices × Fin (2*d)) : ℝ) ≤ ((PortTables.baseGraph (PreprocessingOverlayTables.lazy input)).rejectionCount assignment : ℝ)) : ∀ labels : PoweringTableSemantics.EncodedLabeling vertices (2*d) n, (gain 64 M lambda * min epsilon (1 / ((n+1 : Nat) : ℝ))) * …Mentions: OAI.MaxCutGames.Foundations.PCP.PortTables.Table, OAI.MaxCutGames.Foundations.PCP.PortTables.portGraph, OAI.MaxCutGames.Foundations.PCP.PortTables.baseGraph, OAI.MaxCutGames.Foundations.PCP.PreprocessingOverlayTables.lazy, OAI.MaxCutGames.Foundations.PCP.PoweringTableSemantics.EncodedLabeling
- theorem
OAI.MaxCutGames.Foundations.PCP.RoundTableGap.powered_count_gapline 500 - theorem
OAI.MaxCutGames.Foundations.PCP.RoundTableGap.build_count_gapline 531 - theorem
OAI.MaxCutGames.Foundations.PCP.RoundTableGap.step_gapline 546 - theorem
OAI.MaxCutGames.Foundations.PCP.TableGapReduction.run_gapline 570 - theorem
OAI.MaxCutGames.Foundations.PCP.TableGapReduction.output_gapline 575 - theorem
OAI.MaxCutGames.Foundations.PCP.TableGapReduction.output_satisfiable_iffline 583 - theorem
OAI.MaxCutGames.Foundations.PCP.TableGapReduction.gapMap_satisfiable_iffline 593 - theorem
OAI.MaxCutGames.Foundations.PCP.TableGapReduction.output_count_gapline 597 - theorem
OAI.MaxCutGames.Foundations.PCP.TableGapReduction.gapMap_clauseGapline 617 - theorem
OAI.MaxCutGames.Foundations.PCP.TableGapReduction.finalClauseGap_le_oneline 623