Griffiths–Hurst–Sherman inequality
DefinitionCurrent connecting all three site pairs#✓fully proved
The current m connects each of the pairs {x, y}, {o, y}, {o, x}.
def aie_allConn (ends : ι → Sym2 V) (m : Finset ι) (o x y : V) : Prop :=
connK ends m x y ∧ connK ends m o y ∧ connK ends m o x
def connK (ends : ι → Sym2 V) (K : Finset ι) : V → V → Prop :=
Relation.ReflTransGen (adjStep ends K)
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Switching.lean:90
Uses 1 · Used by 2
DefinitionCurrent connecting none of the three site pairs#✓fully proved
The current m connects none of the pairs {x, y}, {o, y}, {o, x}.
def aie_noneConn (ends : ι → Sym2 V) (m : Finset ι) (o x y : V) : Prop :=
¬ connK ends m x y ∧ ¬ connK ends m o y ∧ ¬ connK ends m o x
def connK (ends : ι → Sym2 V) (K : Finset ι) : V → V → Prop :=
Relation.ReflTransGen (adjStep ends K)
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Switching.lean:90
Uses 1 · Used by 2
DefinitionMass of sub-currents with a given source set#✓fully proved
The total weight ∑ φ(K) of the sub-currents K ⊆ m with source set S.
noncomputable def aie_mass (ends : ι → Sym2 V) (m : Finset ι) (φ : Finset ι → ℝ) (S : Finset V) :
ℝ :=
∑ K ∈ m.powerset.filter (fun K => sources ends K = S), φ K
def sources (ends : ι → Sym2 V) (T : Finset ι) : Finset V :=
(Finset.univ : Finset V).filter (fun x => Odd (degK ends T x))
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Switching.lean:77
Uses 1 · Used by 3
DefinitionGHS gap of a sourced current#✓fully proved
The GHS gap mass(A) - mass(A ∆ {x, y}) - mass(A ∆ {o, y}) - mass(A ∆ {o, x}).
noncomputable def aie_gap (ends : ι → Sym2 V) (m : Finset ι) (φ : Finset ι → ℝ) (A : Finset V)
(o x y : V) : ℝ :=
aie_mass ends m φ A - aie_mass ends m φ (A ∆ {x, y})
- aie_mass ends m φ (A ∆ {o, y}) - aie_mass ends m φ (A ∆ {o, x})
noncomputable def aie_mass (ends : ι → Sym2 V) (m : Finset ι) (φ : Finset ι → ℝ) (S : Finset V) :
ℝ :=
∑ K ∈ m.powerset.filter (fun K => sources ends K = S), φ K
PhyslibExt/CondensedMatter/IsingModel/AizenmanInclusionExclusion.lean:147
Uses 1
DefinitionEdges of the triangle on three vertices#✓fully proved
The three edges {0, 1}, {1, 2}, {0, 2} of the triangle on Fin 3.
Used by 1
LemmaDomination condition fails on the triangle#✓fully proved
The domination condition fails for the single current through all edges of the triangle.
lemma ahd_hdom_refutable :
¬ (2 * (∑ m ∈ ({({0, 1, 2} : Finset (Fin 3))} : Finset (Finset (Fin 3))).filter
(fun m => aie_allConn ahd_triEnds m 0 1 2),
aie_mass ahd_triEnds m (fun _ => 1)
(sources ahd_triEnds ({0, 1, 2} : Finset (Fin 3))))
≤ ∑ m ∈ ({({0, 1, 2} : Finset (Fin 3))} : Finset (Finset (Fin 3))).filter
(fun m => aie_noneConn ahd_triEnds m 0 1 2),
aie_mass ahd_triEnds m (fun _ => 1)
(sources ahd_triEnds ({0, 1, 2} : Finset (Fin 3))))
def aie_allConn (ends : ι → Sym2 V) (m : Finset ι) (o x y : V) : Prop :=
connK ends m x y ∧ connK ends m o y ∧ connK ends m o x
PhyslibExt/CondensedMatter/IsingModel/AizenmanInclusionExclusion.lean:84
def ahd_triEnds : Fin 3 → Sym2 (Fin 3)
| 0 => s(0, 1)
| 1 => s(1, 2)
| 2 => s(0, 2)
PhyslibExt/CondensedMatter/IsingModel/AizenmanHdom.lean:134
noncomputable def aie_mass (ends : ι → Sym2 V) (m : Finset ι) (φ : Finset ι → ℝ) (S : Finset V) :
ℝ :=
∑ K ∈ m.powerset.filter (fun K => sources ends K = S), φ K
PhyslibExt/CondensedMatter/IsingModel/AizenmanInclusionExclusion.lean:147
def sources (ends : ι → Sym2 V) (T : Finset ι) : Finset V :=
(Finset.univ : Finset V).filter (fun x => Odd (degK ends T x))
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Switching.lean:77
def aie_noneConn (ends : ι → Sym2 V) (m : Finset ι) (o x y : V) : Prop :=
¬ connK ends m x y ∧ ¬ connK ends m o y ∧ ¬ connK ends m o x
PhyslibExt/CondensedMatter/IsingModel/AizenmanInclusionExclusion.lean:88
DefinitionThree-site backbone factor#✓fully proved
The three-site backbone factor: 1, 0 or -2 according to how many of the pairs {x, y},
{o, y}, {o, x} the current connects.
noncomputable def asd_ghsBackboneFactor (ends : ι → Sym2 V) (m : Finset ι) (o x y : V) : ℝ :=
1 - (if connK ends m x y then 1 else 0)
- (if connK ends m o y then 1 else 0)
- (if connK ends m o x then 1 else 0)
def connK (ends : ι → Sym2 V) (K : Finset ι) : V → V → Prop :=
Relation.ReflTransGen (adjStep ends K)
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Switching.lean:90
Uses 1
DefinitionBackbone cluster of a site in a current#✓fully proved
The cluster of o in the current m: the sites connected to o by edges of m.
noncomputable def atb_backbone (ends : ι → Sym2 V) (m : Finset ι) (o : V) : Finset V :=
compK ends m o
Uses 1
DefinitionConditioned domination condition over a set of currents#✓fully proved
The domination condition 2 · mass(all connected) ≤ mass(none connected), summed over
currents in M.
def atb_conditionedDom (ends : ι → Sym2 V) (M : Finset (Finset ι)) (A : Finset V)
(o x y : V) (Φ : Finset ι → Finset ι → ℝ) : Prop :=
2 * (∑ m ∈ M.filter (fun m => aie_allConn ends m o x y), aie_mass ends m (Φ m) A)
≤ ∑ m ∈ M.filter (fun m => aie_noneConn ends m o x y), aie_mass ends m (Φ m) A
def aie_allConn (ends : ι → Sym2 V) (m : Finset ι) (o x y : V) : Prop :=
connK ends m x y ∧ connK ends m o y ∧ connK ends m o x
PhyslibExt/CondensedMatter/IsingModel/AizenmanInclusionExclusion.lean:84
noncomputable def aie_mass (ends : ι → Sym2 V) (m : Finset ι) (φ : Finset ι → ℝ) (S : Finset V) :
ℝ :=
∑ K ∈ m.powerset.filter (fun K => sources ends K = S), φ K
PhyslibExt/CondensedMatter/IsingModel/AizenmanInclusionExclusion.lean:147
def aie_noneConn (ends : ι → Sym2 V) (m : Finset ι) (o x y : V) : Prop :=
¬ connK ends m x y ∧ ¬ connK ends m o y ∧ ¬ connK ends m o x
PhyslibExt/CondensedMatter/IsingModel/AizenmanInclusionExclusion.lean:88
DefinitionThree-site duplicated-variable factor#✓fully proved
The three-site duplicated-variable factor 1 - σ_x σ_y - σ_o σ_y - σ_o σ_x of the configuration
q.
noncomputable def ghsQFactor (q : ConfigSpace V) (o x y : V) : ℝ :=
1 - spin q x * spin q y - spin q o * spin q y - spin q o * spin q x
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 2
DefinitionThree-replica Ising expectation#✓fully proved
The expectation of F a b c over three independent copies a, b, c of the Ising Gibbs
distribution.
noncomputable def isingExp3 (β h : ℝ)
(F : ConfigSpace V → ConfigSpace V → ConfigSpace V → ℝ) : ℝ :=
∑ a : ConfigSpace V, ∑ b : ConfigSpace V, ∑ c : ConfigSpace V,
isingProb G β h a * isingProb G β h b * isingProb G β h c * F a b c
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def isingProb (β h : ℝ) (s : ConfigSpace V) : ℝ :=
isingWeight G β h s / isingZ G β h
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:141
Uses 2
DefinitionThree-replica variable \(d_w\)#✓fully proved
The three-replica variable d_w = σ_w(b) + σ_w(c) - 2σ_w(a).
def dvar (a b c : ConfigSpace V) (w : V) : ℝ := spin b w + spin c w - 2 * spin a w
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 2
DefinitionTwo-replica Ising expectation#✓fully proved
The expectation of F a b over two independent copies a, b of the Ising Gibbs
distribution.
noncomputable def isingExp2 (β h : ℝ) (F : ConfigSpace V → ConfigSpace V → ℝ) : ℝ :=
∑ a : ConfigSpace V, ∑ b : ConfigSpace V,
isingProb G β h a * isingProb G β h b * F a b
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def isingProb (β h : ℝ) (s : ConfigSpace V) : ℝ :=
isingWeight G β h s / isingZ G β h
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:141
Uses 2 · Used by 2
DefinitionRotated two-replica difference variable \(u_w\)#✓fully proved
The rotated two-replica variable u_w = (σ_w(a) - σ_w(b)) / 2.
noncomputable def uvar (a b : ConfigSpace V) (w : V) : ℝ := (spin a w - spin b w) / 2
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 2 · Used by 6
DefinitionRotated two-replica sum variable \(t_w\)#✓fully proved
The rotated two-replica variable t_w = (σ_w(a) + σ_w(b)) / 2.
noncomputable def tvar (a b : ConfigSpace V) (w : V) : ℝ := (spin a w + spin b w) / 2
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 2 · Used by 6
DefinitionEdges of the induced subgraph on a vertex set#✓fully proved
The edges of E with both endpoints in S (the edges of the induced subgraph on S).
def ghsvp_vertexEdges (E : Finset (Sym2 W)) (S : Finset W) : Finset (Sym2 W) :=
E.filter (Sharpness.edgeInside S)
def edgeInside (S : Finset V) (e : Sym2 V) : Prop := ∀ v ∈ e, v ∈ S
PhyslibExt/CondensedMatter/IsingModel/Sharpness/CurrentSplitting.lean:60
Uses 1 · Used by 16
DefinitionZero-field partition function of an induced subgraph#✓fully proved
The zero-field partition function of the induced subgraph on S, with couplings J.
noncomputable def ghsvp_vertexZ (E : Finset (Sym2 W)) (J : Sym2 W -> Real)
(S : Finset W) : Real :=
ZJ (ghsvp_vertexEdges E S) J (fun _ => 0)
noncomputable def ZJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ) : ℝ :=
∑ s : ConfigSpace V, wJ E J hf s
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:77
def ghsvp_vertexEdges (E : Finset (Sym2 W)) (S : Finset W) : Finset (Sym2 W) :=
E.filter (Sharpness.edgeInside S)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:211
Uses 2 · Used by 6
DefinitionZero-field correlation in an induced subgraph#✓fully proved
The zero-field correlation ⟨σ_P⟩ in the induced subgraph on S.
noncomputable def ghsvp_vertexCorr (E : Finset (Sym2 W)) (J : Sym2 W -> Real)
(S P : Finset W) : Real :=
expJ (ghsvp_vertexEdges E S) J (fun _ => 0) (spinProd P)
noncomputable def expJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ)
(f : ConfigSpace V → ℝ) : ℝ :=
(∑ s : ConfigSpace V, f s * wJ E J hf s) / ZJ E J hf
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:85
def ghsvp_vertexEdges (E : Finset (Sym2 W)) (S : Finset W) : Finset (Sym2 W) :=
E.filter (Sharpness.edgeInside S)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:211
def spinProd (A : Finset V) (s : ConfigSpace V) : ℝ := ∏ x ∈ A, spin s x
PhyslibExt/CondensedMatter/IsingModel/GKS.lean:186
Uses 3 · Used by 5
DefinitionCorrelation with field in an induced subgraph#✓fully proved
The correlation ⟨σ_P⟩ in the induced subgraph on S, with field hf.
noncomputable def ghsvp_vertexCorrField (E : Finset (Sym2 W)) (J : Sym2 W -> Real)
(hf : W -> Real) (S P : Finset W) : Real :=
expJ (ghsvp_vertexEdges E S) J hf (spinProd P)
noncomputable def expJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ)
(f : ConfigSpace V → ℝ) : ℝ :=
(∑ s : ConfigSpace V, f s * wJ E J hf s) / ZJ E J hf
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:85
def ghsvp_vertexEdges (E : Finset (Sym2 W)) (S : Finset W) : Finset (Sym2 W) :=
E.filter (Sharpness.edgeInside S)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:211
def spinProd (A : Finset V) (s : ConfigSpace V) : ℝ := ∏ x ∈ A, spin s x
PhyslibExt/CondensedMatter/IsingModel/GKS.lean:186
Uses 3 · Used by 5
DefinitionConfiguration merged from two configurations along a set#✓fully proved
The configuration equal to a on S and to b off S.
def ghsvp_mergeConfig (S : Finset W) (a : (↑S -> Bool))
(b : ({v : W // v ∉ S}) -> Bool) : ConfigSpace W :=
fun v => if hv : v ∈ S then a ⟨v, hv⟩ else b ⟨v, hv⟩
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
Uses 1 · Used by 2
DefinitionSplitting a configuration into a set and its complement#✓fully proved
Configurations split into their restrictions to S and to its complement.
def ghsvp_splitConfig (S : Finset W) :
ConfigSpace W ≃ ((↑S -> Bool) × (({v : W // v ∉ S}) -> Bool)) where
toFun s := (fun v => s v, fun v => s v)
invFun p := ghsvp_mergeConfig S p.1 p.2
left_inv s := by
funext v
simp only [ghsvp_mergeConfig]
split <;> rfl
right_inv p := by
apply Prod.ext
· funext v
simp [ghsvp_mergeConfig, v.property]
· funext v
simp [ghsvp_mergeConfig, v.property]
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def ghsvp_mergeConfig (S : Finset W) (a : (↑S -> Bool))
(b : ({v : W // v ∉ S}) -> Bool) : ConfigSpace W :=
fun v => if hv : v ∈ S then a ⟨v, hv⟩ else b ⟨v, hv⟩
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:305
DefinitionInduced-subgraph partition function over the active sites only#✓fully proved
The partition function of the induced subgraph on S, summed over configurations on S
only.
noncomputable def ghsvp_activeZ (E : Finset (Sym2 W)) (J : Sym2 W -> Real)
(S : Finset W) : Real :=
∑ a : ↑S -> Bool,
wJ (ghsvp_vertexEdges E S) J (fun _ => 0)
(ghsvp_mergeConfig S a (fun _ => false))
noncomputable def wJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ)
(s : ConfigSpace V) : ℝ :=
Real.exp ((∑ e ∈ E, J e * bond s e) + ∑ x, hf x * spin s x)
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:64
def ghsvp_vertexEdges (E : Finset (Sym2 W)) (S : Finset W) : Finset (Sym2 W) :=
E.filter (Sharpness.edgeInside S)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:211
def ghsvp_mergeConfig (S : Finset W) (a : (↑S -> Bool))
(b : ({v : W // v ∉ S}) -> Bool) : ConfigSpace W :=
fun v => if hv : v ∈ S then a ⟨v, hv⟩ else b ⟨v, hv⟩
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:305
DefinitionImage of a vertex set among non-ghost vertices#✓fully proved
The image of A among the non-ghost vertices some x of Option V.
def ghsvp_someSet (A : Finset V) : Finset (Option V) := A.map someEmb
def someEmb : V ↪ Option V := ⟨some, Option.some_injective V⟩
PhyslibExt/CondensedMatter/IsingModel/Sharpness/FieldGhostDict.lean:252
Uses 1 · Used by 7
DefinitionImage of a vertex set together with the ghost vertex#✓fully proved
The image of A together with the ghost vertex none.
def ghsvp_fieldSet (A : Finset V) : Finset (Option V) := insert none (ghsvp_someSet A)
def ghsvp_someSet (A : Finset V) : Finset (Option V) := A.map someEmb
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:377
Uses 1 · Used by 6
DefinitionWeight of an agreement set in the duplicated system#✓fully proved
The weight of the agreement set A in the duplicated system: the product of the ghost-vertex
partition functions of A ∪ {ghost} and of Aᶜ, with doubled couplings 2β and field
2βh.
noncomputable def ghsvp_subsetMass (G : SimpleGraph V) [DecidableRel G.Adj]
(β h : Real) (A : Finset V) : Real :=
let J2 := ghostCoupling (2 * β * h) 1 (fun _ => 2 * β)
ghsvp_vertexZ (withGhost G).edgeFinset J2 (ghsvp_fieldSet A) *
ghsvp_vertexZ (withGhost G).edgeFinset J2 (ghsvp_someSet Aᶜ)
noncomputable def ghostCoupling (h _β : ℝ) (J : Sym2 V → ℝ) (e : Sym2 (Option V)) : ℝ :=
Sym2.lift ⟨fun a b =>
match a, b with
| some x, some y => J s(x, y)
| some _, none => h
| none, some _ => h
| none, none => 0,
by rintro (_ | x) (_ | y) <;> simp only []; rw [Sym2.eq_swap]⟩ e
PhyslibExt/CondensedMatter/IsingModel/Sharpness/FieldGhostDict.lean:122
noncomputable def ghsvp_vertexZ (E : Finset (Sym2 W)) (J : Sym2 W -> Real)
(S : Finset W) : Real :=
ZJ (ghsvp_vertexEdges E S) J (fun _ => 0)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:215
def withGhost : SimpleGraph (Option V) where
Adj a b :=
match a, b with
| some x, some y => G.Adj x y
| some _, none => True
| none, some _ => True
| none, none => False
symm := by
constructor
rintro (_ | x) (_ | y) h
· exact h
· exact h
⋯
PhyslibExt/CondensedMatter/IsingModel/Sharpness/RandomCurrent.lean:237
def ghsvp_fieldSet (A : Finset V) : Finset (Option V) := insert none (ghsvp_someSet A)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:380
def ghsvp_someSet (A : Finset V) : Finset (Option V) := A.map someEmb
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:377
DefinitionInduced-subgraph partition function with doubled coupling and field#✓fully proved
The partition function on the induced subgraph on A with couplings 2β and field 2βh.
noncomputable def ghsvp_tZ (G : SimpleGraph V) [DecidableRel G.Adj]
(β h : Real) (A : Finset V) : Real :=
ZJ (ghsvp_vertexEdges G.edgeFinset A) (fun _ => 2 * β)
(fun x => if x ∈ A then 2 * β * h else 0)
noncomputable def ZJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ) : ℝ :=
∑ s : ConfigSpace V, wJ E J hf s
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:77
def ghsvp_vertexEdges (E : Finset (Sym2 W)) (S : Finset W) : Finset (Sym2 W) :=
E.filter (Sharpness.edgeInside S)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:211
DefinitionZero-field induced-subgraph partition function with doubled coupling#✓fully proved
The zero-field partition function on the induced subgraph on A with couplings 2β.
noncomputable def ghsvp_uZ (G : SimpleGraph V) [DecidableRel G.Adj]
(β : Real) (A : Finset V) : Real :=
ZJ (ghsvp_vertexEdges G.edgeFinset A) (fun _ => 2 * β) (fun _ => 0)
noncomputable def ZJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ) : ℝ :=
∑ s : ConfigSpace V, wJ E J hf s
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:77
def ghsvp_vertexEdges (E : Finset (Sym2 W)) (S : Finset W) : Finset (Sym2 W) :=
E.filter (Sharpness.edgeInside S)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:211
DefinitionTwo-point correlation on the disagreement subgraph#✓fully proved
The zero-field correlation ⟨σ_o σ_x⟩ in the subgraph induced by Aᶜ (couplings 2β), or
0 if o or x lies in A.
noncomputable def ghsvp_disagreementCorr (G : SimpleGraph V) [DecidableRel G.Adj]
(β : Real) (o x : V) (A : Finset V) : Real :=
if o ∈ A ∨ x ∈ A then 0
else ghsvp_vertexCorr G.edgeFinset (fun _ => 2 * β) Aᶜ {o, x}
noncomputable def ghsvp_vertexCorr (E : Finset (Sym2 W)) (J : Sym2 W -> Real)
(S P : Finset W) : Real :=
expJ (ghsvp_vertexEdges E S) J (fun _ => 0) (spinProd P)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:259
Uses 1 · Used by 2
DefinitionMagnetization on the agreement subgraph#✓fully proved
The magnetization ⟨σ_y⟩ in the subgraph induced by A (couplings 2β, field 2βh), or
0 if y ∉ A.
noncomputable def ghsvp_agreementMag (G : SimpleGraph V) [DecidableRel G.Adj]
(β h : Real) (y : V) (A : Finset V) : Real :=
if y ∈ A then
ghsvp_vertexCorrField G.edgeFinset (fun _ => 2 * β) (fun _ => 2 * β * h) A {y}
else 0
noncomputable def ghsvp_vertexCorrField (E : Finset (Sym2 W)) (J : Sym2 W -> Real)
(hf : W -> Real) (S P : Finset W) : Real :=
expJ (ghsvp_vertexEdges E S) J hf (spinProd P)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:264
Uses 1 · Used by 2
DefinitionInhomogeneous ghost-vertex couplings#✓fully proved
The ghost-vertex couplings: 2K_{xy} on an edge {x, y}, 2h_x between x and the ghost.
noncomputable def ghsiGhostCoupling (K : Sym2 V → ℝ) (hf : V → ℝ)
(e : Sym2 (Option V)) : ℝ :=
Sym2.lift ⟨fun a b =>
match a, b with
| some x, some y => 2 * K s(x, y)
| some x, none => 2 * hf x
| none, some y => 2 * hf y
| none, none => 0,
by rintro (_ | x) (_ | y) <;> simp only [] <;> rw [Sym2.eq_swap]⟩ e
Used by 2
DefinitionInhomogeneous induced-subgraph partition function with field#✓fully proved
The partition function of the subgraph induced by A with couplings 2K and fields 2h on
A.
noncomputable def ghsiTZ (G : SimpleGraph V) [DecidableRel G.Adj]
(K : Sym2 V → ℝ) (hf : V → ℝ) (A : Finset V) : ℝ :=
ZJ (ghsvp_vertexEdges G.edgeFinset A) (fun e => 2 * K e)
(fun x => if x ∈ A then 2 * hf x else 0)
noncomputable def ZJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ) : ℝ :=
∑ s : ConfigSpace V, wJ E J hf s
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:77
def ghsvp_vertexEdges (E : Finset (Sym2 W)) (S : Finset W) : Finset (Sym2 W) :=
E.filter (Sharpness.edgeInside S)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:211
Uses 2 · Used by 1
DefinitionInhomogeneous zero-field induced-subgraph partition function#✓fully proved
The zero-field partition function of the subgraph induced by A with couplings 2K.
noncomputable def ghsiUZ (G : SimpleGraph V) [DecidableRel G.Adj]
(K : Sym2 V → ℝ) (A : Finset V) : ℝ :=
ZJ (ghsvp_vertexEdges G.edgeFinset A) (fun e => 2 * K e) (fun _ => 0)
noncomputable def ZJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ) : ℝ :=
∑ s : ConfigSpace V, wJ E J hf s
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:77
def ghsvp_vertexEdges (E : Finset (Sym2 W)) (S : Finset W) : Finset (Sym2 W) :=
E.filter (Sharpness.edgeInside S)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:211
Uses 2 · Used by 1
DefinitionInhomogeneous weight of an agreement set#✓fully proved
The weight of the agreement set A in the duplicated inhomogeneous system.
noncomputable def ghsiSubsetMass (G : SimpleGraph V) [DecidableRel G.Adj]
(K : Sym2 V → ℝ) (hf : V → ℝ) (A : Finset V) : ℝ :=
ghsvp_vertexZ (withGhost G).edgeFinset (ghsiGhostCoupling K hf)
(ghsvp_fieldSet A) *
ghsvp_vertexZ (withGhost G).edgeFinset (ghsiGhostCoupling K hf)
(ghsvp_someSet Aᶜ)
noncomputable def ghsvp_vertexZ (E : Finset (Sym2 W)) (J : Sym2 W -> Real)
(S : Finset W) : Real :=
ZJ (ghsvp_vertexEdges E S) J (fun _ => 0)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:215
def withGhost : SimpleGraph (Option V) where
Adj a b :=
match a, b with
| some x, some y => G.Adj x y
| some _, none => True
| none, some _ => True
| none, none => False
symm := by
constructor
rintro (_ | x) (_ | y) h
· exact h
· exact h
⋯
PhyslibExt/CondensedMatter/IsingModel/Sharpness/RandomCurrent.lean:237
noncomputable def ghsiGhostCoupling (K : Sym2 V → ℝ) (hf : V → ℝ)
(e : Sym2 (Option V)) : ℝ :=
Sym2.lift ⟨fun a b =>
match a, b with
| some x, some y => 2 * K s(x, y)
| some x, none => 2 * hf x
| none, some y => 2 * hf y
| none, none => 0,
by rintro (_ | x) (_ | y) <;> simp only [] <;> rw [Sym2.eq_swap]⟩ e
PhyslibExt/CondensedMatter/IsingModel/GHSInhomVertexPartition.lean:58
def ghsvp_fieldSet (A : Finset V) : Finset (Option V) := insert none (ghsvp_someSet A)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:380
def ghsvp_someSet (A : Finset V) : Finset (Option V) := A.map someEmb
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:377
DefinitionTotal spin of a configuration#✓fully proved
The total spin ∑_x σ_x of a configuration.
noncomputable def totalSpin (s : ConfigSpace V) : ℝ := ∑ x, spin s x
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 2 · Used by 8
LemmaGriffiths' second inequality for two distinct spins#✓fully proved
Griffiths' second inequality for two distinct spins: ⟨σ_x⟩⟨σ_y⟩ ≤ ⟨σ_x σ_y⟩ for β, h ≥ 0.
lemma cov_nonneg (β h : ℝ) (hβ : 0 ≤ β) (hh : 0 ≤ h) (x y : V) (hxy : x ≠ y) :
isingExpectation G β h (fun s => spin s x) * isingExpectation G β h (fun s => spin s y)
≤ isingExpectation G β h (fun s => spin s x * spin s y)
noncomputable def isingExpectation (β h : ℝ) (f : ConfigSpace V → ℝ) : ℝ :=
∑ s : ConfigSpace V, isingProb G β h s * f s
PhyslibExt/CondensedMatter/IsingModel/GKS.lean:208
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 5 · Used by 4
LemmaField derivative of the magnetization as a sum of covariances#✓fully proved
The field derivative of the magnetization is β times the sum of the two-point covariances
⟨σ_o σ_x⟩ - ⟨σ_o⟩⟨σ_x⟩ (the finite-volume susceptibility).
lemma deriv_magnetization_eq_sum_cov (β h : ℝ) (o : V) :
deriv (fun h => isingExpectation G β h (fun s => spin s o)) h
= β * ∑ x, (isingExpectation G β h (fun s => spin s o * spin s x)
- isingExpectation G β h (fun s => spin s o) * isingExpectation G β h (fun s => spin s
x))
noncomputable def isingExpectation (β h : ℝ) (f : ConfigSpace V → ℝ) : ℝ :=
∑ s : ConfigSpace V, isingProb G β h s * f s
PhyslibExt/CondensedMatter/IsingModel/GKS.lean:208
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 10 · Used by 4
DefinitionGlobal spin flip#✓fully proved
The global spin flip σ ↦ -σ.
def flipCfg (s : ConfigSpace V) : ConfigSpace V := fun x => ! s x
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
Uses 1 · Used by 2
DefinitionGlobal spin flip as an involutive equivalence#✓fully proved
The global spin flip as an involutive equivalence of configuration space.
def flipEquiv : ConfigSpace V ≃ ConfigSpace V where
toFun := flipCfg
invFun := flipCfg
left_inv := by intro s; funext x; unfold flipCfg; cases s x <;> rfl
right_inv := by intro s; funext x; unfold flipCfg; cases s x <;> rfl
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def flipCfg (s : ConfigSpace V) : ConfigSpace V := fun x => ! s x
PhyslibExt/CondensedMatter/IsingModel/Sharpness/MagnetizationFieldDeriv.lean:278
Uses 2 · Used by 1
LemmaMagnetization vanishes at zero field#✓fully proved
At zero field the global spin-flip symmetry forces the magnetization ⟨σ_o⟩_{β,0} to vanish.
lemma magnetization_zero_field (β : ℝ) (o : V) :
isingExpectation G β 0 (fun s => spin s o) = 0
noncomputable def isingExpectation (β h : ℝ) (f : ConfigSpace V → ℝ) : ℝ :=
∑ s : ConfigSpace V, isingProb G β h s * f s
PhyslibExt/CondensedMatter/IsingModel/GKS.lean:208
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 10 · Used by 4
TheoremSusceptibility bound \(\chi(h) \le m(h)/h \le 1/h\) from concavity#✓fully proved
If the magnetization is concave in the field on [0, ∞) (the GHS inequality), then for h > 0
the susceptibility satisfies χ(h) ≤ m(h)/h ≤ 1/h. Reference: R. B. Griffiths, C. A. Hurst, S.
Sherman, "Concavity of magnetization of an Ising ferromagnet in a positive external field", J. Math.
Phys. 11 (1970) 790–795.
theorem ghs_susceptibility_bound (β h : ℝ) (hh : 0 < h) (o : V)
(hconc : ConcaveOn ℝ (Ici 0) (fun h => isingExpectation G β h (fun s => spin s o))) :
deriv (fun h => isingExpectation G β h (fun s => spin s o)) h
≤ isingExpectation G β h (fun s => spin s o) / h
∧ deriv (fun h => isingExpectation G β h (fun s => spin s o)) h ≤ 1 / h
noncomputable def isingExpectation (β h : ℝ) (f : ConfigSpace V → ℝ) : ℝ :=
∑ s : ConfigSpace V, isingProb G β h s * f s
PhyslibExt/CondensedMatter/IsingModel/GKS.lean:208
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 12 · Used by 1
DefinitionAgreement configuration of two spin configurations#✓fully proved
The agreement configuration of a and b: true exactly at the sites where they agree.
def ghsLAgreement (a b : ConfigSpace V) : ConfigSpace V := fun v => decide (a v = b v)
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
DefinitionMate of a configuration over an agreement pattern#✓fully proved
The configuration that agrees with a where q is true and disagrees where q is
false.
def ghsLMate (q a : ConfigSpace V) : ConfigSpace V := fun v => if q v then a v else !a v
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
Uses 1 · Used by 9
DefinitionSet of sites where an agreement configuration is true#✓fully proved
The set of sites where q is true.
def ghsLAgreeFinset (q : ConfigSpace V) : Finset V :=
Finset.univ.filter (fun v => q v = true)
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
Uses 1 · Used by 10
DefinitionAgreement-set Boltzmann weight with doubled couplings and field#✓fully proved
The Boltzmann weight, with couplings 2β and field 2βh, of a on the subgraph induced by
the agreement set of q.
noncomputable def ghsLTWeight (β h : ℝ) (q a : ConfigSpace V) : ℝ :=
Real.exp (2 * β *
((∑ e ∈ ghsvp_vertexEdges G.edgeFinset (ghsLAgreeFinset q), bond a e)
+ h * ∑ v ∈ ghsLAgreeFinset q, spin a v))
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def ghsvp_vertexEdges (E : Finset (Sym2 W)) (S : Finset W) : Finset (Sym2 W) :=
E.filter (Sharpness.edgeInside S)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:211
def ghsLAgreeFinset (q : ConfigSpace V) : Finset V :=
Finset.univ.filter (fun v => q v = true)
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:165
def bond (s : ConfigSpace V) (e : Sym2 V) : ℝ :=
Sym2.lift ⟨fun x y => spin s x * spin s y, fun _ _ => mul_comm _ _⟩ e
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:85
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
DefinitionZero-field disagreement-set Boltzmann weight#✓fully proved
The zero-field Boltzmann weight, with couplings 2β, of a on the subgraph induced by the
disagreement set of q.
noncomputable def ghsLUWeight (β : ℝ) (q a : ConfigSpace V) : ℝ :=
Real.exp (2 * β *
∑ e ∈ ghsvp_vertexEdges G.edgeFinset (ghsLAgreeFinset q)ᶜ, bond a e)
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def ghsvp_vertexEdges (E : Finset (Sym2 W)) (S : Finset W) : Finset (Sym2 W) :=
E.filter (Sharpness.edgeInside S)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:211
def ghsLAgreeFinset (q : ConfigSpace V) : Finset V :=
Finset.univ.filter (fun v => q v = true)
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:165
def bond (s : ConfigSpace V) (e : Sym2 V) : ℝ :=
Sym2.lift ⟨fun x y => spin s x * spin s y, fun _ _ => mul_comm _ _⟩ e
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:85
DefinitionSplitting a configuration along the agreement set#✓fully proved
Configurations split into their restrictions to the agreement and disagreement sets of q.
def ghsLSplitEquiv (q : ConfigSpace V) :
ConfigSpace V ≃
((v : {v : V // q v = true}) → Bool) ×
((v : {v : V // ¬ q v = true}) → Bool) :=
Equiv.piEquivPiSubtypeProd (fun v => q v = true) (fun _ => Bool)
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
Uses 1 · Used by 6
DefinitionCompletion of agreement-set spins to a full configuration#✓fully proved
The configuration with given spins on the agreement set of q and false elsewhere.
def ghsLTComplete (q : ConfigSpace V)
(st : (v : {v : V // q v = true}) → Bool) : ConfigSpace V :=
(ghsLSplitEquiv q).symm (st, fun _ => false)
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def ghsLSplitEquiv (q : ConfigSpace V) :
ConfigSpace V ≃
((v : {v : V // q v = true}) → Bool) ×
((v : {v : V // ¬ q v = true}) → Bool) :=
Equiv.piEquivPiSubtypeProd (fun v => q v = true) (fun _ => Bool)
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:282
Uses 2 · Used by 4
DefinitionCompletion of disagreement-set spins to a full configuration#✓fully proved
The configuration with given spins on the disagreement set of q and false elsewhere.
def ghsLUComplete (q : ConfigSpace V)
(su : (v : {v : V // ¬ q v = true}) → Bool) : ConfigSpace V :=
(ghsLSplitEquiv q).symm (fun _ => false, su)
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def ghsLSplitEquiv (q : ConfigSpace V) :
ConfigSpace V ≃
((v : {v : V // q v = true}) → Bool) ×
((v : {v : V // ¬ q v = true}) → Bool) :=
Equiv.piEquivPiSubtypeProd (fun v => q v = true) (fun _ => Bool)
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:282
Uses 2 · Used by 4
DefinitionChange of variables from replica pairs to agreement and first copy#✓fully proved
The change of variables (a, b) ↦ (agreement of a and b, a) on pairs of configurations.
def ghsLXnorEquiv :
(ConfigSpace V × ConfigSpace V) ≃ (ConfigSpace V × ConfigSpace V) where
toFun p := (ghsLAgreement p.1 p.2, p.1)
invFun p := (p.2, ghsLMate p.1 p.2)
left_inv p := by ext <;> simp
right_inv p := by ext <;> simp
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def ghsLAgreement (a b : ConfigSpace V) : ConfigSpace V := fun v => decide (a v = b v)
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:132
def ghsLMate (q a : ConfigSpace V) : ConfigSpace V := fun v => if q v then a v else !a v
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:136
Uses 3 · Used by 3
DefinitionTwo-replica weight in an agreement fibre#✓fully proved
The two-replica weight of a and its mate in the fibre over the agreement configuration q.
noncomputable def ghsLFibreWeight (β h : ℝ) (q a : ConfigSpace V) : ℝ :=
isingWeight G β h a * isingWeight G β h (ghsLMate q a)
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def isingWeight (β h : ℝ) (s : ConfigSpace V) : ℝ :=
Real.exp (-β * hamiltonian G h s)
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:112
def ghsLMate (q a : ConfigSpace V) : ConfigSpace V := fun v => if q v then a v else !a v
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:136
Uses 3 · Used by 5
DefinitionTotal two-replica weight of an agreement fibre#✓fully proved
The total two-replica weight of the fibre over q.
noncomputable def ghsLFibreMass (β h : ℝ) (q : ConfigSpace V) : ℝ :=
∑ a : ConfigSpace V, ghsLFibreWeight G β h q a
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsLFibreWeight (β h : ℝ) (q a : ConfigSpace V) : ℝ :=
isingWeight G β h a * isingWeight G β h (ghsLMate q a)
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:448
Uses 2 · Used by 5
DefinitionFibre weight integrated against an observable#✓fully proved
The two-replica weight of the fibre over q, integrated against F.
noncomputable def ghsLFibreNumerator (β h : ℝ)
(F : ConfigSpace V → ConfigSpace V → ℝ) (q : ConfigSpace V) : ℝ :=
∑ a : ConfigSpace V,
ghsLFibreWeight G β h q a * F a (ghsLMate q a)
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsLFibreWeight (β h : ℝ) (q a : ConfigSpace V) : ℝ :=
isingWeight G β h a * isingWeight G β h (ghsLMate q a)
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:448
def ghsLMate (q a : ConfigSpace V) : ConfigSpace V := fun v => if q v then a v else !a v
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:136
Uses 3 · Used by 3
DefinitionConditional two-replica expectation given the agreement#✓fully proved
The conditional two-replica expectation of F given the agreement configuration q.
noncomputable def ghsLFibreExpectation (β h : ℝ)
(F : ConfigSpace V → ConfigSpace V → ℝ) (q : ConfigSpace V) : ℝ :=
ghsLFibreNumerator G β h F q / ghsLFibreMass G β h q
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsLFibreNumerator (β h : ℝ)
(F : ConfigSpace V → ConfigSpace V → ℝ) (q : ConfigSpace V) : ℝ :=
∑ a : ConfigSpace V,
ghsLFibreWeight G β h q a * F a (ghsLMate q a)
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:456
noncomputable def ghsLFibreMass (β h : ℝ) (q : ConfigSpace V) : ℝ :=
∑ a : ConfigSpace V, ghsLFibreWeight G β h q a
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:452
Uses 3 · Used by 5
DefinitionLaw of the agreement configuration of two replicas#✓fully proved
The probability that two independent replicas have agreement configuration q.
noncomputable def ghsLAgreementProb (β h : ℝ) (q : ConfigSpace V) : ℝ :=
ghsLFibreMass G β h q / (isingZ G β h) ^ 2
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsLFibreMass (β h : ℝ) (q : ConfigSpace V) : ℝ :=
∑ a : ConfigSpace V, ghsLFibreWeight G β h q a
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:452
noncomputable def isingZ (β h : ℝ) : ℝ := ∑ s : ConfigSpace V, isingWeight G β h s
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:126
Uses 3 · Used by 3
DefinitionConditional expectation of \(u_o u_x\) given the agreement#✓fully proved
The conditional expectation E[u_o u_x | q] of the rotated variables.
noncomputable def ghsLFibreUU (β h : ℝ) (o x : V) (q : ConfigSpace V) : ℝ :=
ghsLFibreExpectation G β h (fun a b => uvar a b o * uvar a b x) q
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsLFibreExpectation (β h : ℝ)
(F : ConfigSpace V → ConfigSpace V → ℝ) (q : ConfigSpace V) : ℝ :=
ghsLFibreNumerator G β h F q / ghsLFibreMass G β h q
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:462
noncomputable def uvar (a b : ConfigSpace V) (w : V) : ℝ := (spin a w - spin b w) / 2
PhyslibExt/CondensedMatter/IsingModel/GHSDuplicateVariable.lean:328
Uses 3 · Used by 3
DefinitionConditional expectation of \(t_y\) given the agreement#✓fully proved
The conditional expectation E[t_y | q] of the rotated variable.
noncomputable def ghsLFibreT (β h : ℝ) (y : V) (q : ConfigSpace V) : ℝ :=
ghsLFibreExpectation G β h (fun a b => tvar a b y) q
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsLFibreExpectation (β h : ℝ)
(F : ConfigSpace V → ConfigSpace V → ℝ) (q : ConfigSpace V) : ℝ :=
ghsLFibreNumerator G β h F q / ghsLFibreMass G β h q
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:462
noncomputable def tvar (a b : ConfigSpace V) (w : V) : ℝ := (spin a w + spin b w) / 2
PhyslibExt/CondensedMatter/IsingModel/GHSDuplicateVariable.lean:331
Uses 3 · Used by 3
DefinitionZero-field expectation on the disagreement set#✓fully proved
The zero-field expectation on the subgraph induced by the disagreement set of q.
noncomputable def ghsLUExpectation (β : ℝ) (q : ConfigSpace V)
(F : ConfigSpace V → ℝ) : ℝ :=
(∑ a : ConfigSpace V, ghsLUWeight G β q a * F a) /
∑ a : ConfigSpace V, ghsLUWeight G β q a
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsLUWeight (β : ℝ) (q a : ConfigSpace V) : ℝ :=
Real.exp (2 * β *
∑ e ∈ ghsvp_vertexEdges G.edgeFinset (ghsLAgreeFinset q)ᶜ, bond a e)
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:254
Uses 2 · Used by 2
DefinitionExpectation with field on the agreement set#✓fully proved
The expectation with field on the subgraph induced by the agreement set of q.
noncomputable def ghsLTExpectation (β h : ℝ) (q : ConfigSpace V)
(F : ConfigSpace V → ℝ) : ℝ :=
(∑ a : ConfigSpace V, ghsLTWeight G β h q a * F a) /
∑ a : ConfigSpace V, ghsLTWeight G β h q a
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsLTWeight (β h : ℝ) (q a : ConfigSpace V) : ℝ :=
Real.exp (2 * β *
((∑ e ∈ ghsvp_vertexEdges G.edgeFinset (ghsLAgreeFinset q), bond a e)
+ h * ∑ v ∈ ghsLAgreeFinset q, spin a v))
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:247
Uses 2 · Used by 2
DefinitionField weight of the disagreement set#✓fully proved
The field weight exp (2βh ∑_{v ∉ q} σ_v(a)) of the disagreement set.
noncomputable def ghsLOutFieldWeight (β h : ℝ) (q a : ConfigSpace V) : ℝ :=
Real.exp (2 * β * h * ∑ v ∈ (ghsLAgreeFinset q)ᶜ, spin a v)
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def ghsLAgreeFinset (q : ConfigSpace V) : Finset V :=
Finset.univ.filter (fun v => q v = true)
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:165
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
DefinitionFibre factorization property for \(u_o u_x\) and \(t_y\)#✓fully proved
Conditioned on the agreement configuration, u_o u_x and t_y are uncorrelated.
def GHSLebowitzFibreFactorization (β h : ℝ) (o x y : V) : Prop :=
∀ q : ConfigSpace V,
ghsLFibreExpectation G β h
(fun a b => uvar a b o * uvar a b x * tvar a b y) q
= ghsLFibreUU G β h o x q * ghsLFibreT G β h y q
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsLFibreExpectation (β h : ℝ)
(F : ConfigSpace V → ConfigSpace V → ℝ) (q : ConfigSpace V) : ℝ :=
ghsLFibreNumerator G β h F q / ghsLFibreMass G β h q
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:462
noncomputable def uvar (a b : ConfigSpace V) (w : V) : ℝ := (spin a w - spin b w) / 2
PhyslibExt/CondensedMatter/IsingModel/GHSDuplicateVariable.lean:328
noncomputable def tvar (a b : ConfigSpace V) (w : V) : ℝ := (spin a w + spin b w) / 2
PhyslibExt/CondensedMatter/IsingModel/GHSDuplicateVariable.lean:331
noncomputable def ghsLFibreUU (β h : ℝ) (o x : V) (q : ConfigSpace V) : ℝ :=
ghsLFibreExpectation G β h (fun a b => uvar a b o * uvar a b x) q
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:663
noncomputable def ghsLFibreT (β h : ℝ) (y : V) (q : ConfigSpace V) : ℝ :=
ghsLFibreExpectation G β h (fun a b => tvar a b y) q
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:667
Uses 6 · Used by 2
LemmaFKG lattice condition for the agreement law when \(\beta, h \ge 0\)#✓fully proved
The law of the agreement configuration satisfies the FKG lattice condition.
lemma ghsLAgreementProb_fkg (β h : ℝ) (hβ : 0 ≤ β) (hh : 0 ≤ h) :
FKGLatticeCondition (ghsLAgreementProb G β h)
def FKGLatticeCondition (π : ConfigSpace E → ℝ) : Prop :=
∀ ω ω', π ω * π ω' ≤ π (ω ⊔ ω') * π (ω ⊓ ω')
MathlibExt/Probability/CorrelationInequalities/IncreasingEvent.lean:111
noncomputable def ghsLAgreementProb (β h : ℝ) (q : ConfigSpace V) : ℝ :=
ghsLFibreMass G β h q / (isingZ G β h) ^ 2
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:467
Uses 40
TheoremGriffiths–Hurst–Sherman inequality#✓fully proved
For the ferromagnetic Ising model on a finite graph
with β ≥ 0 and h ≥ 0, the third Ursell function u₃(o, x, y) is nonpositive for every
site o and every edge {x, y}.
R. B. Griffiths, C. A. Hurst and S. Sherman, Concavity of magnetization of an Ising ferromagnet in a positive external field, J. Math. Phys. 11 (1970); proof of J. L. Lebowitz, GHS and other inequalities, Comm. Math. Phys. 35 (1974).
theorem ghsL_threePoint (β h : ℝ) (hβ : 0 ≤ β) (hh : 0 ≤ h) (o : V) :
GHSThreePointSym G β h o
def GHSThreePointSym (β h : ℝ) (o : V) : Prop :=
∀ e ∈ G.edgeFinset, cov3sym G β h o e ≤ ghsBoundSym G β h o e
PhyslibExt/CondensedMatter/IsingModel/AizenmanBarsky.lean:287
Uses 69
LemmaDerivative of a concave function is at most a secant slope#✓fully proved
The derivative of a concave function on [0, ∞) at b is at most the secant slope over [a,
b]. This is Mathlib's ConcaveOn.deriv_le_slope with the slope unfolded.
LemmaConcavity from a nonpositive second derivative on \((0, \infty)\)#✓fully proved
A continuous function on [0, ∞), twice differentiable on the interior with nonpositive second
derivative, is concave. This is Mathlib's concaveOn_of_deriv2_nonpos specialized to Ici 0.
lemma shg_concaveOn_of_deriv2_nonpos (φ : ℝ → ℝ)
(hcont : ContinuousOn φ (Ici 0))
(hf' : DifferentiableOn ℝ φ (interior (Ici 0)))
(hf'' : DifferentiableOn ℝ (deriv φ) (interior (Ici 0)))
(hu3 : ∀ h ∈ interior (Ici (0 : ℝ)), deriv^[2] φ h ≤ 0) :
ConcaveOn ℝ (Ici 0) φ
Used by 2
LemmaSusceptibility bound \(\chi(h) \le m(h)/h\) from concavity#✓fully proved
If the magnetization is concave in the field on [0, ∞), then for h > 0 the susceptibility
satisfies χ(h) ≤ m(h)/h, using m(0) = 0.
lemma shg_magnetization_susceptibility_bound (β h : ℝ) (hh : 0 < h) (o : V)
(hconc : ConcaveOn ℝ (Ici 0) (fun h => isingExpectation G β h (fun s => spin s o))) :
deriv (fun h => isingExpectation G β h (fun s => spin s o)) h
≤ isingExpectation G β h (fun s => spin s o) / h
noncomputable def isingExpectation (β h : ℝ) (f : ConfigSpace V → ℝ) : ℝ :=
∑ s : ConfigSpace V, isingProb G β h s * f s
PhyslibExt/CondensedMatter/IsingModel/GKS.lean:208
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 12
DefinitionTwo-replica expectation with inhomogeneous couplings#✓fully proved
The expectation of F a b over two independent copies of the Ising model with couplings K
and fields hf.
noncomputable def ghsiExp2 (K : Sym2 V → ℝ) (hf : V → ℝ)
(F : ConfigSpace V → ConfigSpace V → ℝ) : ℝ :=
(∑ a : ConfigSpace V, ∑ b : ConfigSpace V,
wJ G.edgeFinset K hf a * wJ G.edgeFinset K hf b * F a b) /
(ZJ G.edgeFinset K hf) ^ 2
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def wJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ)
(s : ConfigSpace V) : ℝ :=
Real.exp ((∑ e ∈ E, J e * bond s e) + ∑ x, hf x * spin s x)
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:64
noncomputable def ZJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ) : ℝ :=
∑ s : ConfigSpace V, wJ E J hf s
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:77
DefinitionInhomogeneous agreement-set Boltzmann weight#✓fully proved
The Boltzmann weight, with couplings 2K and fields 2hf, of a on the subgraph induced by
the agreement set of q.
noncomputable def ghsiTWeight (K : Sym2 V → ℝ) (hf : V → ℝ)
(q a : ConfigSpace V) : ℝ :=
Real.exp
((∑ e ∈ ghsvp_vertexEdges G.edgeFinset (ghsLAgreeFinset q),
(2 * K e) * bond a e) +
∑ v ∈ ghsLAgreeFinset q, (2 * hf v) * spin a v)
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def ghsvp_vertexEdges (E : Finset (Sym2 W)) (S : Finset W) : Finset (Sym2 W) :=
E.filter (Sharpness.edgeInside S)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:211
def ghsLAgreeFinset (q : ConfigSpace V) : Finset V :=
Finset.univ.filter (fun v => q v = true)
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:165
def bond (s : ConfigSpace V) (e : Sym2 V) : ℝ :=
Sym2.lift ⟨fun x y => spin s x * spin s y, fun _ _ => mul_comm _ _⟩ e
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:85
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
DefinitionInhomogeneous zero-field disagreement-set weight#✓fully proved
The zero-field Boltzmann weight, with couplings 2K, of a on the subgraph induced by the
disagreement set of q.
noncomputable def ghsiUWeight (K : Sym2 V → ℝ) (q a : ConfigSpace V) : ℝ :=
Real.exp
(∑ e ∈ ghsvp_vertexEdges G.edgeFinset (ghsLAgreeFinset q)ᶜ,
(2 * K e) * bond a e)
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def ghsvp_vertexEdges (E : Finset (Sym2 W)) (S : Finset W) : Finset (Sym2 W) :=
E.filter (Sharpness.edgeInside S)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:211
def ghsLAgreeFinset (q : ConfigSpace V) : Finset V :=
Finset.univ.filter (fun v => q v = true)
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:165
def bond (s : ConfigSpace V) (e : Sym2 V) : ℝ :=
Sym2.lift ⟨fun x y => spin s x * spin s y, fun _ _ => mul_comm _ _⟩ e
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:85
DefinitionInhomogeneous two-replica weight in an agreement fibre#✓fully proved
The two-replica weight of a and its mate in the fibre over the agreement configuration q.
noncomputable def ghsiFibreWeight (K : Sym2 V → ℝ) (hf : V → ℝ)
(q a : ConfigSpace V) : ℝ :=
wJ G.edgeFinset K hf a * wJ G.edgeFinset K hf (ghsLMate q a)
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def wJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ)
(s : ConfigSpace V) : ℝ :=
Real.exp ((∑ e ∈ E, J e * bond s e) + ∑ x, hf x * spin s x)
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:64
def ghsLMate (q a : ConfigSpace V) : ConfigSpace V := fun v => if q v then a v else !a v
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:136
DefinitionInhomogeneous total weight of an agreement fibre#✓fully proved
The total two-replica weight of the fibre over q.
noncomputable def ghsiFibreMass (K : Sym2 V → ℝ) (hf : V → ℝ)
(q : ConfigSpace V) : ℝ :=
∑ a : ConfigSpace V, ghsiFibreWeight G K hf q a
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsiFibreWeight (K : Sym2 V → ℝ) (hf : V → ℝ)
(q a : ConfigSpace V) : ℝ :=
wJ G.edgeFinset K hf a * wJ G.edgeFinset K hf (ghsLMate q a)
PhyslibExt/CondensedMatter/IsingModel/GHSInhomogeneous.lean:371
DefinitionInhomogeneous fibre weight integrated against an observable#✓fully proved
The two-replica weight of the fibre over q, integrated against F.
noncomputable def ghsiFibreNumerator (K : Sym2 V → ℝ) (hf : V → ℝ)
(F : ConfigSpace V → ConfigSpace V → ℝ) (q : ConfigSpace V) : ℝ :=
∑ a : ConfigSpace V, ghsiFibreWeight G K hf q a * F a (ghsLMate q a)
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsiFibreWeight (K : Sym2 V → ℝ) (hf : V → ℝ)
(q a : ConfigSpace V) : ℝ :=
wJ G.edgeFinset K hf a * wJ G.edgeFinset K hf (ghsLMate q a)
PhyslibExt/CondensedMatter/IsingModel/GHSInhomogeneous.lean:371
def ghsLMate (q a : ConfigSpace V) : ConfigSpace V := fun v => if q v then a v else !a v
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:136
DefinitionInhomogeneous conditional expectation given the agreement#✓fully proved
The conditional two-replica expectation of F given the agreement configuration q.
noncomputable def ghsiFibreExpectation (K : Sym2 V → ℝ) (hf : V → ℝ)
(F : ConfigSpace V → ConfigSpace V → ℝ) (q : ConfigSpace V) : ℝ :=
ghsiFibreNumerator G K hf F q / ghsiFibreMass G K hf q
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsiFibreNumerator (K : Sym2 V → ℝ) (hf : V → ℝ)
(F : ConfigSpace V → ConfigSpace V → ℝ) (q : ConfigSpace V) : ℝ :=
∑ a : ConfigSpace V, ghsiFibreWeight G K hf q a * F a (ghsLMate q a)
PhyslibExt/CondensedMatter/IsingModel/GHSInhomogeneous.lean:381
noncomputable def ghsiFibreMass (K : Sym2 V → ℝ) (hf : V → ℝ)
(q : ConfigSpace V) : ℝ :=
∑ a : ConfigSpace V, ghsiFibreWeight G K hf q a
PhyslibExt/CondensedMatter/IsingModel/GHSInhomogeneous.lean:376
DefinitionInhomogeneous law of the agreement configuration#✓fully proved
The probability that two independent replicas have agreement configuration q.
noncomputable def ghsiAgreementProb (K : Sym2 V → ℝ) (hf : V → ℝ)
(q : ConfigSpace V) : ℝ :=
ghsiFibreMass G K hf q / (ZJ G.edgeFinset K hf) ^ 2
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsiFibreMass (K : Sym2 V → ℝ) (hf : V → ℝ)
(q : ConfigSpace V) : ℝ :=
∑ a : ConfigSpace V, ghsiFibreWeight G K hf q a
PhyslibExt/CondensedMatter/IsingModel/GHSInhomogeneous.lean:376
noncomputable def ZJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ) : ℝ :=
∑ s : ConfigSpace V, wJ E J hf s
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:77
DefinitionInhomogeneous conditional expectation of \(u_o u_x\)#✓fully proved
The conditional expectation E[u_o u_x | q] of the rotated variables.
noncomputable def ghsiFibreUU (K : Sym2 V → ℝ) (hf : V → ℝ)
(o x : V) (q : ConfigSpace V) : ℝ :=
ghsiFibreExpectation G K hf (fun a b => uvar a b o * uvar a b x) q
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsiFibreExpectation (K : Sym2 V → ℝ) (hf : V → ℝ)
(F : ConfigSpace V → ConfigSpace V → ℝ) (q : ConfigSpace V) : ℝ :=
ghsiFibreNumerator G K hf F q / ghsiFibreMass G K hf q
PhyslibExt/CondensedMatter/IsingModel/GHSInhomogeneous.lean:386
noncomputable def uvar (a b : ConfigSpace V) (w : V) : ℝ := (spin a w - spin b w) / 2
PhyslibExt/CondensedMatter/IsingModel/GHSDuplicateVariable.lean:328
DefinitionInhomogeneous conditional expectation of \(t_y\)#✓fully proved
The conditional expectation E[t_y | q] of the rotated variable.
noncomputable def ghsiFibreT (K : Sym2 V → ℝ) (hf : V → ℝ)
(y : V) (q : ConfigSpace V) : ℝ :=
ghsiFibreExpectation G K hf (fun a b => tvar a b y) q
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsiFibreExpectation (K : Sym2 V → ℝ) (hf : V → ℝ)
(F : ConfigSpace V → ConfigSpace V → ℝ) (q : ConfigSpace V) : ℝ :=
ghsiFibreNumerator G K hf F q / ghsiFibreMass G K hf q
PhyslibExt/CondensedMatter/IsingModel/GHSInhomogeneous.lean:386
noncomputable def tvar (a b : ConfigSpace V) (w : V) : ℝ := (spin a w + spin b w) / 2
PhyslibExt/CondensedMatter/IsingModel/GHSDuplicateVariable.lean:331
Uses 3 · Used by 1
DefinitionInhomogeneous zero-field expectation on the disagreement set#✓fully proved
The zero-field expectation on the subgraph induced by the disagreement set of q.
noncomputable def ghsiUExpectation (K : Sym2 V → ℝ) (q : ConfigSpace V)
(F : ConfigSpace V → ℝ) : ℝ :=
(∑ a : ConfigSpace V, ghsiUWeight G K q a * F a) /
∑ a : ConfigSpace V, ghsiUWeight G K q a
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsiUWeight (K : Sym2 V → ℝ) (q a : ConfigSpace V) : ℝ :=
Real.exp
(∑ e ∈ ghsvp_vertexEdges G.edgeFinset (ghsLAgreeFinset q)ᶜ,
(2 * K e) * bond a e)
PhyslibExt/CondensedMatter/IsingModel/GHSInhomogeneous.lean:245
Uses 2 · Used by 1
DefinitionInhomogeneous expectation with field on the agreement set#✓fully proved
The expectation with field on the subgraph induced by the agreement set of q.
noncomputable def ghsiTExpectation (K : Sym2 V → ℝ) (hf : V → ℝ)
(q : ConfigSpace V) (F : ConfigSpace V → ℝ) : ℝ :=
(∑ a : ConfigSpace V, ghsiTWeight G K hf q a * F a) /
∑ a : ConfigSpace V, ghsiTWeight G K hf q a
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def ghsiTWeight (K : Sym2 V → ℝ) (hf : V → ℝ)
(q a : ConfigSpace V) : ℝ :=
Real.exp
((∑ e ∈ ghsvp_vertexEdges G.edgeFinset (ghsLAgreeFinset q),
(2 * K e) * bond a e) +
∑ v ∈ ghsLAgreeFinset q, (2 * hf v) * spin a v)
PhyslibExt/CondensedMatter/IsingModel/GHSInhomogeneous.lean:236
Uses 2 · Used by 1
DefinitionInhomogeneous two-point correlation on the disagreement set#✓fully proved
The zero-field correlation ⟨σ_o σ_x⟩ on the subgraph induced by Aᶜ (couplings 2K), or
0 if o or x lies in A.
noncomputable def ghsiDisagreementCorr (K : Sym2 V → ℝ)
(o x : V) (A : Finset V) : ℝ :=
if o ∈ A ∨ x ∈ A then 0
else ghsvp_vertexCorr G.edgeFinset (fun e => 2 * K e) Aᶜ {o, x}
noncomputable def ghsvp_vertexCorr (E : Finset (Sym2 W)) (J : Sym2 W -> Real)
(S P : Finset W) : Real :=
expJ (ghsvp_vertexEdges E S) J (fun _ => 0) (spinProd P)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:259
Uses 1 · Used by 1
DefinitionInhomogeneous magnetization on the agreement set#✓fully proved
The magnetization ⟨σ_y⟩ on the subgraph induced by A (couplings 2K, fields 2hf), or
0 if y ∉ A.
noncomputable def ghsiAgreementMag (K : Sym2 V → ℝ) (hf : V → ℝ)
(y : V) (A : Finset V) : ℝ :=
if y ∈ A then
ghsvp_vertexCorrField G.edgeFinset (fun e => 2 * K e) (fun x => 2 * hf x) A {y}
else 0
noncomputable def ghsvp_vertexCorrField (E : Finset (Sym2 W)) (J : Sym2 W -> Real)
(hf : W -> Real) (S P : Finset W) : Real :=
expJ (ghsvp_vertexEdges E S) J hf (spinProd P)
PhyslibExt/CondensedMatter/IsingModel/GHSVertexPartition.lean:264
Uses 1 · Used by 1
DefinitionInhomogeneous field weight of the disagreement set#✓fully proved
The field weight exp (∑_{v ∉ q} 2h_v σ_v(a)) of the disagreement set.
noncomputable def ghsiOutFieldWeight (hf : V → ℝ)
(q a : ConfigSpace V) : ℝ :=
Real.exp (∑ v ∈ (ghsLAgreeFinset q)ᶜ, (2 * hf v) * spin a v)
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def ghsLAgreeFinset (q : ConfigSpace V) : Finset V :=
Finset.univ.filter (fun v => q v = true)
PhyslibExt/CondensedMatter/IsingModel/GHSLebowitz.lean:165
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 3 · Used by 1
TheoremGriffiths–Hurst–Sherman inequality#✓fully proved
Griffiths–Hurst–Sherman inequality for inhomogeneous ferromagnets. With couplings K_e ≥
0 and fields h_x ≥ 0, the third Ursell function satisfies u₃(o, x, y) ≤ 0 whenever o ≠
x.
R. B. Griffiths, C. A. Hurst and S. Sherman, Concavity of magnetization of an Ising ferromagnet in a positive external field, J. Math. Phys. 11 (1970); J. L. Lebowitz, GHS and other inequalities, Comm. Math. Phys. 35 (1974).
theorem ghsi_ursell_nonpos (K : Sym2 V → ℝ) (hf : V → ℝ)
(hK : ∀ e, 0 ≤ K e) (hhf : ∀ x, 0 ≤ hf x)
(o x y : V) (hox : o ≠ x) :
expJ G.edgeFinset K hf (fun s => spin s o * (spin s x * spin s y)) -
expJ G.edgeFinset K hf (fun s => spin s o) *
expJ G.edgeFinset K hf (fun s => spin s x * spin s y) -
expJ G.edgeFinset K hf (fun s => spin s o * spin s x) *
expJ G.edgeFinset K hf (fun s => spin s y) -
expJ G.edgeFinset K hf (fun s => spin s o * spin s y) *
expJ G.edgeFinset K hf (fun s => spin s x) +
2 * expJ G.edgeFinset K hf (fun s => spin s o) *
expJ G.edgeFinset K hf (fun s => spin s x) *
expJ G.edgeFinset K hf (fun s => spin s y) ≤ 0
noncomputable def expJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ)
(f : ConfigSpace V → ℝ) : ℝ :=
(∑ s : ConfigSpace V, f s * wJ E J hf s) / ZJ E J hf
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:85
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 54 · Used by 1
DefinitionDirectional spin sum#✓fully proved
The directional spin ∑_x d_x σ_x.
noncomputable def ghsiDirSpin (dir : V → ℝ) (s : ConfigSpace V) : ℝ :=
∑ x, dir x * spin s x
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 2 · Used by 3
DefinitionField line in a given direction#✓fully proved
The field line t ↦ h + t·d.
def ghsiFieldLine (hf dir : V → ℝ) (t : ℝ) : V → ℝ :=
fun x => hf x + t * dir x
Used by 2
DefinitionCovariance with the directional spin#✓fully proved
The covariance of f with the directional spin ∑_x d_x σ_x.
noncomputable def ghsiDirectionalCovariance (K : Sym2 V → ℝ)
(hf dir : V → ℝ) (f : ConfigSpace V → ℝ) : ℝ :=
expJ G.edgeFinset K hf (fun s => f s * ghsiDirSpin dir s) -
expJ G.edgeFinset K hf f * expJ G.edgeFinset K hf (ghsiDirSpin dir)
abbrev ConfigSpace (E : Type*) : Type _ := E → Bool
MathlibExt/Probability/Configuration/ConfigSpace.lean:40
noncomputable def expJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ)
(f : ConfigSpace V → ℝ) : ℝ :=
(∑ s : ConfigSpace V, f s * wJ E J hf s) / ZJ E J hf
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:85
noncomputable def ghsiDirSpin (dir : V → ℝ) (s : ConfigSpace V) : ℝ :=
∑ x, dir x * spin s x
PhyslibExt/CondensedMatter/IsingModel/GHSInhomogeneous.lean:1105
DefinitionThird Ursell function of the inhomogeneous Ising model#✓fully proved
The third Ursell function u₃(o, x, y) of the inhomogeneous Ising model.
noncomputable def ghsiUrsell3 (K : Sym2 V → ℝ) (hf : V → ℝ)
(o x y : V) : ℝ :=
expJ G.edgeFinset K hf (fun s => spin s o * (spin s x * spin s y)) -
expJ G.edgeFinset K hf (fun s => spin s o) *
expJ G.edgeFinset K hf (fun s => spin s x * spin s y) -
expJ G.edgeFinset K hf (fun s => spin s o * spin s x) *
expJ G.edgeFinset K hf (fun s => spin s y) -
expJ G.edgeFinset K hf (fun s => spin s o * spin s y) *
expJ G.edgeFinset K hf (fun s => spin s x) +
2 * expJ G.edgeFinset K hf (fun s => spin s o) *
expJ G.edgeFinset K hf (fun s => spin s x) *
expJ G.edgeFinset K hf (fun s => spin s y)
noncomputable def expJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ)
(f : ConfigSpace V → ℝ) : ℝ :=
(∑ s : ConfigSpace V, f s * wJ E J hf s) / ZJ E J hf
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:85
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
LemmaThird Ursell function nonpositive for \(o \ne x\), \(K, h \ge 0\)#✓fully proved
The GHS inequality u₃(o, x, y) ≤ 0 for the named Ursell function, when o ≠ x.
lemma ghsiUrsell3_nonpos (K : Sym2 V → ℝ) (hf : V → ℝ)
(hK : ∀ e, 0 ≤ K e) (hhf : ∀ x, 0 ≤ hf x)
(o x y : V) (hox : o ≠ x) :
ghsiUrsell3 G K hf o x y ≤ 0
noncomputable def ghsiUrsell3 (K : Sym2 V → ℝ) (hf : V → ℝ)
(o x y : V) : ℝ :=
expJ G.edgeFinset K hf (fun s => spin s o * (spin s x * spin s y)) -
expJ G.edgeFinset K hf (fun s => spin s o) *
expJ G.edgeFinset K hf (fun s => spin s x * spin s y) -
expJ G.edgeFinset K hf (fun s => spin s o * spin s x) *
expJ G.edgeFinset K hf (fun s => spin s y) -
expJ G.edgeFinset K hf (fun s => spin s o * spin s y) *
expJ G.edgeFinset K hf (fun s => spin s x) +
2 * expJ G.edgeFinset K hf (fun s => spin s o) *
expJ G.edgeFinset K hf (fun s => spin s x) *
expJ G.edgeFinset K hf (fun s => spin s y)
PhyslibExt/CondensedMatter/IsingModel/GHSInhomogeneous.lean:1238
Uses 2 · Used by 1
DefinitionSpin covariance in the inhomogeneous model#✓fully proved
The spin covariance ⟨σ_o σ_x⟩ - ⟨σ_o⟩⟨σ_x⟩.
noncomputable def ghsiCovariance (K : Sym2 V → ℝ) (hf : V → ℝ)
(o x : V) : ℝ :=
expJ G.edgeFinset K hf (fun s => spin s o * spin s x) -
expJ G.edgeFinset K hf (fun s => spin s o) *
expJ G.edgeFinset K hf (fun s => spin s x)
noncomputable def expJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ)
(f : ConfigSpace V → ℝ) : ℝ :=
(∑ s : ConfigSpace V, f s * wJ E J hf s) / ZJ E J hf
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:85
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
TheoremGHS concavity#✓fully proved
With couplings K_e ≥ 0, fields h_x ≥ 0, and a direction d ≥ 0 with
d_o = 0, the magnetization ⟨σ_o⟩ is concave in t ≥ 0 along the field line h + t·d.
R. B. Griffiths, C. A. Hurst and S. Sherman, J. Math. Phys. 11 (1970).
theorem ghsi_magnetization_fieldLine_concaveOn
(K : Sym2 V → ℝ) (hf dir : V → ℝ)
(hK : ∀ e, 0 ≤ K e) (hhf : ∀ x, 0 ≤ hf x)
(hdir : ∀ x, 0 ≤ dir x) (o : V) (hdiro : dir o = 0) :
ConcaveOn ℝ (Ici 0)
(fun t => expJ G.edgeFinset K (ghsiFieldLine hf dir t)
(fun s => spin s o))
noncomputable def expJ (E : Finset (Sym2 V)) (J : Sym2 V → ℝ) (hf : V → ℝ)
(f : ConfigSpace V → ℝ) : ℝ :=
(∑ s : ConfigSpace V, f s * wJ E J hf s) / ZJ E J hf
PhyslibExt/CondensedMatter/IsingModel/Sharpness/Simon.lean:85
def ghsiFieldLine (hf dir : V → ℝ) (t : ℝ) : V → ℝ :=
fun x => hf x + t * dir x
PhyslibExt/CondensedMatter/IsingModel/GHSInhomogeneous.lean:1109
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 13 · Used by 1
DefinitionField derivative of the two-point covariance#✓fully proved
The field derivative of the two-point covariance Cov(σ_o, σ_x), written as a combination of
Gibbs expectations involving the total spin.
noncomputable def gc4_D1cov2 (β h : ℝ) (o x : V) : ℝ :=
(β * (isingExpectation G β h (fun s => (spin s o * spin s x) * totalSpin s)
- isingExpectation G β h (fun s => spin s o * spin s x) * isingExpectation G β h totalSpin))
- ((β * (isingExpectation G β h (fun s => spin s o * totalSpin s)
- isingExpectation G β h (fun s => spin s o) * isingExpectation G β h totalSpin))
* isingExpectation G β h (fun s => spin s x)
+ isingExpectation G β h (fun s => spin s o)
* (β * (isingExpectation G β h (fun s => spin s x * totalSpin s)
- isingExpectation G β h (fun s => spin s x) * isingExpectation G β h totalSpin)))
noncomputable def isingExpectation (β h : ℝ) (f : ConfigSpace V → ℝ) : ℝ :=
∑ s : ConfigSpace V, isingProb G β h s * f s
PhyslibExt/CondensedMatter/IsingModel/GKS.lean:208
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
noncomputable def totalSpin (s : ConfigSpace V) : ℝ := ∑ x, spin s x
PhyslibExt/CondensedMatter/IsingModel/Sharpness/MagnetizationFieldDeriv.lean:73
Uses 4 · Used by 1
DefinitionThird Ursell function of the finite Ising model#✓fully proved
The third Ursell function (joint cumulant) u₃(o, x, y) = ⟨σ_oσ_xσ_y⟩ - ⟨σ_o⟩⟨σ_xσ_y⟩ -
⟨σ_oσ_x⟩⟨σ_y⟩ - ⟨σ_oσ_y⟩⟨σ_x⟩ + 2⟨σ_o⟩⟨σ_x⟩⟨σ_y⟩ of the finite Ising model at (β, h).
noncomputable def eg_ursell3 (β h : ℝ) (o x y : V) : ℝ :=
isingExpectation G β h (fun s => spin s o * (spin s x * spin s y))
- isingExpectation G β h (fun s => spin s o)
* isingExpectation G β h (fun s => spin s x * spin s y)
- isingExpectation G β h (fun s => spin s o * spin s x)
* isingExpectation G β h (fun s => spin s y)
- isingExpectation G β h (fun s => spin s o * spin s y)
* isingExpectation G β h (fun s => spin s x)
+ 2 * isingExpectation G β h (fun s => spin s o)
* isingExpectation G β h (fun s => spin s x)
* isingExpectation G β h (fun s => spin s y)
noncomputable def isingExpectation (β h : ℝ) (f : ConfigSpace V → ℝ) : ℝ :=
∑ s : ConfigSpace V, isingProb G β h s * f s
PhyslibExt/CondensedMatter/IsingModel/GKS.lean:208
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 3 · Used by 4
LemmaGHS gap on an edge equals the third Ursell function#✓fully proved
On an edge {x, y}, the gap cov3sym - ghsBoundSym between the three-point covariance and the
GHS bound equals u₃(o, x, y).
lemma ghc_ursellEqGap (β h : ℝ) (o x y : V) :
cov3sym G β h o s(x, y) - ghsBoundSym G β h o s(x, y) = eg_ursell3 G β h o x y
noncomputable def cov3sym (β h : ℝ) (o : V) (e : Sym2 V) : ℝ :=
Sym2.lift ⟨fun x y =>
isingExpectation G β h (fun s => spin s o * (spin s x * spin s y))
- isingExpectation G β h (fun s => spin s o)
* isingExpectation G β h (fun s => spin s x * spin s y), by
intro a b
have h1 : (fun s : ConfigSpace V => spin s o * (spin s a * spin s b))
= (fun s => spin s o * (spin s b * spin s a)) := by funext s; ring
have h2 : (fun s : ConfigSpace V => spin s a * spin s b)
= (fun s => spin s b * spin s a) := by funext s; ring
simp only []; rw [h1, h2]⟩ e
PhyslibExt/CondensedMatter/IsingModel/AizenmanBarsky.lean:172
noncomputable def ghsBoundSym (β h : ℝ) (o : V) (e : Sym2 V) : ℝ :=
Sym2.lift ⟨fun x y => cov2 G β h o x * onePt G β h y + cov2 G β h o y * onePt G β h x,
fun a b => by ring⟩ e
PhyslibExt/CondensedMatter/IsingModel/AizenmanBarsky.lean:191
noncomputable def eg_ursell3 (β h : ℝ) (o x y : V) : ℝ :=
isingExpectation G β h (fun s => spin s o * (spin s x * spin s y))
- isingExpectation G β h (fun s => spin s o)
* isingExpectation G β h (fun s => spin s x * spin s y)
- isingExpectation G β h (fun s => spin s o * spin s x)
* isingExpectation G β h (fun s => spin s y)
- isingExpectation G β h (fun s => spin s o * spin s y)
* isingExpectation G β h (fun s => spin s x)
+ 2 * isingExpectation G β h (fun s => spin s o)
* isingExpectation G β h (fun s => spin s x)
* isingExpectation G β h (fun s => spin s y)
PhyslibExt/CondensedMatter/IsingModel/Sharpness/UrsellGap.lean:72
Uses 8 · Used by 1
DefinitionLebowitz four-point combination#✓fully proved
Lebowitz's four-point combination ⟨σ_oσ_xσ_y⟩ - ⟨σ_oσ_x⟩⟨σ_y⟩ - ⟨σ_oσ_y⟩⟨σ_x⟩ - ⟨σ_o⟩⟨σ_xσ_y⟩,
which differs from the Ursell function u₃ by 2⟨σ_o⟩⟨σ_x⟩⟨σ_y⟩.
noncomputable def gc_lebowitzU4 (β h : ℝ) (o x y : V) : ℝ :=
isingExpectation G β h (fun s => spin s o * (spin s x * spin s y))
- isingExpectation G β h (fun s => spin s o * spin s x)
* isingExpectation G β h (fun s => spin s y)
- isingExpectation G β h (fun s => spin s o * spin s y)
* isingExpectation G β h (fun s => spin s x)
- isingExpectation G β h (fun s => spin s o)
* isingExpectation G β h (fun s => spin s x * spin s y)
noncomputable def isingExpectation (β h : ℝ) (f : ConfigSpace V → ℝ) : ℝ :=
∑ s : ConfigSpace V, isingProb G β h s * f s
PhyslibExt/CondensedMatter/IsingModel/GKS.lean:208
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
DefinitionGHS sign condition#✓fully proved
The GHS sign condition at (β, h) and root o: the third Ursell function u₃(o, x, y) is
nonpositive for all pairwise distinct o, x, y. It holds for β, h ≥ 0 by ghs_signDominance.
def GHSSignDominance (β h : ℝ) (o : V) : Prop :=
∀ x y : V, o ≠ x → o ≠ y → x ≠ y → eg_ursell3 G β h o x y ≤ 0
noncomputable def eg_ursell3 (β h : ℝ) (o x y : V) : ℝ :=
isingExpectation G β h (fun s => spin s o * (spin s x * spin s y))
- isingExpectation G β h (fun s => spin s o)
* isingExpectation G β h (fun s => spin s x * spin s y)
- isingExpectation G β h (fun s => spin s o * spin s x)
* isingExpectation G β h (fun s => spin s y)
- isingExpectation G β h (fun s => spin s o * spin s y)
* isingExpectation G β h (fun s => spin s x)
+ 2 * isingExpectation G β h (fun s => spin s o)
* isingExpectation G β h (fun s => spin s x)
* isingExpectation G β h (fun s => spin s y)
PhyslibExt/CondensedMatter/IsingModel/Sharpness/UrsellGap.lean:72
Uses 1 · Used by 2
TheoremGriffiths–Hurst–Sherman inequality#✓fully proved
The Griffiths–Hurst–Sherman inequality: for β ≥ 0 and h ≥ 0, the third Ursell function
u₃(o, x, y) is nonpositive for distinct sites. Reference: R. B. Griffiths, C. A. Hurst, S.
Sherman, "Concavity of magnetization of an Ising ferromagnet in a positive external field", J. Math.
Phys. 11 (1970) 790–795; proof via J. L. Lebowitz, "GHS and other inequalities", Comm. Math. Phys.
35 (1974) 87–92.
theorem ghs_signDominance (β h : ℝ) (hβ : 0 ≤ β) (hh : 0 ≤ h) (o : V) :
GHSSignDominance G β h o
def GHSSignDominance (β h : ℝ) (o : V) : Prop :=
∀ x y : V, o ≠ x → o ≠ y → x ≠ y → eg_ursell3 G β h o x y ≤ 0
PhyslibExt/CondensedMatter/IsingModel/Sharpness/GHSSignCondition.lean:113
Uses 68 · Used by 1
TheoremGHS concavity of the magnetization in the field#✓fully proved
GHS concavity: for β ≥ 0 the magnetization h ↦ ⟨σ_o⟩_{β,h} is concave on [0, ∞).
Reference: R. B. Griffiths, C. A. Hurst, S. Sherman, "Concavity of magnetization of an Ising
ferromagnet in a positive external field", J. Math. Phys. 11 (1970) 790–795.
theorem ghs_magnetization_concaveOn (β : ℝ) (hβ : 0 ≤ β) (o : V) :
ConcaveOn ℝ (Ici 0) (fun h => isingExpectation G β h (fun s => spin s o))
noncomputable def isingExpectation (β h : ℝ) (f : ConfigSpace V → ℝ) : ℝ :=
∑ s : ConfigSpace V, isingProb G β h s * f s
PhyslibExt/CondensedMatter/IsingModel/GKS.lean:208
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Uses 25 · Used by 1
TheoremGHS susceptibility bound \(\chi \le 1/h\)#✓fully proved
The GHS susceptibility bound: for β ≥ 0 and h > 0, the field derivative of the magnetization
equals β ∑_x (⟨σ_oσ_x⟩ - ⟨σ_o⟩⟨σ_x⟩) and is at most 1/h. Reference: R. B. Griffiths, C. A.
Hurst, S. Sherman, "Concavity of magnetization of an Ising ferromagnet in a positive external
field", J. Math. Phys. 11 (1970) 790–795. Scope: a finite graph with uniform coupling and uniform
field, the field entering the Hamiltonian as -h ∑ σ_x and the derivative taken in h; χ ≤ 1/h
is the weak form of the GHS consequence χ ≤ m/h, the full concavity statement being
ghs_magnetization_concaveOn. The hypotheses β ≥ 0, h > 0 are satisfiable.
theorem ghs_susceptibility_le_inv_field (β h : ℝ) (hβ : 0 ≤ β) (hh : 0 < h)
(o : V) :
deriv (fun t => isingExpectation G β t (fun s => spin s o)) h =
β * ∑ x, (isingExpectation G β h (fun s => spin s o * spin s x) -
isingExpectation G β h (fun s => spin s o) *
isingExpectation G β h (fun s => spin s x))
∧ deriv (fun t => isingExpectation G β t (fun s => spin s o)) h ≤ 1 / h
noncomputable def isingExpectation (β h : ℝ) (f : ConfigSpace V → ℝ) : ℝ :=
∑ s : ConfigSpace V, isingProb G β h s * f s
PhyslibExt/CondensedMatter/IsingModel/GKS.lean:208
def spin (s : ConfigSpace V) (x : V) : ℝ := if s x then 1 else -1
PhyslibExt/CondensedMatter/IsingModel/Gibbs.lean:70
Loading dependency explorer…
The dependency explorer needs JavaScript and its graph data.