LIDT definitions: LDT / SelfImprovement / Theorems / Results / AddInUDiagonalAndDefs / ScalarChain
DefinitionLIDT_MIPStarRE_LDT_SelfImprovement_Theorems_Results_AddInUDiagonalAndDefs_ScalarChainlidt
This bundle defines the selected and averaged scalar Cauchy-Schwarz chains for add-in-. The four transitions move right point projectors to the left and then replace one evaluation point by another, with endpoint identities connecting the chain to matching-mass quantities.
Principal definitions:
addInUSelectedCSChainQ0throughaddInUSelectedCSChainQ4: The selected endpoint and its four successive transport stages.addInUCSChainQ0throughaddInUCSChainQ4: The corresponding averaged chain with the sandwiched family.addInUSelectedCSChainQ0_eq_leftQuantity_averagedSandwiched: Identifies the selected left endpoint with the averaged-sandwich left quantity.addInUSelectedCSChainQ4_eq_rightQuantity: Identifies the selected right endpoint with the right quantity.add_in_u_cs_chain_q0_eq_match_mass: Rewrites the averaged left endpoint as diagonal matching mass.add_in_u_cs_chain_q4_eq_simplified_rhs: Collapses the raw right endpoint using projectivity.
Formalization credit: Sirui Lu, Ruixuan Deng, Yanqiao Zhu, and Zhengfeng Ji (arXiv:2609.19814). Mathematical source: Zhengfeng Ji, Anand Natarajan, Thomas Vidick, John Wright, and Henry Yuen (arXiv:2009.12982).
Definition code
-- Generated from MIPStarRE/LDT/SelfImprovement/Theorems/Results/AddInUDiagonalAndDefs/ScalarChain.lean by skeleton subtraction.
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_AxisParallelLine
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_DiagonalLine
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_Distribution
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_DistributionAvg
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_DistributionUniformSums
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_LinePolynomials
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_LowDegreePolynomial
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_MeasurementLift
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_OpFamily
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_OperatorExpectations
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_PMFAverages
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_PMFUniformAverages
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_ParametersBase
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_ParametersFiniteAnswers
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_QuantumState
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_SqrtBounds
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_SubMeasurementCore
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_SubMeasurementFamilies
import Definitions.Def_LIDT_MIPStarRE_LDT_Basic_TensorPlacement
import Definitions.Def_LIDT_MIPStarRE_LDT_ExpansionHypercubeGraph_Defs_Core
import Definitions.Def_LIDT_MIPStarRE_LDT_GlobalVariance_Defs_Core
import Definitions.Def_LIDT_MIPStarRE_LDT_GlobalVariance_Defs_Families
import Definitions.Def_LIDT_MIPStarRE_LDT_GlobalVariance_Defs_Operators
import Definitions.Def_LIDT_MIPStarRE_LDT_MainInductionStep_Defs
import Definitions.Def_LIDT_MIPStarRE_LDT_MakingMeasurementsProjective_Defs
import Definitions.Def_LIDT_MIPStarRE_LDT_MakingMeasurementsProjective_NaimarkCore
import Definitions.Def_LIDT_MIPStarRE_LDT_MakingMeasurementsProjective_Projectivization
import Definitions.Def_LIDT_MIPStarRE_LDT_MakingMeasurementsProjective_Statements
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_BipartiteSelfConsistency_Completion
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_BipartiteSelfConsistency_Core
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_BipartiteSelfConsistency_Local
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_CauchySchwarz
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_ComparisonCore
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_Defs
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_DistanceBounds
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_PolynomialAgreement
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_Polynomials
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_SelfConsistency_DataProcessing
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_SelfConsistency_Extensions
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_SwitchSandwichMain_Completeness
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_SwitchSandwichPrep_ApproxDelta
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_SwitchSandwichPrep_Core
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_SwitchSandwichPrep_InnerProduct
import Definitions.Def_LIDT_MIPStarRE_LDT_Preliminaries_Triangles_Core
import Definitions.Def_LIDT_MIPStarRE_LDT_SelfImprovement_Defs
import Definitions.Def_LIDT_MIPStarRE_LDT_SelfImprovement_Theorems_Results_AddInUDiagonalAndDefs_Selection
import Definitions.Def_LIDT_MIPStarRE_LDT_SelfImprovement_Theorems_Results_CommonHelpers
import Definitions.Def_LIDT_MIPStarRE_LDT_SelfImprovement_Theorems_Statements
import Definitions.Def_LIDT_MIPStarRE_LDT_SelfImprovement_Theorems_Thresholds_Final
import Definitions.Def_LIDT_MIPStarRE_LDT_SelfImprovement_Theorems_Thresholds_Helper
import Definitions.Def_LIDT_MIPStarRE_LDT_Tactic_QuantumNonneg
import Definitions.Def_LIDT_MIPStarRE_LDT_Test_Defs
import Definitions.Def_LIDT_MIPStarRE_LDT_Test_StrategyCore
import Definitions.Def_LIDT_MIPStarRE_LDT_Test_StrategyFailures
import Definitions.Def_LIDT_MIPStarRE_LDT_Test_StrategyRole_Algebra
import Definitions.Def_LIDT_MIPStarRE_LDT_Test_StrategyRole_Core
import Definitions.Def_LIDT_MIPStarRE_Quantum_FiniteMatrix_Basic
import Definitions.Def_LIDT_MIPStarRE_Quantum_FiniteMatrix_BlockDiagonal
import Definitions.Def_LIDT_MIPStarRE_Quantum_FiniteMatrix_NormalizedTrace
import Definitions.Def_LIDT_MIPStarRE_Quantum_FiniteMatrix_Order
import Definitions.Def_LIDT_MIPStarRE_Quantum_FiniteMatrix_TracePairing
import Definitions.Def_LIDT_MIPStarRE_Quantum_Measurement
import Mathlib
import Mathlib.Algebra.MvPolynomial.CommRing
import Mathlib.Algebra.MvPolynomial.SchwartzZippel
import Mathlib.Algebra.Polynomial.Roots
import Mathlib.Analysis.MeanInequalitiesPow
import Mathlib.Analysis.SpecialFunctions.Pow.Real
import Mathlib.Probability.ProbabilityMassFunction.Integrals
import Mathlib.Probability.ProbabilityMassFunction.Monad
import Mathlib.RingTheory.MvPolynomial.Basic
namespace LIDT
namespace MIPStarRE.LDT.Preliminaries
end MIPStarRE.LDT.Preliminaries
/-!
# Scalar chain for the diagonal add-in-u transfer
This module contains the selected and diagonal Q₀--Q₄ scalar chains used for the
projection-simplified diagonal add-in-`u` transfer, together with the endpoint
identifications needed by the helper strong-self-consistency proof.
## References
- `references/ldt-paper/self_improvement.tex` lines 247--252
- `blueprint/src/chapter/ch07_self_improvement.tex`
-/
@[expose] public section
namespace MIPStarRE.LDT.SelfImprovement
open MIPStarRE.LDT
open MIPStarRE.LDT.ExpansionHypercubeGraph
open MIPStarRE.LDT.GlobalVariance
open MIPStarRE.LDT.MakingMeasurementsProjective
open scoped BigOperators MatrixOrder _root_.Matrix ComplexOrder
variable {ι : Type*} [Fintype ι] [DecidableEq ι]
/-! ### Scalar chain for the projection-simplified diagonal add-in-u transfer -/
/-- Strong self-consistency for the point measurement, pulled back to the second
coordinate of the independent `(u, v)` average used by the add-in-`u` scalar
chain.
This is the distributional self-consistency input for the `A^v_{h(v)}` moves in
`self_improvement.tex`, lines 255--297: the point measurement sampled at `v`
has the same `2δ` left/right state-dependent distance after the product average
over `(u, v)`. -/
lemma addInU_pointMeasurement_snd_selfConsistency
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(delta : Error)
(hssc : BipartiteSSCRel strategy.state (uniformDistribution (Point params))
(IdxProjMeas.toIdxSubMeas strategy.pointMeasurement) delta) :
SDDRel strategy.state (uniformDistribution (Point params × Point params))
(IdxSubMeas.liftLeft
(fun uv : Point params × Point params =>
(strategy.pointMeasurement uv.2).toSubMeas))
(IdxSubMeas.liftRight
(fun uv : Point params × Point params =>
(strategy.pointMeasurement uv.2).toSubMeas))
(2 * delta) := by
classical
have hssc_pair :
BipartiteSSCRel strategy.state
(uniformDistribution (Point params × Point params))
(fun uv : Point params × Point params =>
(strategy.pointMeasurement uv.2).toSubMeas)
delta := by
rcases hssc with ⟨hssc⟩
constructor
calc
avgOver (uniformDistribution (Point params × Point params))
(fun uv : Point params × Point params =>
qBipartiteSSCDefect strategy.state
((strategy.pointMeasurement uv.2).toSubMeas))
=
avgOver (uniformDistribution (Point params))
(fun v : Point params =>
qBipartiteSSCDefect strategy.state
((strategy.pointMeasurement v).toSubMeas)) := by
exact avgOver_uniform_snd
(α := Point params) (β := Point params)
(fun v : Point params =>
qBipartiteSSCDefect strategy.state
((strategy.pointMeasurement v).toSubMeas))
_ ≤ delta := by
simpa [bipartiteSSCError, IdxProjMeas.toIdxSubMeas] using hssc
have hraw :=
Preliminaries.twoNotionsOfSelfConsistencyAfterEvaluation
strategy.state strategy.permInvState
(uniformDistribution (Point params × Point params))
(fun uv : Point params × Point params =>
(strategy.pointMeasurement uv.2).toSubMeas)
delta
(fun _uv (a : Fq params) => a)
hssc_pair
simpa using hraw
/-- The grouped tensor mass over a fiber `h(v)=a` is a contraction.
This is the submeasurement bound used inside the first Cauchy--Schwarz square
root in `self_improvement.tex`, lines 267--272: after grouping by the value
`a = h(v)`, the selected operators
`H^u_h ⊗ T_h` are dominated by the total mass of the sandwiched polynomial
submeasurement at `u`, hence by `I`. -/
lemma addInU_filtered_sandwiched_tensor_sum_le_one
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(T : SubMeas (Polynomial params) ι)
(u v : Point params) (a : Fq params) :
∑ h ∈ Finset.univ.filter (fun h : Polynomial params => h v = a),
opTensor ((sandwichedPolynomialSubMeasAt params strategy T u).outcome h)
(T.outcome h) ≤
(1 : MIPStarRE.Quantum.Op (ι × ι)) := by
classical
exact SubMeas.opTensor_sum_filter_le_one
(sandwichedPolynomialSubMeasAt params strategy T u)
T
(fun h : Polynomial params => h v = a)
/-! ### Selection-parametrized add-in-u scalar chain
The paper proves `lem:add-in-u` for an arbitrary outcome family `M` and
selection `S_u ⊆ 𝒪 × polyfunc`. The diagonal helper strong-self-consistency
application below is one specialization of this statement. The following
definitions record the same five scalar quantities before specializing to the
diagonal case, so that the off-diagonal point-consistency selection can reuse
the Cauchy--Schwarz chain rather than restating the transfer hypothesis. -/
/-- The selected-chain left endpoint `Q₀`.
For a selected pair `(o, h) ∈ S_u`, this is the expectation of
`M^u_o ⊗ H^v_h`, where `H^v_h = A^v_{h(v)} T_h A^v_{h(v)}`. -/
noncomputable def addInUSelectedCSChainQ0
{Outcome : Type*} [Fintype Outcome]
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(M : IdxSubMeas (Point params) Outcome ι)
(T : SubMeas (Polynomial params) ι)
(S : AddInUSelection params Outcome) : Error :=
avgOver (uniformDistribution (Point params × Point params)) (fun uv =>
∑ ah ∈ addInUSelectionPairs params S uv.1,
ev strategy.state
(opTensor ((M uv.1).outcome ah.1)
((sandwichedPolynomialSubMeasAt params strategy T uv.2).outcome ah.2)))
/-- The selected-chain scalar `Q₁`, after moving the right point projector
`A^v_{h(v)}` to the left tensor factor once. -/
noncomputable def addInUSelectedCSChainQ1
{Outcome : Type*} [Fintype Outcome]
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(M : IdxSubMeas (Point params) Outcome ι)
(T : SubMeas (Polynomial params) ι)
(S : AddInUSelection params Outcome) : Error :=
avgOver (uniformDistribution (Point params × Point params)) (fun uv =>
∑ ah ∈ addInUSelectionPairs params S uv.1,
let Av := pointConditionedOutcomeOperatorAtPolynomial params strategy ah.2 uv.2
ev strategy.state
(opTensor (Av * (M uv.1).outcome ah.1) (T.outcome ah.2 * Av)))
/-- The selected-chain scalar `Q₂`, after moving both right point projectors to
the left tensor factor. -/
noncomputable def addInUSelectedCSChainQ2
{Outcome : Type*} [Fintype Outcome]
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(M : IdxSubMeas (Point params) Outcome ι)
(T : SubMeas (Polynomial params) ι)
(S : AddInUSelection params Outcome) : Error :=
avgOver (uniformDistribution (Point params × Point params)) (fun uv =>
∑ ah ∈ addInUSelectionPairs params S uv.1,
let Av := pointConditionedOutcomeOperatorAtPolynomial params strategy ah.2 uv.2
ev strategy.state
(opTensor (Av * (M uv.1).outcome ah.1 * Av) (T.outcome ah.2)))
/-- The selected-chain scalar `Q₃`, after replacing the first point projector
at `v` by the corresponding point projector at `u`. -/
noncomputable def addInUSelectedCSChainQ3
{Outcome : Type*} [Fintype Outcome]
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(M : IdxSubMeas (Point params) Outcome ι)
(T : SubMeas (Polynomial params) ι)
(S : AddInUSelection params Outcome) : Error :=
avgOver (uniformDistribution (Point params × Point params)) (fun uv =>
∑ ah ∈ addInUSelectionPairs params S uv.1,
let Au := pointConditionedOutcomeOperatorAtPolynomial params strategy ah.2 uv.1
let Av := pointConditionedOutcomeOperatorAtPolynomial params strategy ah.2 uv.2
ev strategy.state
(opTensor (Au * (M uv.1).outcome ah.1 * Av) (T.outcome ah.2)))
/-- The selected-chain scalar `Q₄`, after replacing both point projectors at
`v` by the corresponding point projectors at `u`. -/
noncomputable def addInUSelectedCSChainQ4
{Outcome : Type*} [Fintype Outcome]
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(M : IdxSubMeas (Point params) Outcome ι)
(T : SubMeas (Polynomial params) ι)
(S : AddInUSelection params Outcome) : Error :=
avgOver (uniformDistribution (Point params × Point params)) (fun uv =>
∑ ah ∈ addInUSelectionPairs params S uv.1,
let Au := pointConditionedOutcomeOperatorAtPolynomial params strategy ah.2 uv.1
ev strategy.state
(opTensor (Au * (M uv.1).outcome ah.1 * Au) (T.outcome ah.2)))
/-- The selected-chain endpoint `Q₀` is the generic add-in-u left quantity when
the second measurement is the averaged sandwiched polynomial submeasurement. -/
theorem addInUSelectedCSChainQ0_eq_leftQuantity_averagedSandwiched
{Outcome : Type*} [Fintype Outcome]
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(M : IdxSubMeas (Point params) Outcome ι)
(T : SubMeas (Polynomial params) ι)
(S : AddInUSelection params Outcome) :
addInULeftQuantity params strategy M
(averagedSandwichedPolynomialSubMeas params strategy T)
S =
addInUSelectedCSChainQ0 params strategy M T S := by
classical
change avgOver (uniformDistribution (Point params)) (fun u =>
ev strategy.state
(addInULeftOperatorAtPoint params strategy M
(averagedSandwichedPolynomialSubMeas params strategy T) S u)) =
addInUSelectedCSChainQ0 params strategy M T S
unfold addInULeftOperatorAtPoint addInUSelectedCSChainQ0
rw [avgOver_uniform_prod (α := Point params) (β := Point params)
(f := fun u v =>
∑ ah ∈ addInUSelectionPairs params S u,
ev strategy.state
(opTensor ((M u).outcome ah.1)
((sandwichedPolynomialSubMeasAt params strategy T v).outcome ah.2)))]
refine avgOver_congr (uniformDistribution (Point params)) _ _ ?_
intro u
calc
ev strategy.state
(∑ ah ∈ addInUSelectionPairs params S u,
opTensor ((M u).outcome ah.1)
((averagedSandwichedPolynomialSubMeas params strategy T).outcome ah.2))
=
∑ ah ∈ addInUSelectionPairs params S u,
ev strategy.state
(opTensor ((M u).outcome ah.1)
((averagedSandwichedPolynomialSubMeas params strategy T).outcome ah.2)) := by
rw [ev_finset_sum]
_ =
∑ ah ∈ addInUSelectionPairs params S u,
avgOver (uniformDistribution (Point params)) (fun v =>
ev strategy.state
(opTensor ((M u).outcome ah.1)
((sandwichedPolynomialSubMeasAt params strategy T v).outcome ah.2))) := by
refine Finset.sum_congr rfl ?_
intro ah _
simpa [averagedSandwichedPolynomialSubMeas, sandwichedPolynomialSubMeasAt] using
ev_opTensor_averageOperatorOverDistribution_right strategy.state
(uniformDistribution (Point params))
((M u).outcome ah.1)
(fun v => (sandwichedPolynomialSubMeasAt params strategy T v).outcome ah.2)
_ =
avgOver (uniformDistribution (Point params)) (fun v =>
∑ ah ∈ addInUSelectionPairs params S u,
ev strategy.state
(opTensor ((M u).outcome ah.1)
((sandwichedPolynomialSubMeasAt params strategy T v).outcome ah.2))) := by
exact (MIPStarRE.LDT.avgOver_finset_sum (uniformDistribution (Point params))
(addInUSelectionPairs params S u)
(fun v ah =>
ev strategy.state
(opTensor ((M u).outcome ah.1)
((sandwichedPolynomialSubMeasAt params strategy T v).outcome ah.2)))).symm
/-- The selected-chain endpoint `Q₄` is the generic add-in-u right quantity. -/
theorem addInUSelectedCSChainQ4_eq_rightQuantity
{Outcome : Type*} [Fintype Outcome]
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(M : IdxSubMeas (Point params) Outcome ι)
(T : SubMeas (Polynomial params) ι)
(S : AddInUSelection params Outcome) :
addInURightQuantity params strategy M T S =
addInUSelectedCSChainQ4 params strategy M T S := by
classical
change avgOver (uniformDistribution (Point params)) (fun u =>
ev strategy.state
(addInURightOperatorAtPoint params strategy M T S u)) =
addInUSelectedCSChainQ4 params strategy M T S
unfold addInURightOperatorAtPoint addInUSelectedCSChainQ4
rw [avgOver_uniform_prod (α := Point params) (β := Point params)
(f := fun u _ =>
∑ ah ∈ addInUSelectionPairs params S u,
let Au := pointConditionedOutcomeOperatorAtPolynomial params strategy ah.2 u
ev strategy.state
(opTensor (Au * (M u).outcome ah.1 * Au) (T.outcome ah.2)))]
refine avgOver_congr (uniformDistribution (Point params)) _ _ ?_
intro u
rw [ev_finset_sum]
exact (avgOver_uniform_const
(∑ ah ∈ addInUSelectionPairs params S u,
let Au := pointConditionedOutcomeOperatorAtPolynomial params strategy ah.2 u
ev strategy.state
(opTensor (Au * (M u).outcome ah.1 * Au) (T.outcome ah.2)))).symm
/-- The expanded left endpoint `Q₀` of the four-step scalar chain in
`self_improvement.tex`, lines 247--252, after setting `M^u = H^u` and averaging
the second tensor factor `H = E_v H^v`. -/
noncomputable def addInUCSChainQ0
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(T : SubMeas (Polynomial params) ι) : Error :=
avgOver (uniformDistribution (Point params × Point params)) (fun uv =>
∑ h : Polynomial params,
ev strategy.state
(opTensor ((sandwichedPolynomialSubMeasAt params strategy T uv.1).outcome h)
((sandwichedPolynomialSubMeasAt params strategy T uv.2).outcome h)))
/-- The scalar `Q₁` obtained from `Q₀` by moving the right point projection
`A^v_{h(v)}` to the left tensor factor; this is the target of
`eq:move-one`. -/
noncomputable def addInUCSChainQ1
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(T : SubMeas (Polynomial params) ι) : Error :=
avgOver (uniformDistribution (Point params × Point params)) (fun uv =>
∑ h : Polynomial params,
let Av := pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.2
ev strategy.state
(opTensor (Av * (sandwichedPolynomialSubMeasAt params strategy T uv.1).outcome h)
(T.outcome h * Av)))
/-- The scalar `Q₂` obtained from `Q₁` by moving the second right point
projection to the left tensor factor; this is the target of `eq:move-another`. -/
noncomputable def addInUCSChainQ2
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(T : SubMeas (Polynomial params) ι) : Error :=
avgOver (uniformDistribution (Point params × Point params)) (fun uv =>
∑ h : Polynomial params,
let Av := pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.2
ev strategy.state
(opTensor (Av * (sandwichedPolynomialSubMeasAt params strategy T uv.1).outcome h *
Av)
(T.outcome h)))
/-- The scalar `Q₃` obtained from `Q₂` by replacing the first point projection
`A^v_{h(v)}` by `A^u_{h(u)}`; this is the target of `eq:change-one`. -/
noncomputable def addInUCSChainQ3
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(T : SubMeas (Polynomial params) ι) : Error :=
avgOver (uniformDistribution (Point params × Point params)) (fun uv =>
∑ h : Polynomial params,
let Au := pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.1
let Av := pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.2
ev strategy.state
(opTensor (Au * (sandwichedPolynomialSubMeasAt params strategy T uv.1).outcome h *
Av)
(T.outcome h)))
/-- The scalar `Q₄` obtained from `Q₃` by replacing the second point projection
`A^v_{h(v)}` by `A^u_{h(u)}`; after the projection collapse, this is the
projection-simplified right endpoint of the diagonal add-in-u transfer. -/
noncomputable def addInUCSChainQ4
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(T : SubMeas (Polynomial params) ι) : Error :=
avgOver (uniformDistribution (Point params × Point params)) (fun uv =>
∑ h : Polynomial params,
let Au := pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.1
ev strategy.state
(opTensor (Au * (sandwichedPolynomialSubMeasAt params strategy T uv.1).outcome h *
Au)
(T.outcome h)))
/-- The expanded chain endpoint `Q₀` is the existing diagonal match-mass left
side used by `selfConsistencyDiagonalAddInU_of_simplifiedTransfer`. -/
lemma add_in_u_cs_chain_q0_eq_match_mass
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(T : SubMeas (Polynomial params) ι) :
qBipartiteMatchMass strategy.state
(averagedSandwichedPolynomialSubMeas params strategy T)
(averagedSandwichedPolynomialSubMeas params strategy T) =
addInUCSChainQ0 params strategy T := by
classical
calc
qBipartiteMatchMass strategy.state
(averagedSandwichedPolynomialSubMeas params strategy T)
(averagedSandwichedPolynomialSubMeas params strategy T)
=
∑ h : Polynomial params,
avgOver (uniformDistribution (Point params)) (fun u =>
avgOver (uniformDistribution (Point params)) (fun v =>
ev strategy.state
(opTensor ((sandwichedPolynomialSubMeasAt params strategy T u).outcome h)
((sandwichedPolynomialSubMeasAt params strategy T v).outcome h)))) := by
unfold qBipartiteMatchMass averagedSandwichedPolynomialSubMeas
refine Finset.sum_congr rfl ?_
intro h _
rw [ev_opTensor_averageOperatorOverDistribution_left]
refine avgOver_congr _ _ _ ?_
intro u
simpa [sandwichedPolynomialSubMeasAt] using
ev_opTensor_averageOperatorOverDistribution_right strategy.state
(uniformDistribution (Point params))
((sandwichedPolynomialSubMeasAt params strategy T u).outcome h)
(fun v => (sandwichedPolynomialSubMeasAt params strategy T v).outcome h)
_ = addInUCSChainQ0 params strategy T := by
symm
unfold addInUCSChainQ0
rw [avgOver_uniform_prod (α := Point params) (β := Point params)
(f := fun u v =>
∑ h : Polynomial params,
ev strategy.state
(opTensor ((sandwichedPolynomialSubMeasAt params strategy T u).outcome h)
((sandwichedPolynomialSubMeasAt params strategy T v).outcome h)))]
calc
avgOver (uniformDistribution (Point params)) (fun u =>
avgOver (uniformDistribution (Point params)) (fun v =>
∑ h : Polynomial params,
ev strategy.state
(opTensor
((sandwichedPolynomialSubMeasAt params strategy T u).outcome h)
((sandwichedPolynomialSubMeasAt params strategy T v).outcome h))))
=
avgOver (uniformDistribution (Point params)) (fun u =>
∑ h : Polynomial params,
avgOver (uniformDistribution (Point params)) (fun v =>
ev strategy.state
(opTensor
((sandwichedPolynomialSubMeasAt params strategy T u).outcome h)
((sandwichedPolynomialSubMeasAt params strategy T v).outcome h)))) := by
refine avgOver_congr _ _ _ ?_
intro u
rw [avgOver_sum]
_ =
∑ h : Polynomial params,
avgOver (uniformDistribution (Point params)) (fun u =>
avgOver (uniformDistribution (Point params)) (fun v =>
ev strategy.state
(opTensor
((sandwichedPolynomialSubMeasAt params strategy T u).outcome h)
((sandwichedPolynomialSubMeasAt params strategy T v).outcome h)))) := by
rw [avgOver_sum]
/-- The raw chain endpoint `Q₄` collapses to the projection-simplified scalar
right side used by `selfConsistencyDiagonalAddInU_of_simplifiedTransfer`. -/
lemma add_in_u_cs_chain_q4_eq_simplified_rhs
(params : Parameters) [FieldModel params.q]
(strategy : SymStrat params ι)
(T : SubMeas (Polynomial params) ι) :
addInUCSChainQ4 params strategy T =
avgOver (uniformDistribution (Point params)) (fun u =>
∑ h : Polynomial params,
ev strategy.state
(opTensor ((sandwichedPolynomialSubMeasAt params strategy T u).outcome h)
(T.outcome h))) := by
classical
calc
addInUCSChainQ4 params strategy T =
avgOver (uniformDistribution (Point params × Point params)) (fun uv =>
∑ h : Polynomial params,
ev strategy.state
(opTensor ((sandwichedPolynomialSubMeasAt params strategy T uv.1).outcome h)
(T.outcome h))) := by
unfold addInUCSChainQ4
refine avgOver_congr _ _ _ ?_
intro uv
refine Finset.sum_congr rfl ?_
intro h _
have hproj :
pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.1 *
pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.1 =
pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.1 := by
simpa [pointConditionedOutcomeOperatorAtPolynomial] using
(strategy.pointMeasurement uv.1).proj (h uv.1)
have hcollapse :
pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.1 *
(sandwichedPolynomialSubMeasAt params strategy T uv.1).outcome h *
pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.1 =
(sandwichedPolynomialSubMeasAt params strategy T uv.1).outcome h := by
change
pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.1 *
(pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.1 *
T.outcome h *
pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.1) *
pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.1 =
pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.1 *
T.outcome h *
pointConditionedOutcomeOperatorAtPolynomial params strategy h uv.1
exact proj_outer_sandwich_eq _ _ hproj
simp [hcollapse]
_ =
avgOver (uniformDistribution (Point params)) (fun u =>
∑ h : Polynomial params,
ev strategy.state
(opTensor ((sandwichedPolynomialSubMeasAt params strategy T u).outcome h)
(T.outcome h))) := by
exact avgOver_uniform_fst (α := Point params) (β := Point params)
(fun u =>
∑ h : Polynomial params,
ev strategy.state
(opTensor ((sandwichedPolynomialSubMeasAt params strategy T u).outcome h)
(T.outcome h)))
end MIPStarRE.LDT.SelfImprovement
end
end LIDT
Source