Benchmark relations, reductions, operators and unitary representations
DefinitionBenchmarkRelationsbenchmark-comparisonsborel-reducibilitydescriptive-set-theoryfunctional-analysis
The shared interfaces use real-sequence difference relations, the full shift, countable products and jumps, actual Cantor probability measures and mutual absolute continuity, and genuine bounded operators and unitary representations on . Reductions require measurability and preservation and reflection on every pair. Operator and representation Borel structures are explicitly induced by pointwise norm convergence. The final six propositions name the exact target comparisons without assuming or proving them.
Definition code
import Mathlib.MeasureTheory.Measure.ProbabilityMeasure
import Mathlib.Analysis.SpecialFunctions.Pow.Real
import Mathlib.Analysis.InnerProductSpace.l2Space
import Mathlib.GroupTheory.FreeGroup.Basic
open MeasureTheory Filter
open scoped Topology
noncomputable section
namespace BenchmarkComparisons
abbrev Sequences := ℕ → ℝ
abbrev Cantor := ℕ → Bool
abbrev FreeTwo := FreeGroup (Fin 2)
abbrev ShiftSpace := FreeTwo → Bool
abbrev Hilbert := lp (fun _ : ℕ => ℂ) 2
abbrev BoundedOperator := Hilbert →L[ℂ] Hilbert
abbrev Unitary := Hilbert ≃ₗᵢ[ℂ] Hilbert
def BorelReduces {X Y : Type*} [MeasurableSpace X] [MeasurableSpace Y]
(E : X → X → Prop) (F : Y → Y → Prop) : Prop :=
∃ f : X → Y, Measurable f ∧ ∀ x y, E x y ↔ F (f x) (f y)
def Incomparable {X Y : Type*} [MeasurableSpace X] [MeasurableSpace Y]
(E : X → X → Prop) (F : Y → Y → Prop) : Prop :=
¬ BorelReduces E F ∧ ¬ BorelReduces F E
def ProductRelation {X : Type*} (E : X → X → Prop) (x y : ℕ → X) : Prop :=
∀ n, E (x n) (y n)
def Jump {X : Type*} (E : X → X → Prop) (x y : ℕ → X) : Prop :=
(∀ i, ∃ j, E (x i) (y j)) ∧ (∀ j, ∃ i, E (x i) (y j))
def EInfinity (x y : ShiftSpace) : Prop :=
∃ g : FreeTwo, ∀ h : FreeTwo, y h = x (g⁻¹ * h)
def LpDifference (p : ℝ) (x y : Sequences) : Prop :=
Summable (fun n => |x n - y n| ^ p)
def NullDifference (x y : Sequences) : Prop :=
Tendsto (fun n => x n - y n) atTop (𝓝 0)
def BoundedDifference (x y : Sequences) : Prop :=
∃ C : ℝ, ∀ n, |x n - y n| ≤ C
def ExtendedLpDifference (p : WithTop ℝ) (x y : Sequences) : Prop :=
match p with
| ⊤ => BoundedDifference x y
| (r : ℝ) => LpDifference r x y
def MeasureEquivalent (μ ν : ProbabilityMeasure Cantor) : Prop :=
(μ : Measure Cantor) ≪ (ν : Measure Cantor) ∧
(ν : Measure Cantor) ≪ (μ : Measure Cantor)
def OperatorEquivalent (T S : BoundedOperator) : Prop :=
∃ U : Unitary, ∀ v : Hilbert, U (T v) = S (U v)
def operatorStrongTopology : TopologicalSpace BoundedOperator :=
TopologicalSpace.induced (fun T : BoundedOperator => fun v : Hilbert => T v) inferInstance
def operatorBorel : MeasurableSpace BoundedOperator :=
@borel BoundedOperator operatorStrongTopology
structure Representation (Γ : Type*) [Group Γ] where
operator : Γ → Unitary
map_one : ∀ v : Hilbert, operator 1 v = v
map_mul : ∀ g h : Γ, ∀ v : Hilbert, operator (g * h) v = operator g (operator h v)
def representationStrongTopology (Γ : Type*) [Group Γ] :
TopologicalSpace (Representation Γ) :=
TopologicalSpace.induced
(fun ρ : Representation Γ => fun g : Γ => fun v : Hilbert => ρ.operator g v) inferInstance
def representationBorel (Γ : Type*) [Group Γ] : MeasurableSpace (Representation Γ) :=
@borel (Representation Γ) (representationStrongTopology Γ)
def RepresentationEquivalent {Γ : Type*} [Group Γ] (ρ σ : Representation Γ) : Prop :=
∃ U : Unitary, ∀ g : Γ, ∀ v : Hilbert, U (ρ.operator g v) = σ.operator g (U v)
def BorelReducesWith {X Y : Type*} (mX : MeasurableSpace X) (mY : MeasurableSpace Y)
(E : X → X → Prop) (F : Y → Y → Prop) : Prop :=
∃ f : X → Y, @Measurable X Y mX mY f ∧ ∀ x y, E x y ↔ F (f x) (f y)
def BorelBireducibleWith {X Y : Type*} (mX : MeasurableSpace X) (mY : MeasurableSpace Y)
(E : X → X → Prop) (F : Y → Y → Prop) : Prop :=
BorelReducesWith mX mY E F ∧ BorelReducesWith mY mX F E
def BorelStrictlyReducesWith {X Y : Type*} (mX : MeasurableSpace X) (mY : MeasurableSpace Y)
(E : X → X → Prop) (F : Y → Y → Prop) : Prop :=
BorelReducesWith mX mY E F ∧ ¬ BorelReducesWith mY mX F E
def CountableComparisons : Prop :=
BorelReduces (ProductRelation EInfinity) NullDifference ∧
∀ p : ℝ, 1 ≤ p → Incomparable EInfinity (LpDifference p)
def ExponentThreshold : Prop :=
∀ p : WithTop ℝ, (1 : WithTop ℝ) ≤ p →
(BorelReduces (ExtendedLpDifference p) MeasureEquivalent ↔ p ≤ (2 : WithTop ℝ))
def NullMeasureObstruction : Prop :=
¬ BorelReduces NullDifference MeasureEquivalent
def JumpObstructions : Prop :=
¬ BorelReduces NullDifference (Jump BoundedDifference) ∧
¬ BorelReduces MeasureEquivalent (Jump BoundedDifference)
def OperatorStrictComparison : Prop :=
BorelStrictlyReducesWith inferInstance operatorBorel MeasureEquivalent OperatorEquivalent
def RepresentationComparisons : Prop :=
BorelBireducibleWith (representationBorel (Multiplicative ℤ)) inferInstance
RepresentationEquivalent MeasureEquivalent ∧
BorelStrictlyReducesWith inferInstance (representationBorel (FreeGroup ℕ))
MeasureEquivalent RepresentationEquivalent ∧
BorelBireducibleWith (representationBorel (FreeGroup ℕ)) operatorBorel
RepresentationEquivalent OperatorEquivalent
end BenchmarkComparisons
Source
Ruijun Wang, Comparisons of benchmark equivalence relations, supplied Benchmark.tex, Section 1; source SHA256 c3c80147a17922e19044935df97b27a2a2a0be5f86665e0869d6a376250d6330; repository designated by the author: https://github.com/RayjunWang/Benchmark. Definitions from Sections 1 and 2.
Human review
Confirmed by the mission captain (proposal self-audit).