InformationTheory

InformationTheory.Shannon.BroadcastChannel.Achievability.Setup

source

Broadcast channel — superposition achievability setup and infrastructure #

Cover–Thomas superposition coding. The structural setup: the per-coordinate joint distribution and its i.i.d. ambient measure, the auxiliary-variable informations, the two-tier (cloud / satellite) random codebook, the i.i.d. coordinate facts and positivity of the BC ambient law, the (U, X) marginal factorization, typical-set relabeling invariance, the two exponential ingredients of the covering bound, the conditional-slice satellite typicality bound, and the two-tier decoders assembling the broadcast code.

Per-coordinate joint distribution #

noncomputable def

InformationTheory.Shannon.BroadcastChannel.bcJointDistribution

source
{U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [MeasurableSpace U] [MeasurableSpace α] [MeasurableSpace β₁] [MeasurableSpace β₂] (pU : MeasureTheory.Measure U) (K : ProbabilityTheory.Kernel U α) (W : BCChannel α β₁ β₂) :
MeasureTheory.Measure (U × α × β₁ × β₂)

The per-coordinate broadcast joint law on U × α × β₁ × β₂: the compProd chain pU → K → W (U ∼ pU, X ∣ U ∼ K, (Y₁, Y₂) ∣ X ∼ W), reshaped from the left-nested (U × α) × (β₁ × β₂) to the right-nested quadruple.

Equations
  • One or more equations did not get rendered due to their size.
Instances For
    Used by
      instance

      InformationTheory.Shannon.BroadcastChannel.bcJointDistribution.instIsProbabilityMeasure

      source
      Used by

        I.i.d. ambient measure on ℕ → U × α × β₁ × β₂ #

        noncomputable def

        InformationTheory.Shannon.BroadcastChannel.bcAmbientMeasure

        source
        {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [MeasurableSpace U] [MeasurableSpace α] [MeasurableSpace β₁] [MeasurableSpace β₂] (pU : MeasureTheory.Measure U) (K : ProbabilityTheory.Kernel U α) (W : BCChannel α β₁ β₂) :
        MeasureTheory.Measure (U × α × β₁ × β₂)

        The i.i.d. broadcast ambient measure: Measure.infinitePi (fun _ ↦ bcJointDistribution pU K W).

        Equations
        Instances For
          Used by
            instance

            InformationTheory.Shannon.BroadcastChannel.bcAmbientMeasure.instIsProbabilityMeasure

            source
            Used by
              def

              InformationTheory.Shannon.BroadcastChannel.bcUs

              source
              {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} :
              (U × α × β₁ × β₂)U

              The cloud coordinate ω ↦ (ω i).1.

              Equations
              Instances For
                Used by
                  def

                  InformationTheory.Shannon.BroadcastChannel.bcXs

                  source
                  {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} :
                  (U × α × β₁ × β₂)α

                  The satellite input coordinate ω ↦ (ω i).2.1.

                  Equations
                  Instances For
                    Used by
                      def

                      InformationTheory.Shannon.BroadcastChannel.bcY₁s

                      source
                      {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} :
                      (U × α × β₁ × β₂)β₁

                      The first-receiver output coordinate ω ↦ (ω i).2.2.1.

                      Equations
                      Instances For
                        Used by
                          def

                          InformationTheory.Shannon.BroadcastChannel.bcY₂s

                          source
                          {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} :
                          (U × α × β₁ × β₂)β₂

                          The second-receiver output coordinate ω ↦ (ω i).2.2.2.

                          Equations
                          Instances For
                            Used by

                              Auxiliary-variable informations #

                              noncomputable def

                              InformationTheory.Shannon.BroadcastChannel.bcInfo₂

                              source
                              {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [Fintype U] [MeasurableSpace U] [MeasurableSpace α] [MeasurableSpace β₁] [Fintype β₂] [MeasurableSpace β₂] (pU : MeasureTheory.Measure U) (K : ProbabilityTheory.Kernel U α) (W : BCChannel α β₁ β₂) :

                              The cloud information I(U; Y₂) = H(U) + H(Y₂) − H(U, Y₂) of the per-coordinate joint law. This is the achievable rate of receiver 2, which decodes the cloud tier alone.

                              Equations
                              • One or more equations did not get rendered due to their size.
                              Instances For
                                Used by
                                  noncomputable def

                                  InformationTheory.Shannon.BroadcastChannel.bcInfo₁

                                  source
                                  {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [Fintype U] [MeasurableSpace U] [Fintype α] [MeasurableSpace α] [Fintype β₁] [MeasurableSpace β₁] [MeasurableSpace β₂] (pU : MeasureTheory.Measure U) (K : ProbabilityTheory.Kernel U α) (W : BCChannel α β₁ β₂) :

                                  The satellite conditional information I(X; Y₁ ∣ U) = H(U, X) + H(U, Y₁) − H(U, X, Y₁) − H(U) of the per-coordinate joint law. This is the achievable rate of receiver 1, which decodes the satellite tier on top of the cloud U. Unlike the MAC macInfo, this is a genuine four-entropy conditional mutual information, not a plain three-term unconditional one.

                                  Equations
                                  • One or more equations did not get rendered due to their size.
                                  Instances For
                                    Used by

                                      Two-tier (cloud / satellite) random codebook #

                                      abbrev

                                      InformationTheory.Shannon.BroadcastChannel.BCCloudCodebook

                                      source
                                      @[reducible, inline]
                                      (M₂ n : ) (U : Type u_5) :
                                      Type u_5

                                      A length-n cloud codebook: for each cloud message w₂ a cloud codeword Uⁿ(w₂).

                                      Equations
                                      Instances For
                                        Used by
                                          abbrev

                                          InformationTheory.Shannon.BroadcastChannel.BCSatelliteCodebook

                                          source
                                          @[reducible, inline]
                                          (M₁ M₂ n : ) (α : Type u_5) :
                                          Type u_5

                                          A length-n satellite codebook: for each message pair (w₁, w₂) a satellite codeword Xⁿ(w₁, w₂). Definitionally the BroadcastCode joint encoder.

                                          Equations
                                          Instances For
                                            Used by
                                              noncomputable def

                                              InformationTheory.Shannon.BroadcastChannel.bcCloudCodebookMeasure

                                              source

                                              The cloud codebook law: pU-i.i.d. over all M₂ · n cloud letters.

                                              Equations
                                              Instances For
                                                Used by
                                                  noncomputable def

                                                  InformationTheory.Shannon.BroadcastChannel.bcSatelliteCodebookMeasure

                                                  source
                                                  {U : Type u_1} {α : Type u_2} [MeasurableSpace U] [MeasurableSpace α] (K : ProbabilityTheory.Kernel U α) (M₁ M₂ n : ) (u : BCCloudCodebook M₂ n U) :

                                                  The satellite codebook law conditional on the cloud codebook u: each satellite letter Xₗ(w₁, w₂) is drawn from K (u w₂ l), independently across pairs and letters. This is the conditional product Πᵢ K(Uᵢ) at the heart of superposition coding — the single point of departure from the MAC flat-product ensemble.

                                                  Equations
                                                  Instances For
                                                    Used by
                                                      noncomputable def

                                                      InformationTheory.Shannon.BroadcastChannel.bcCodebookMeasure

                                                      source
                                                      {U : Type u_1} {α : Type u_2} [MeasurableSpace U] [MeasurableSpace α] (pU : MeasureTheory.Measure U) (K : ProbabilityTheory.Kernel U α) (M₁ M₂ n : ) :

                                                      The joint two-tier codebook law on (cloud, satellite) pairs: draw the cloud codebook from bcCloudCodebookMeasure, then the satellite codebook conditionally from bcSatelliteCodebookMeasure.

                                                      Equations
                                                      • One or more equations did not get rendered due to their size.
                                                      Instances For
                                                        Used by

                                                          I.i.d. coordinate facts for the BC ambient measure #

                                                          Every random variable consumed by the covering bound has the form fun ω ↦ g (ω i) for a measurable coordinate selector g : U × α × β₁ × β₂ → γ. These are the BC analogues of the InformationTheory.Shannon.MAC macAmbient_* lemmas (IIDAmbient.lean), proven the same way via Measure.infinitePi_map_eval / iIndepFun_infinitePi.

                                                          theorem

                                                          InformationTheory.Shannon.BroadcastChannel.bcAmbient_map_coord

                                                          source
                                                          {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [Fintype U] [DecidableEq U] [Nonempty U] [MeasurableSpace U] [MeasurableSingletonClass U] [Fintype α] [DecidableEq α] [Nonempty α] [MeasurableSpace α] [MeasurableSingletonClass α] [Fintype β₁] [DecidableEq β₁] [Nonempty β₁] [MeasurableSpace β₁] [MeasurableSingletonClass β₁] [Fintype β₂] [DecidableEq β₂] [Nonempty β₂] [MeasurableSpace β₂] [MeasurableSingletonClass β₂] {γ : Type u_5} [MeasurableSpace γ] (pU : MeasureTheory.Measure U) [MeasureTheory.IsProbabilityMeasure pU] (K : ProbabilityTheory.Kernel U α) [ProbabilityTheory.IsMarkovKernel K] (W : BCChannel α β₁ β₂) [ProbabilityTheory.IsMarkovKernel W] (g : U × α × β₁ × β₂γ) (hg : Measurable g) (i : ) :
                                                          MeasureTheory.Measure.map (fun (ω : U × α × β₁ × β₂) => g (ω i)) (bcAmbientMeasure pU K W) = MeasureTheory.Measure.map g (bcJointDistribution pU K W)

                                                          The map of a coordinate selector under the BC ambient measure equals the map of the selector under the per-coordinate joint law.

                                                          Used by
                                                            theorem

                                                            InformationTheory.Shannon.BroadcastChannel.bcAmbient_iIndepFun_coord

                                                            source
                                                            {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [Fintype U] [DecidableEq U] [Nonempty U] [MeasurableSpace U] [MeasurableSingletonClass U] [Fintype α] [DecidableEq α] [Nonempty α] [MeasurableSpace α] [MeasurableSingletonClass α] [Fintype β₁] [DecidableEq β₁] [Nonempty β₁] [MeasurableSpace β₁] [MeasurableSingletonClass β₁] [Fintype β₂] [DecidableEq β₂] [Nonempty β₂] [MeasurableSpace β₂] [MeasurableSingletonClass β₂] {γ : Type u_5} [MeasurableSpace γ] (pU : MeasureTheory.Measure U) [MeasureTheory.IsProbabilityMeasure pU] (K : ProbabilityTheory.Kernel U α) [ProbabilityTheory.IsMarkovKernel K] (W : BCChannel α β₁ β₂) [ProbabilityTheory.IsMarkovKernel W] (g : U × α × β₁ × β₂γ) (hg : Measurable g) :
                                                            ProbabilityTheory.iIndepFun (fun (i : ) (ω : U × α × β₁ × β₂) => g (ω i)) (bcAmbientMeasure pU K W)

                                                            Mutual independence of any coordinate selector under the BC ambient measure.

                                                            Used by
                                                              theorem

                                                              InformationTheory.Shannon.BroadcastChannel.bcAmbient_identDistrib_coord

                                                              source
                                                              {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [Fintype U] [DecidableEq U] [Nonempty U] [MeasurableSpace U] [MeasurableSingletonClass U] [Fintype α] [DecidableEq α] [Nonempty α] [MeasurableSpace α] [MeasurableSingletonClass α] [Fintype β₁] [DecidableEq β₁] [Nonempty β₁] [MeasurableSpace β₁] [MeasurableSingletonClass β₁] [Fintype β₂] [DecidableEq β₂] [Nonempty β₂] [MeasurableSpace β₂] [MeasurableSingletonClass β₂] {γ : Type u_5} [MeasurableSpace γ] (pU : MeasureTheory.Measure U) [MeasureTheory.IsProbabilityMeasure pU] (K : ProbabilityTheory.Kernel U α) [ProbabilityTheory.IsMarkovKernel K] (W : BCChannel α β₁ β₂) [ProbabilityTheory.IsMarkovKernel W] (g : U × α × β₁ × β₂γ) (hg : Measurable g) (i : ) :
                                                              ProbabilityTheory.IdentDistrib (fun (ω : U × α × β₁ × β₂) => g (ω i)) (fun (ω : U × α × β₁ × β₂) => g (ω 0)) (bcAmbientMeasure pU K W) (bcAmbientMeasure pU K W)

                                                              Identical distribution of a coordinate selector across indices.

                                                              Used by
                                                                theorem

                                                                InformationTheory.Shannon.BroadcastChannel.bcAmbient_entropy_coord

                                                                source
                                                                {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [Fintype U] [DecidableEq U] [Nonempty U] [MeasurableSpace U] [MeasurableSingletonClass U] [Fintype α] [DecidableEq α] [Nonempty α] [MeasurableSpace α] [MeasurableSingletonClass α] [Fintype β₁] [DecidableEq β₁] [Nonempty β₁] [MeasurableSpace β₁] [MeasurableSingletonClass β₁] [Fintype β₂] [DecidableEq β₂] [Nonempty β₂] [MeasurableSpace β₂] [MeasurableSingletonClass β₂] {γ : Type u_5} [Fintype γ] [Nonempty γ] [MeasurableSpace γ] [MeasurableSingletonClass γ] (pU : MeasureTheory.Measure U) [MeasureTheory.IsProbabilityMeasure pU] (K : ProbabilityTheory.Kernel U α) [ProbabilityTheory.IsMarkovKernel K] (W : BCChannel α β₁ β₂) [ProbabilityTheory.IsMarkovKernel W] (g : U × α × β₁ × β₂γ) (hg : Measurable g) (i : ) :
                                                                (entropy (bcAmbientMeasure pU K W) fun (ω : U × α × β₁ × β₂) => g (ω i)) = entropy (bcJointDistribution pU K W) g

                                                                Entropy of a coordinate selector under the BC ambient measure equals its entropy under the per-coordinate joint law.

                                                                Used by

                                                                  Positivity of the BC per-coordinate joint law and coordinate marginals #

                                                                  theorem

                                                                  InformationTheory.Shannon.BroadcastChannel.bcJointDistribution_singleton_pos

                                                                  source
                                                                  {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [Fintype U] [DecidableEq U] [Nonempty U] [MeasurableSpace U] [MeasurableSingletonClass U] [Fintype α] [DecidableEq α] [Nonempty α] [MeasurableSpace α] [MeasurableSingletonClass α] [Fintype β₁] [DecidableEq β₁] [Nonempty β₁] [MeasurableSpace β₁] [MeasurableSingletonClass β₁] [Fintype β₂] [DecidableEq β₂] [Nonempty β₂] [MeasurableSpace β₂] [MeasurableSingletonClass β₂] (pU : MeasureTheory.Measure U) [MeasureTheory.IsProbabilityMeasure pU] (K : ProbabilityTheory.Kernel U α) [ProbabilityTheory.IsMarkovKernel K] (W : BCChannel α β₁ β₂) [ProbabilityTheory.IsMarkovKernel W] (hpU : ∀ (a : U), 0 < pU.real {a}) (hK : ∀ (a : U) (b : α), 0 < (K a).real {b}) (hW : ∀ (a : α) (b : β₁ × β₂), 0 < (W a).real {b}) (q : U × α × β₁ × β₂) :

                                                                  The per-coordinate BC joint law has positive singleton mass.

                                                                  Used by
                                                                    theorem

                                                                    InformationTheory.Shannon.BroadcastChannel.bcAmbient_coord_marginal_pos

                                                                    source
                                                                    {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [Fintype U] [DecidableEq U] [Nonempty U] [MeasurableSpace U] [MeasurableSingletonClass U] [Fintype α] [DecidableEq α] [Nonempty α] [MeasurableSpace α] [MeasurableSingletonClass α] [Fintype β₁] [DecidableEq β₁] [Nonempty β₁] [MeasurableSpace β₁] [MeasurableSingletonClass β₁] [Fintype β₂] [DecidableEq β₂] [Nonempty β₂] [MeasurableSpace β₂] [MeasurableSingletonClass β₂] {γ : Type u_5} [MeasurableSpace γ] [MeasurableSingletonClass γ] (pU : MeasureTheory.Measure U) [MeasureTheory.IsProbabilityMeasure pU] (K : ProbabilityTheory.Kernel U α) [ProbabilityTheory.IsMarkovKernel K] (W : BCChannel α β₁ β₂) [ProbabilityTheory.IsMarkovKernel W] (hpU : ∀ (a : U), 0 < pU.real {a}) (hK : ∀ (a : U) (b : α), 0 < (K a).real {b}) (hW : ∀ (a : α) (b : β₁ × β₂), 0 < (W a).real {b}) (g : U × α × β₁ × β₂γ) (hg : Measurable g) (i : ) (c : γ) (r : U × α × β₁ × β₂) (hr : g r = c) :
                                                                    0 < (MeasureTheory.Measure.map (fun (ω : U × α × β₁ × β₂) => g (ω i)) (bcAmbientMeasure pU K W)).real {c}

                                                                    Positivity of any coordinate-selector marginal singleton, reduced to the per-coordinate joint positivity via a chosen fiber witness.

                                                                    Used by

                                                                      (U, X) marginal factorization #

                                                                      theorem

                                                                      InformationTheory.Shannon.BroadcastChannel.bcJointDistribution_map_fst

                                                                      source

                                                                      The U-marginal of the BC joint law is pU.

                                                                      Used by
                                                                        theorem

                                                                        InformationTheory.Shannon.BroadcastChannel.bcJointDistribution_map_UX_singleton

                                                                        source
                                                                        {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [Fintype U] [DecidableEq U] [Nonempty U] [MeasurableSpace U] [MeasurableSingletonClass U] [Fintype α] [DecidableEq α] [Nonempty α] [MeasurableSpace α] [MeasurableSingletonClass α] [Fintype β₁] [DecidableEq β₁] [Nonempty β₁] [MeasurableSpace β₁] [MeasurableSingletonClass β₁] [Fintype β₂] [DecidableEq β₂] [Nonempty β₂] [MeasurableSpace β₂] [MeasurableSingletonClass β₂] (pU : MeasureTheory.Measure U) [MeasureTheory.IsProbabilityMeasure pU] (K : ProbabilityTheory.Kernel U α) [ProbabilityTheory.IsMarkovKernel K] (W : BCChannel α β₁ β₂) [ProbabilityTheory.IsMarkovKernel W] (a : U) (x : α) :
                                                                        (MeasureTheory.Measure.map (fun (q : U × α × β₁ × β₂) => (q.1, q.2.1)) (bcJointDistribution pU K W)).real {(a, x)} = pU.real {a} * (K a).real {x}

                                                                        The (U, X)-marginal singleton mass of the BC joint law factorizes as pU {u} · K u {x}.

                                                                        Used by

                                                                          Relabeling invariance of the typical set (BC-local copy of the MAC helper) #

                                                                          The two exponential ingredients of the covering bound #

                                                                          theorem

                                                                          InformationTheory.Shannon.BroadcastChannel.bc_perseq_mass_le

                                                                          source
                                                                          {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [Fintype U] [DecidableEq U] [Nonempty U] [MeasurableSpace U] [MeasurableSingletonClass U] [Fintype α] [DecidableEq α] [Nonempty α] [MeasurableSpace α] [MeasurableSingletonClass α] [Fintype β₁] [DecidableEq β₁] [Nonempty β₁] [MeasurableSpace β₁] [MeasurableSingletonClass β₁] [Fintype β₂] [DecidableEq β₂] [Nonempty β₂] [MeasurableSpace β₂] [MeasurableSingletonClass β₂] (pU : MeasureTheory.Measure U) [MeasureTheory.IsProbabilityMeasure pU] (K : ProbabilityTheory.Kernel U α) [ProbabilityTheory.IsMarkovKernel K] (W : BCChannel α β₁ β₂) [ProbabilityTheory.IsMarkovKernel W] (hpU : ∀ (a : U), 0 < pU.real {a}) (hK : ∀ (a : U) (b : α), 0 < (K a).real {b}) {n : } {ε : } (u : Fin nU) (x : Fin nα) (hu : u typicalSet (bcAmbientMeasure pU K W) bcUs n ε) (hux : (fun (i : Fin n) => (u i, x i)) typicalSet (bcAmbientMeasure pU K W) (ChannelCoding.jointSequence bcUs bcXs) n ε) :

                                                                          Per-sequence conditional mass bound: for a typical cloud u and a satellite x whose (U, X)-pair sequence is typical, the conditional-product mass of x is at most exp(−n (H(U, X) − H(U) − 2ε)).

                                                                          Used by
                                                                            theorem

                                                                            InformationTheory.Shannon.BroadcastChannel.bc_slice_card_le

                                                                            source
                                                                            {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [Fintype U] [DecidableEq U] [Nonempty U] [MeasurableSpace U] [MeasurableSingletonClass U] [Fintype α] [DecidableEq α] [Nonempty α] [MeasurableSpace α] [MeasurableSingletonClass α] [Fintype β₁] [DecidableEq β₁] [Nonempty β₁] [MeasurableSpace β₁] [MeasurableSingletonClass β₁] [Fintype β₂] [DecidableEq β₂] [Nonempty β₂] [MeasurableSpace β₂] [MeasurableSingletonClass β₂] (pU : MeasureTheory.Measure U) [MeasureTheory.IsProbabilityMeasure pU] (K : ProbabilityTheory.Kernel U α) [ProbabilityTheory.IsMarkovKernel K] (W : BCChannel α β₁ β₂) [ProbabilityTheory.IsMarkovKernel W] (hpU : ∀ (a : U), 0 < pU.real {a}) (hK : ∀ (a : U) (b : α), 0 < (K a).real {b}) (hW : ∀ (a : α) (b : β₁ × β₂), 0 < (W a).real {b}) {n : } {ε : } (u : Fin nU) (y₁ : Fin nβ₁) :

                                                                            Slice-cardinality bound: the number of satellites x making (u, x, y₁) jointly typical is at most exp(n (H(X, (U, Y₁)) − H(U, Y₁) + 2ε)).

                                                                            Used by

                                                                              Conditional-slice satellite typicality bound #

                                                                              theorem

                                                                              InformationTheory.Shannon.BroadcastChannel.bc_conditional_slice_prob_le

                                                                              source
                                                                              {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [Fintype U] [DecidableEq U] [Nonempty U] [MeasurableSpace U] [MeasurableSingletonClass U] [Fintype α] [DecidableEq α] [Nonempty α] [MeasurableSpace α] [MeasurableSingletonClass α] [Fintype β₁] [DecidableEq β₁] [Nonempty β₁] [MeasurableSpace β₁] [MeasurableSingletonClass β₁] [Fintype β₂] [DecidableEq β₂] [Nonempty β₂] [MeasurableSpace β₂] [MeasurableSingletonClass β₂] (pU : MeasureTheory.Measure U) [MeasureTheory.IsProbabilityMeasure pU] (K : ProbabilityTheory.Kernel U α) [ProbabilityTheory.IsMarkovKernel K] (W : BCChannel α β₁ β₂) [ProbabilityTheory.IsMarkovKernel W] (hpU : ∀ (a : U), 0 < pU.real {a}) (hK : ∀ (a : U) (b : α), 0 < (K a).real {b}) (hW : ∀ (a : α) (b : β₁ × β₂), 0 < (W a).real {b}) {n : } {ε : } (u : Fin nU) (y₁ : Fin nβ₁) (hu : u typicalSet (bcAmbientMeasure pU K W) bcUs n ε) (hy₁ : y₁ typicalSet (bcAmbientMeasure pU K W) bcY₁s n ε) :
                                                                              (MeasureTheory.Measure.pi fun (l : Fin n) => K (u l)).real {x : Fin nα | (u, x, y₁) MAC.macJointlyTypicalSet (bcAmbientMeasure pU K W) bcUs bcXs bcY₁s n ε} Real.exp (-n * (bcInfo₁ pU K W - 4 * ε))

                                                                              Conditional-slice satellite typicality probability bound for the superposition covering argument. For a fixed typical cloud codeword u and a fixed typical received word y₁, the probability under the conditional product law Πᵢ K(uᵢ) that an independently drawn satellite x is jointly typical with (u, y₁) is at most exp(−n (I(X; Y₁ ∣ U) − 4ε)). This is the receiver-1 "wrong satellite, correct cloud" sub-event of the superposition random-coding argument (Cover–Thomas); the exponent matches bcInfo₁, with the slack the sum of the four entropy-typicality windows (matching the slack of the MAC lemmas macJTS_indep_prob_le_*). Full support (hpU/hK/hW) is a regularity precondition of the AEP mass bounds, not load-bearing. @audit:ok

                                                                              Used by

                                                                                Two-tier decoders and the assembled broadcast code #

                                                                                noncomputable def

                                                                                InformationTheory.Shannon.BroadcastChannel.bcCloudTypicalDecoder

                                                                                source
                                                                                {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [Fintype U] [MeasurableSpace U] [MeasurableSpace α] [MeasurableSpace β₁] [Fintype β₂] [MeasurableSpace β₂] (pU : MeasureTheory.Measure U) (K : ProbabilityTheory.Kernel U α) (W : BCChannel α β₁ β₂) {M₂ n : } (hM₂ : 0 < M₂) (ε : ) (cU : BCCloudCodebook M₂ n U) :
                                                                                (Fin nβ₂)Fin M₂

                                                                                Receiver-2 (cloud tier) joint-typical decoder. Given a received word y₂, returns the unique cloud message w₂ whose codeword Uⁿ(w₂) is jointly typical with y₂, falling back to ⟨0, hM₂⟩ if no such w₂ exists or it is not unique. This is a single-user joint-typical decoder over the cloud codebook — receiver 2 never needs the satellite tier.

                                                                                Equations
                                                                                • One or more equations did not get rendered due to their size.
                                                                                Instances For
                                                                                  Used by
                                                                                    noncomputable def

                                                                                    InformationTheory.Shannon.BroadcastChannel.bcJointTypicalDecoder

                                                                                    source
                                                                                    {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [Fintype U] [MeasurableSpace U] [Fintype α] [MeasurableSpace α] [Fintype β₁] [MeasurableSpace β₁] [MeasurableSpace β₂] (pU : MeasureTheory.Measure U) (K : ProbabilityTheory.Kernel U α) (W : BCChannel α β₁ β₂) {M₁ M₂ n : } (hM₁ : 0 < M₁) (hM₂ : 0 < M₂) (ε : ) (cU : BCCloudCodebook M₂ n U) (cX : BCSatelliteCodebook M₁ M₂ n α) :
                                                                                    (Fin nβ₁)Fin M₁ × Fin M₂

                                                                                    Receiver-1 superposition joint-typical decoder. Given a received word y₁, returns the unique message pair (w₁, w₂) such that the cloud/satellite/output triple (Uⁿ(w₂), Xⁿ(w₁, w₂), y₁) is jointly typical, falling back to (⟨0, hM₁⟩, ⟨0, hM₂⟩) otherwise. The typical-set argument order bcUs, bcXs, bcY₁s matches the covering bound bc_conditional_slice_prob_le.

                                                                                    Equations
                                                                                    • One or more equations did not get rendered due to their size.
                                                                                    Instances For
                                                                                      Used by
                                                                                        noncomputable def

                                                                                        InformationTheory.Shannon.BroadcastChannel.bcCodebookToCode

                                                                                        source
                                                                                        {U : Type u_1} {α : Type u_2} {β₁ : Type u_3} {β₂ : Type u_4} [Fintype U] [MeasurableSpace U] [Fintype α] [MeasurableSpace α] [Fintype β₁] [MeasurableSpace β₁] [Fintype β₂] [MeasurableSpace β₂] (pU : MeasureTheory.Measure U) (K : ProbabilityTheory.Kernel U α) (W : BCChannel α β₁ β₂) {M₁ M₂ n : } (hM₁ : 0 < M₁) (hM₂ : 0 < M₂) (ε : ) (cU : BCCloudCodebook M₂ n U) (cX : BCSatelliteCodebook M₁ M₂ n α) :
                                                                                        BroadcastCode M₁ M₂ n α β₁ β₂

                                                                                        Bundle a cloud codebook cU and satellite codebook cX into a BroadcastCode: cX is the joint encoder, receiver 1 uses the superposition joint decoder, receiver 2 the cloud decoder.

                                                                                        Equations
                                                                                        • One or more equations did not get rendered due to their size.
                                                                                        Instances For
                                                                                          Used by