Angle between column types, row types and a global signed type on a punctured rectangle
ProvedOAI.SignedSweeps.occupied_block_type_angle_exp_allThis is the angle estimate between prescribed column types, prescribed row types and a prescribed global signed type on a rectangle with marked cells removed.
Throughout, acts on the word space , with two-coloured into even and odd letters, by the signed action , where is the sign of the permutation induces between the odd positions of . For and partitions , , the operator is the sum, over the placements of even sites among the sites, of the tensor product of the orthogonal projections onto the -isotypic part of the even sites and the -isotypic part of the odd sites. Its entropy is , and a local type on sites is such a pair of total size . For an enumeration of the sites by blocks with sites each and local types , is the corresponding tensor product of local type projections, with . For a coefficient function on a finite group and a representation , .
Let be finite with nonempty, let be a bijection, and let . Fix an injection of marks into the positions, and let enumerate the remaining positions. Let enumerate these sites by column, with sites in column , and let enumerate them by row, with sites in row , compatibly with and . Let and be local types on these blocks. Let with , having at most rows, and let . Then
The estimate converts the entropies of line types into the global entropy; the removed cells cost only the factor , with no assumption on their arrangement.
Formalization Note The norm is the operator norm on . The upstream statement is reproduced verbatim from BlockAngle.lean.
import Mathlib import Definitions.Def_SignedSweepMoment import Definitions.Def_OAISignedSweepsProofDefs open scoped BigOperators Classical
theorem OAI.SignedSweeps.occupied_block_type_angle_exp_all {A B : Type} [Fintype A] [Fintype B] [Nonempty B]
{q p l n s m R : ℕ} (hq : 0 < q)
(hn : p+l=n) (e : Fin n ≃ A × B) (z : MarkedAssignment l n)
{kH : B → ℕ} {kK : A → ℕ}
(eH : (Σ j, Fin (kH j)) ≃ Fin p)
(hHroute : ∀ x, (eH.symm x).1 = (e (markedInjection hn z x)).2)
(eK : (Σ i, Fin (kK i)) ≃ Fin p)
(hKroute : ∀ x, (eK.symm x).1 = (e (markedInjection hn z x)).1)
(H : ∀ j, PairType (kH j)) (K : ∀ i, PairType (kK i))
(hkH : ∀ j, kH j ≤ s) (hkK : ∀ i, kK i ≤ m)
(hR : max s m+1 ≤ R)
{u v : ℕ} (h : u+v=p) (a : Partition u) (b : Partition v)
(ha : a.1.colLen 0 ≤ q) (hb : b.1.colLen 0 ≤ q) :
‖(blockTypeProjection (C:=Fin q) eH H * blockTypeProjection eK K *
pairTypeProjection h a b (Fin q)).toContinuousLinearMap‖^2 ≤
min 1 (Real.exp (blockEntropy H+blockEntropy K-signedEntropy a b+l+
((Fintype.card A : ℝ)+Fintype.card B)*(2*(q:ℝ)^2*Real.log R))) := by sorry