descriptive combinatorics and ergodic theoremsanush tserunyan university of illinois at...

101
Descriptive combinatorics and ergodic theorems Anush Tserunyan University of Illinois at Urbana-Champaign

Upload: others

Post on 24-Jan-2021

2 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Descriptive combinatorics and ergodictheorems

Anush Tserunyan

University of Illinois at Urbana-Champaign

Page 2: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

“Gee Professor Kechris, descriptive set theory sure ispowerful, and beautiful too!”

Page 3: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Descriptive set theory and equivalence relations

I Descriptive set theory studies definable sets/functions in Polish spaces.

I In the last 30 years: definable (Borel, analytic) equivalence relations.

— View an equivalence relation E on a Polish space X as E ⊆ X 2.

I Such equivalence relations arise naturally all over mathematics:

Orbit equivalence relations of Borel actions of Polish groups on Polishspaces are analytic.

Many mathematical objects (such as Riemann surfaces, Banach spaces,measure-preserving transformations, etc.) can be encoded as points inPolish spaces.

Thinking of classifying these points up to some notion of equivalence (e.g.,conformal equivalence, isomorphism, conjugacy), we define what it meansfor one such classification problem to be no harder than another: the Borelreducibility relation 6B between two equivalence relations.

I We will focus on countable Borel equivalence relations (CBERs).

— Here, “countable” means each equivalence class is countable.

Page 4: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Descriptive set theory and equivalence relations

I Descriptive set theory studies definable sets/functions in Polish spaces.

I In the last 30 years: definable (Borel, analytic) equivalence relations.

— View an equivalence relation E on a Polish space X as E ⊆ X 2.

I Such equivalence relations arise naturally all over mathematics:

Orbit equivalence relations of Borel actions of Polish groups on Polishspaces are analytic.

Many mathematical objects (such as Riemann surfaces, Banach spaces,measure-preserving transformations, etc.) can be encoded as points inPolish spaces.

Thinking of classifying these points up to some notion of equivalence (e.g.,conformal equivalence, isomorphism, conjugacy), we define what it meansfor one such classification problem to be no harder than another: the Borelreducibility relation 6B between two equivalence relations.

I We will focus on countable Borel equivalence relations (CBERs).

— Here, “countable” means each equivalence class is countable.

Page 5: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Descriptive set theory and equivalence relations

I Descriptive set theory studies definable sets/functions in Polish spaces.

I In the last 30 years: definable (Borel, analytic) equivalence relations.

— View an equivalence relation E on a Polish space X as E ⊆ X 2.

I Such equivalence relations arise naturally all over mathematics:

Orbit equivalence relations of Borel actions of Polish groups on Polishspaces are analytic.

Many mathematical objects (such as Riemann surfaces, Banach spaces,measure-preserving transformations, etc.) can be encoded as points inPolish spaces.

Thinking of classifying these points up to some notion of equivalence (e.g.,conformal equivalence, isomorphism, conjugacy), we define what it meansfor one such classification problem to be no harder than another: the Borelreducibility relation 6B between two equivalence relations.

I We will focus on countable Borel equivalence relations (CBERs).

— Here, “countable” means each equivalence class is countable.

Page 6: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Descriptive set theory and equivalence relations

I Descriptive set theory studies definable sets/functions in Polish spaces.

I In the last 30 years: definable (Borel, analytic) equivalence relations.

— View an equivalence relation E on a Polish space X as E ⊆ X 2.

I Such equivalence relations arise naturally all over mathematics:

Orbit equivalence relations of Borel actions of Polish groups on Polishspaces are analytic.

Many mathematical objects (such as Riemann surfaces, Banach spaces,measure-preserving transformations, etc.) can be encoded as points inPolish spaces.

Thinking of classifying these points up to some notion of equivalence (e.g.,conformal equivalence, isomorphism, conjugacy), we define what it meansfor one such classification problem to be no harder than another: the Borelreducibility relation 6B between two equivalence relations.

I We will focus on countable Borel equivalence relations (CBERs).

— Here, “countable” means each equivalence class is countable.

Page 7: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Descriptive set theory and equivalence relations

I Descriptive set theory studies definable sets/functions in Polish spaces.

I In the last 30 years: definable (Borel, analytic) equivalence relations.

— View an equivalence relation E on a Polish space X as E ⊆ X 2.

I Such equivalence relations arise naturally all over mathematics:

Orbit equivalence relations of Borel actions of Polish groups on Polishspaces are analytic.

Many mathematical objects (such as Riemann surfaces, Banach spaces,measure-preserving transformations, etc.) can be encoded as points inPolish spaces.

Thinking of classifying these points up to some notion of equivalence (e.g.,conformal equivalence, isomorphism, conjugacy), we define what it meansfor one such classification problem to be no harder than another: the Borelreducibility relation 6B between two equivalence relations.

I We will focus on countable Borel equivalence relations (CBERs).

— Here, “countable” means each equivalence class is countable.

Page 8: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Descriptive set theory and equivalence relations

I Descriptive set theory studies definable sets/functions in Polish spaces.

I In the last 30 years: definable (Borel, analytic) equivalence relations.

— View an equivalence relation E on a Polish space X as E ⊆ X 2.

I Such equivalence relations arise naturally all over mathematics:

Orbit equivalence relations of Borel actions of Polish groups on Polishspaces are analytic.

Many mathematical objects (such as Riemann surfaces, Banach spaces,measure-preserving transformations, etc.) can be encoded as points inPolish spaces.

Thinking of classifying these points up to some notion of equivalence (e.g.,conformal equivalence, isomorphism, conjugacy), we define what it meansfor one such classification problem to be no harder than another: the Borelreducibility relation 6B between two equivalence relations.

I We will focus on countable Borel equivalence relations (CBERs).

— Here, “countable” means each equivalence class is countable.

Page 9: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Descriptive set theory and equivalence relations

I Descriptive set theory studies definable sets/functions in Polish spaces.

I In the last 30 years: definable (Borel, analytic) equivalence relations.

— View an equivalence relation E on a Polish space X as E ⊆ X 2.

I Such equivalence relations arise naturally all over mathematics:

Orbit equivalence relations of Borel actions of Polish groups on Polishspaces are analytic.

Many mathematical objects (such as Riemann surfaces, Banach spaces,measure-preserving transformations, etc.) can be encoded as points inPolish spaces.

Thinking of classifying these points up to some notion of equivalence (e.g.,conformal equivalence, isomorphism, conjugacy), we define what it meansfor one such classification problem to be no harder than another: the Borelreducibility relation 6B between two equivalence relations.

I We will focus on countable Borel equivalence relations (CBERs).

— Here, “countable” means each equivalence class is countable.

Page 10: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Countable Borel equivalence relations via group actions

I These come from countable group actions: orbit equivalence relationsinduced by Borel actions of countable groups are countable Borel.

I Conversely, the Feldman–Moore theorem states that these are all of them!

I Thus, taking {γn} =.. Γ y X with E = EΓ, each x ∈ X can refer to otherpoints in its E -class by names: γ0x , γ1x , γ2x . . .

I The abundance of Borel actions of countable groups makes the class ofCBERs extremely rich and 6B on it very complicated:

Adams–Kechris (using Zimmer’s cocycle superrigidity): N ⊇ A 7→ EA suchthat for all A,B ⊆ N, A ⊆ B ⇔ EA 6B EB .

Page 11: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Countable Borel equivalence relations via group actions

I These come from countable group actions: orbit equivalence relationsinduced by Borel actions of countable groups are countable Borel.

I Conversely, the Feldman–Moore theorem states that these are all of them!

I Thus, taking {γn} =.. Γ y X with E = EΓ, each x ∈ X can refer to otherpoints in its E -class by names: γ0x , γ1x , γ2x . . .

I The abundance of Borel actions of countable groups makes the class ofCBERs extremely rich and 6B on it very complicated:

Adams–Kechris (using Zimmer’s cocycle superrigidity): N ⊇ A 7→ EA suchthat for all A,B ⊆ N, A ⊆ B ⇔ EA 6B EB .

Page 12: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Countable Borel equivalence relations via group actions

I These come from countable group actions: orbit equivalence relationsinduced by Borel actions of countable groups are countable Borel.

I Conversely, the Feldman–Moore theorem states that these are all of them!

I Thus, taking {γn} =.. Γ y X with E = EΓ, each x ∈ X can refer to otherpoints in its E -class by names: γ0x , γ1x , γ2x . . .

I The abundance of Borel actions of countable groups makes the class ofCBERs extremely rich and 6B on it very complicated:

Adams–Kechris (using Zimmer’s cocycle superrigidity): N ⊇ A 7→ EA suchthat for all A,B ⊆ N, A ⊆ B ⇔ EA 6B EB .

Page 13: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Countable Borel equivalence relations via group actions

I These come from countable group actions: orbit equivalence relationsinduced by Borel actions of countable groups are countable Borel.

I Conversely, the Feldman–Moore theorem states that these are all of them!

I Thus, taking {γn} =.. Γ y X with E = EΓ, each x ∈ X can refer to otherpoints in its E -class by names: γ0x , γ1x , γ2x . . .

I The abundance of Borel actions of countable groups makes the class ofCBERs extremely rich and 6B on it very complicated:

Adams–Kechris (using Zimmer’s cocycle superrigidity): N ⊇ A 7→ EA suchthat for all A,B ⊆ N, A ⊆ B ⇔ EA 6B EB .

Page 14: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Countable Borel equivalence relations via group actions

I These come from countable group actions: orbit equivalence relationsinduced by Borel actions of countable groups are countable Borel.

I Conversely, the Feldman–Moore theorem states that these are all of them!

I Thus, taking {γn} =.. Γ y X with E = EΓ, each x ∈ X can refer to otherpoints in its E -class by names: γ0x , γ1x , γ2x . . .

I The abundance of Borel actions of countable groups makes the class ofCBERs extremely rich and 6B on it very complicated:

Adams–Kechris (using Zimmer’s cocycle superrigidity): N ⊇ A 7→ EA suchthat for all A,B ⊆ N, A ⊆ B ⇔ EA 6B EB .

Page 15: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Countable Borel equivalence relations via graphs

I CBERs also come from graphs as the connectedness relations EG oflocally countable Borel graphs G .

— By a graph G on a space X we just mean a symmetric irreflexive G ⊆ X 2.

I Every CBER E on X comes from such a graph:

Take G ..= E \ {(x , x) : x ∈ X}, the complete graph for E .

Even better: let Γ yα X such that E = EΓ, take a symmetric generatingset Γ = 〈S〉, and define the Cayley–Schreier graph:

xGSy ..⇔ σ ·α x = y for some σ ∈ S .

Then E = EGS.

Page 16: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Countable Borel equivalence relations via graphs

I CBERs also come from graphs as the connectedness relations EG oflocally countable Borel graphs G .

— By a graph G on a space X we just mean a symmetric irreflexive G ⊆ X 2.

I Every CBER E on X comes from such a graph:

Take G ..= E \ {(x , x) : x ∈ X}, the complete graph for E .

Even better: let Γ yα X such that E = EΓ, take a symmetric generatingset Γ = 〈S〉, and define the Cayley–Schreier graph:

xGSy ..⇔ σ ·α x = y for some σ ∈ S .

Then E = EGS.

Page 17: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Countable Borel equivalence relations via graphs

I CBERs also come from graphs as the connectedness relations EG oflocally countable Borel graphs G .

— By a graph G on a space X we just mean a symmetric irreflexive G ⊆ X 2.

I Every CBER E on X comes from such a graph:

Take G ..= E \ {(x , x) : x ∈ X}, the complete graph for E .

Even better: let Γ yα X such that E = EΓ, take a symmetric generatingset Γ = 〈S〉, and define the Cayley–Schreier graph:

xGSy ..⇔ σ ·α x = y for some σ ∈ S .

Then E = EGS.

Page 18: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Countable Borel equivalence relations via graphs

I CBERs also come from graphs as the connectedness relations EG oflocally countable Borel graphs G .

— By a graph G on a space X we just mean a symmetric irreflexive G ⊆ X 2.

I Every CBER E on X comes from such a graph:

Take G ..= E \ {(x , x) : x ∈ X}, the complete graph for E .

Even better: let Γ yα X such that E = EΓ, take a symmetric generatingset Γ = 〈S〉, and define the Cayley–Schreier graph:

xGSy ..⇔ σ ·α x = y for some σ ∈ S .

Then E = EGS.

Page 19: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Countable Borel equivalence relations via graphs

I CBERs also come from graphs as the connectedness relations EG oflocally countable Borel graphs G .

— By a graph G on a space X we just mean a symmetric irreflexive G ⊆ X 2.

I Every CBER E on X comes from such a graph:

Take G ..= E \ {(x , x) : x ∈ X}, the complete graph for E .

Even better: let Γ yα X such that E = EΓ, take a symmetric generatingset Γ = 〈S〉, and define the Cayley–Schreier graph:

xGSy ..⇔ σ ·α x = y for some σ ∈ S .

Then E = EGS.

Page 20: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Countable Borel equivalence relations via graphs

I CBERs also come from graphs as the connectedness relations EG oflocally countable Borel graphs G .

— By a graph G on a space X we just mean a symmetric irreflexive G ⊆ X 2.

I Every CBER E on X comes from such a graph:

Take G ..= E \ {(x , x) : x ∈ X}, the complete graph for E .

Even better: let Γ yα X such that E = EΓ, take a symmetric generatingset Γ = 〈S〉, and define the Cayley–Schreier graph:

xGSy ..⇔ σ ·α x = y for some σ ∈ S .

Then E = EGS.

Page 21: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Interplay with other subjects

I Any countable Borel group action Γ y X can be turned into acontinuous one by replacing the Polish topology on X .

— Enables topological tools such as Baire category.

I Borel group actions and graphs naturally occur on measure spaces (X , µ).

— Enables measure-theoretic tools such as the Borel–Cantelli lemma andmuch much more.

I All these together has created extremely active two-way traffic betweenthe study of CBERs and

ergodic theory

(ergodic theorems, mixing, entropy, . . . )

measured group theory

(measure equivalence, cost, `2-(co)homology, . . . )

graph combinatorics

(colorings, matchings, flows, . . . )

geometric group theory

(tree constructions, growth,. . . )

percolation theory

(graph/action constructions, amenability barriers, . . . )

probabilistic combinatorics

(Lovasz Local Lemma, concentration bounds, . . . )

topological dynamics

(subshifts, topological entropy, compact actions, . . . )

von Neumann algebras

(spectral gap, superrigidity, vN-dimension, . . . )

Page 22: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Interplay with other subjects

I Any countable Borel group action Γ y X can be turned into acontinuous one by replacing the Polish topology on X .

— Enables topological tools such as Baire category.

I Borel group actions and graphs naturally occur on measure spaces (X , µ).

— Enables measure-theoretic tools such as the Borel–Cantelli lemma andmuch much more.

I All these together has created extremely active two-way traffic betweenthe study of CBERs and

ergodic theory

(ergodic theorems, mixing, entropy, . . . )

measured group theory

(measure equivalence, cost, `2-(co)homology, . . . )

graph combinatorics

(colorings, matchings, flows, . . . )

geometric group theory

(tree constructions, growth,. . . )

percolation theory

(graph/action constructions, amenability barriers, . . . )

probabilistic combinatorics

(Lovasz Local Lemma, concentration bounds, . . . )

topological dynamics

(subshifts, topological entropy, compact actions, . . . )

von Neumann algebras

(spectral gap, superrigidity, vN-dimension, . . . )

Page 23: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Interplay with other subjects

I Any countable Borel group action Γ y X can be turned into acontinuous one by replacing the Polish topology on X .

— Enables topological tools such as Baire category.

I Borel group actions and graphs naturally occur on measure spaces (X , µ).

— Enables measure-theoretic tools such as the Borel–Cantelli lemma andmuch much more.

I All these together has created extremely active two-way traffic betweenthe study of CBERs and

ergodic theory

(ergodic theorems, mixing, entropy, . . . )

measured group theory

(measure equivalence, cost, `2-(co)homology, . . . )

graph combinatorics

(colorings, matchings, flows, . . . )

geometric group theory

(tree constructions, growth,. . . )

percolation theory

(graph/action constructions, amenability barriers, . . . )

probabilistic combinatorics

(Lovasz Local Lemma, concentration bounds, . . . )

topological dynamics

(subshifts, topological entropy, compact actions, . . . )

von Neumann algebras

(spectral gap, superrigidity, vN-dimension, . . . )

Page 24: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Interplay with other subjects

I Any countable Borel group action Γ y X can be turned into acontinuous one by replacing the Polish topology on X .

— Enables topological tools such as Baire category.

I Borel group actions and graphs naturally occur on measure spaces (X , µ).

— Enables measure-theoretic tools such as the Borel–Cantelli lemma andmuch much more.

I All these together has created extremely active two-way traffic betweenthe study of CBERs and

ergodic theory (ergodic theorems, mixing, entropy, . . . )

measured group theory (measure equivalence, cost, `2-(co)homology, . . . )

graph combinatorics (colorings, matchings, flows, . . . )

geometric group theory (tree constructions, growth,. . . )

percolation theory (graph/action constructions, amenability barriers, . . . )

probabilistic combinatorics (Lovasz Local Lemma, concentration bounds, . . . )

topological dynamics (subshifts, topological entropy, compact actions, . . . )

von Neumann algebras (spectral gap, superrigidity, vN-dimension, . . . )

Page 25: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Our secret weapon

But what’s our contribution? What’s the difference between descriptive settheoretic and analytic thinking?

I We think pointwise, analyzing the local combinatorics at a point, whereasanalysts analyze the space through the prism of functions on it.

I What allows us to do this is the Luzin–Novikov uniformization theorem!

— Every Borel set B ⊆ X × Y with countable X -fibers is a countable unionof Borel partial functions X ⇀ Y .

— Allows each x ∈ X to quantify over its (countable) equivalence class andgive (Borel) names to the other guys in its class.

I We are not allowed to choose a point from each class! (Measure Theory 101)

I Thus, our way of thinking is best described as originless combinatorics.

Page 26: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Our secret weapon

But what’s our contribution? What’s the difference between descriptive settheoretic and analytic thinking?

I We think pointwise, analyzing the local combinatorics at a point, whereasanalysts analyze the space through the prism of functions on it.

I What allows us to do this is the Luzin–Novikov uniformization theorem!

— Every Borel set B ⊆ X × Y with countable X -fibers is a countable unionof Borel partial functions X ⇀ Y .

— Allows each x ∈ X to quantify over its (countable) equivalence class andgive (Borel) names to the other guys in its class.

I We are not allowed to choose a point from each class! (Measure Theory 101)

I Thus, our way of thinking is best described as originless combinatorics.

Page 27: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Our secret weapon

But what’s our contribution? What’s the difference between descriptive settheoretic and analytic thinking?

I We think pointwise, analyzing the local combinatorics at a point, whereasanalysts analyze the space through the prism of functions on it.

I What allows us to do this is the Luzin–Novikov uniformization theorem!

— Every Borel set B ⊆ X × Y with countable X -fibers is a countable unionof Borel partial functions X ⇀ Y .

— Allows each x ∈ X to quantify over its (countable) equivalence class andgive (Borel) names to the other guys in its class.

I We are not allowed to choose a point from each class! (Measure Theory 101)

I Thus, our way of thinking is best described as originless combinatorics.

Page 28: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Our secret weapon

But what’s our contribution? What’s the difference between descriptive settheoretic and analytic thinking?

I We think pointwise, analyzing the local combinatorics at a point, whereasanalysts analyze the space through the prism of functions on it.

I What allows us to do this is the Luzin–Novikov uniformization theorem!

— Every Borel set B ⊆ X × Y with countable X -fibers is a countable unionof Borel partial functions X ⇀ Y .

— Allows each x ∈ X to quantify over its (countable) equivalence class andgive (Borel) names to the other guys in its class.

I We are not allowed to choose a point from each class! (Measure Theory 101)

I Thus, our way of thinking is best described as originless combinatorics.

Page 29: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Our secret weapon

But what’s our contribution? What’s the difference between descriptive settheoretic and analytic thinking?

I We think pointwise, analyzing the local combinatorics at a point, whereasanalysts analyze the space through the prism of functions on it.

I What allows us to do this is the Luzin–Novikov uniformization theorem!

— Every Borel set B ⊆ X × Y with countable X -fibers is a countable unionof Borel partial functions X ⇀ Y .

— Allows each x ∈ X to quantify over its (countable) equivalence class

andgive (Borel) names to the other guys in its class.

I We are not allowed to choose a point from each class! (Measure Theory 101)

I Thus, our way of thinking is best described as originless combinatorics.

Page 30: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Our secret weapon

But what’s our contribution? What’s the difference between descriptive settheoretic and analytic thinking?

I We think pointwise, analyzing the local combinatorics at a point, whereasanalysts analyze the space through the prism of functions on it.

I What allows us to do this is the Luzin–Novikov uniformization theorem!

— Every Borel set B ⊆ X × Y with countable X -fibers is a countable unionof Borel partial functions X ⇀ Y .

— Allows each x ∈ X to quantify over its (countable) equivalence class andgive (Borel) names to the other guys in its class.

I We are not allowed to choose a point from each class! (Measure Theory 101)

I Thus, our way of thinking is best described as originless combinatorics.

Page 31: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Our secret weapon

But what’s our contribution? What’s the difference between descriptive settheoretic and analytic thinking?

I We think pointwise, analyzing the local combinatorics at a point, whereasanalysts analyze the space through the prism of functions on it.

I What allows us to do this is the Luzin–Novikov uniformization theorem!

— Every Borel set B ⊆ X × Y with countable X -fibers is a countable unionof Borel partial functions X ⇀ Y .

— Allows each x ∈ X to quantify over its (countable) equivalence class andgive (Borel) names to the other guys in its class.

I We are not allowed to choose a point from each class! (Measure Theory 101)

I Thus, our way of thinking is best described as originless combinatorics.

Page 32: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Our secret weapon

But what’s our contribution? What’s the difference between descriptive settheoretic and analytic thinking?

I We think pointwise, analyzing the local combinatorics at a point, whereasanalysts analyze the space through the prism of functions on it.

I What allows us to do this is the Luzin–Novikov uniformization theorem!

— Every Borel set B ⊆ X × Y with countable X -fibers is a countable unionof Borel partial functions X ⇀ Y .

— Allows each x ∈ X to quantify over its (countable) equivalence class andgive (Borel) names to the other guys in its class.

I We are not allowed to choose a point from each class! (Measure Theory 101)

I Thus, our way of thinking is best described as originless combinatorics.

Page 33: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

pmp actions of groups and ergodicity

As an example, today we’ll discuss ergodic theorems.

I (X , µ) — a standard probability space, e.g. [0, 1] with Lebesgue measure

I Γ — a countable (discrete) group

I An action Γ y (X , µ) is ergodic if every invariant measurable subset isnull or conull.

I A probability measure preserving (pmp) action of Γ is a Borel actionΓ y (X , µ) such that µ(γ · A) = µ(A), for each γ ∈ Γ and A ⊆ X .

I Dating back to Birkhoff, pointwise ergodic theorems for pmp actionsΓ y (X , µ) are bridges between the global condition of ergodicity and thea.e. local combinatorics of the action.

Page 34: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

pmp actions of groups and ergodicity

As an example, today we’ll discuss ergodic theorems.

I (X , µ) — a standard probability space, e.g. [0, 1] with Lebesgue measure

I Γ — a countable (discrete) group

I An action Γ y (X , µ) is ergodic if every invariant measurable subset isnull or conull.

I A probability measure preserving (pmp) action of Γ is a Borel actionΓ y (X , µ) such that µ(γ · A) = µ(A), for each γ ∈ Γ and A ⊆ X .

I Dating back to Birkhoff, pointwise ergodic theorems for pmp actionsΓ y (X , µ) are bridges between the global condition of ergodicity and thea.e. local combinatorics of the action.

Page 35: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

pmp actions of groups and ergodicity

As an example, today we’ll discuss ergodic theorems.

I (X , µ) — a standard probability space, e.g. [0, 1] with Lebesgue measure

I Γ — a countable (discrete) group

I An action Γ y (X , µ) is ergodic if every invariant measurable subset isnull or conull.

I A probability measure preserving (pmp) action of Γ is a Borel actionΓ y (X , µ) such that µ(γ · A) = µ(A), for each γ ∈ Γ and A ⊆ X .

I Dating back to Birkhoff, pointwise ergodic theorems for pmp actionsΓ y (X , µ) are bridges between the global condition of ergodicity and thea.e. local combinatorics of the action.

Page 36: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

pmp actions of groups and ergodicity

As an example, today we’ll discuss ergodic theorems.

I (X , µ) — a standard probability space, e.g. [0, 1] with Lebesgue measure

I Γ — a countable (discrete) group

I An action Γ y (X , µ) is ergodic if every invariant measurable subset isnull or conull.

I A probability measure preserving (pmp) action of Γ is a Borel actionΓ y (X , µ) such that µ(γ · A) = µ(A), for each γ ∈ Γ and A ⊆ X .

I Dating back to Birkhoff, pointwise ergodic theorems for pmp actionsΓ y (X , µ) are bridges between the global condition of ergodicity and thea.e. local combinatorics of the action.

Page 37: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

pmp actions of groups and ergodicity

As an example, today we’ll discuss ergodic theorems.

I (X , µ) — a standard probability space, e.g. [0, 1] with Lebesgue measure

I Γ — a countable (discrete) group

I An action Γ y (X , µ) is ergodic if every invariant measurable subset isnull or conull.

I A probability measure preserving (pmp) action of Γ is a Borel actionΓ y (X , µ) such that µ(γ · A) = µ(A), for each γ ∈ Γ and A ⊆ X .

I Dating back to Birkhoff, pointwise ergodic theorems for pmp actionsΓ y (X , µ) are bridges between the global condition of ergodicity and thea.e. local combinatorics of the action.

Page 38: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

pmp actions of groups and ergodicity

As an example, today we’ll discuss ergodic theorems.

I (X , µ) — a standard probability space, e.g. [0, 1] with Lebesgue measure

I Γ — a countable (discrete) group

I An action Γ y (X , µ) is ergodic if every invariant measurable subset isnull or conull.

I A probability measure preserving (pmp) action of Γ is a Borel actionΓ y (X , µ) such that µ(γ · A) = µ(A), for each γ ∈ Γ and A ⊆ X .

I Dating back to Birkhoff, pointwise ergodic theorems for pmp actionsΓ y (X , µ) are bridges between the global condition of ergodicity and thea.e. local combinatorics of the action.

Page 39: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Pointwise ergodic theorems for pmp actions

Theorem (Pointwise ergodic, Birkhoff 1931)

A pmp action Z yα (X , µ) is ergodic if and only if for each f ∈ L1(X , µ)and for a.e. x ∈ X ,

limit of local averages limn→∞

Af [Fn ·α x ] =

∫Xf dµ global average

where Fn ..= [0, n] ⊆ Z and Af [Fn ·α x ] is the average of f over Fn ·α x .

I This was generalized to amenable groups by Lindenstrauss (2001).I Analogous statements for free groups have been proven by Grigorchuk

(1987), Nevo (1994), Nevo–Stein (1994), Bufetov (2002), andBowen–Nevo (2013).

I Bowen–Nevo (2013) also have pointwise ergodic theorems for othernonamenable groups, but for special kinds of actions.

Page 40: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Pointwise ergodic theorems for pmp actions

Theorem (Pointwise ergodic, Birkhoff 1931)

A pmp action Z yα (X , µ) is ergodic if and only if for each f ∈ L1(X , µ)and for a.e. x ∈ X ,

limit of local averages limn→∞

Af [Fn ·α x ] =

∫Xf dµ global average

where Fn ..= [0, n] ⊆ Z and Af [Fn ·α x ] is the average of f over Fn ·α x .

I This was generalized to amenable groups by Lindenstrauss (2001).

I Analogous statements for free groups have been proven by Grigorchuk(1987), Nevo (1994), Nevo–Stein (1994), Bufetov (2002), andBowen–Nevo (2013).

I Bowen–Nevo (2013) also have pointwise ergodic theorems for othernonamenable groups, but for special kinds of actions.

Page 41: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Pointwise ergodic theorems for pmp actions

Theorem (Pointwise ergodic, Birkhoff 1931)

A pmp action Z yα (X , µ) is ergodic if and only if for each f ∈ L1(X , µ)and for a.e. x ∈ X ,

limit of local averages limn→∞

Af [Fn ·α x ] =

∫Xf dµ global average

where Fn ..= [0, n] ⊆ Z and Af [Fn ·α x ] is the average of f over Fn ·α x .

I This was generalized to amenable groups by Lindenstrauss (2001).I Analogous statements for free groups have been proven by Grigorchuk

(1987), Nevo (1994), Nevo–Stein (1994), Bufetov (2002), andBowen–Nevo (2013).

I Bowen–Nevo (2013) also have pointwise ergodic theorems for othernonamenable groups, but for special kinds of actions.

Page 42: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Pointwise ergodic theorems for pmp actions

Theorem (Pointwise ergodic, Birkhoff 1931)

A pmp action Z yα (X , µ) is ergodic if and only if for each f ∈ L1(X , µ)and for a.e. x ∈ X ,

limit of local averages limn→∞

Af [Fn ·α x ] =

∫Xf dµ global average

where Fn ..= [0, n] ⊆ Z and Af [Fn ·α x ] is the average of f over Fn ·α x .

I This was generalized to amenable groups by Lindenstrauss (2001).I Analogous statements for free groups have been proven by Grigorchuk

(1987), Nevo (1994), Nevo–Stein (1994), Bufetov (2002), andBowen–Nevo (2013).

I Bowen–Nevo (2013) also have pointwise ergodic theorems for othernonamenable groups, but for special kinds of actions.

Page 43: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Pointwise ergodic theorems for quasi-pmp actions

I An action Γ y (X , µ) is quasi-pmp if every γ ∈ Γ is nonsingular, i.e.maps nonnull sets to nonnull sets (possibly changing their measure).

I Why would one consider quasi-pmp actions?

Theorem (Ratio Pointwise Ergodic, Hopf 1937)

Let Z yα (X , µ) be an ergodic pmp action. For each f , g ∈ L1(X , µ) withg > 0 and for a.e. x ∈ X ,

limn→∞

Af [Fn ·α x ]

Ag [Fn ·α x ]=

∫X f dµ∫X gdµ

.

I Replacing the measure µ with dµg..= gdµ makes the action Γ y (X , µg )

quasi-pmp and turns the ratio ergodic theorem into a usual one.I A quasi-pmp ergodic theorem for Zd was proven by Feldman (2007) and

generalized by Hochman (2010) and Dooley–Jarrett (2016).I Also known for lattice actions on homogeneous spaces (Nevo and co.), a

weaker form for groups of polynomial growth (Hochman 2013), anindirect version for free groups (Bowen–Nevo 2014).

Page 44: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Pointwise ergodic theorems for quasi-pmp actions

I An action Γ y (X , µ) is quasi-pmp if every γ ∈ Γ is nonsingular, i.e.maps nonnull sets to nonnull sets (possibly changing their measure).

I Why would one consider quasi-pmp actions?

Theorem (Ratio Pointwise Ergodic, Hopf 1937)

Let Z yα (X , µ) be an ergodic pmp action. For each f , g ∈ L1(X , µ) withg > 0 and for a.e. x ∈ X ,

limn→∞

Af [Fn ·α x ]

Ag [Fn ·α x ]=

∫X f dµ∫X gdµ

.

I Replacing the measure µ with dµg..= gdµ makes the action Γ y (X , µg )

quasi-pmp and turns the ratio ergodic theorem into a usual one.I A quasi-pmp ergodic theorem for Zd was proven by Feldman (2007) and

generalized by Hochman (2010) and Dooley–Jarrett (2016).I Also known for lattice actions on homogeneous spaces (Nevo and co.), a

weaker form for groups of polynomial growth (Hochman 2013), anindirect version for free groups (Bowen–Nevo 2014).

Page 45: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Pointwise ergodic theorems for quasi-pmp actions

I An action Γ y (X , µ) is quasi-pmp if every γ ∈ Γ is nonsingular, i.e.maps nonnull sets to nonnull sets (possibly changing their measure).

I Why would one consider quasi-pmp actions?

Theorem (Ratio Pointwise Ergodic, Hopf 1937)

Let Z yα (X , µ) be an ergodic pmp action. For each f , g ∈ L1(X , µ) withg > 0 and for a.e. x ∈ X ,

limn→∞

Af [Fn ·α x ]

Ag [Fn ·α x ]=

∫X f dµ∫X gdµ

.

I Replacing the measure µ with dµg..= gdµ makes the action Γ y (X , µg )

quasi-pmp and turns the ratio ergodic theorem into a usual one.I A quasi-pmp ergodic theorem for Zd was proven by Feldman (2007) and

generalized by Hochman (2010) and Dooley–Jarrett (2016).I Also known for lattice actions on homogeneous spaces (Nevo and co.), a

weaker form for groups of polynomial growth (Hochman 2013), anindirect version for free groups (Bowen–Nevo 2014).

Page 46: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Pointwise ergodic theorems for quasi-pmp actions

I An action Γ y (X , µ) is quasi-pmp if every γ ∈ Γ is nonsingular, i.e.maps nonnull sets to nonnull sets (possibly changing their measure).

I Why would one consider quasi-pmp actions?

Theorem (Ratio Pointwise Ergodic, Hopf 1937)

Let Z yα (X , µ) be an ergodic pmp action. For each f , g ∈ L1(X , µ) withg > 0 and for a.e. x ∈ X ,

limn→∞

Af [Fn ·α x ]

Ag [Fn ·α x ]=

∫X f dµ∫X gdµ

.

I Replacing the measure µ with dµg..= gdµ makes the action Γ y (X , µg )

quasi-pmp and turns the ratio ergodic theorem into a usual one.

I A quasi-pmp ergodic theorem for Zd was proven by Feldman (2007) andgeneralized by Hochman (2010) and Dooley–Jarrett (2016).

I Also known for lattice actions on homogeneous spaces (Nevo and co.), aweaker form for groups of polynomial growth (Hochman 2013), anindirect version for free groups (Bowen–Nevo 2014).

Page 47: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Pointwise ergodic theorems for quasi-pmp actions

I An action Γ y (X , µ) is quasi-pmp if every γ ∈ Γ is nonsingular, i.e.maps nonnull sets to nonnull sets (possibly changing their measure).

I Why would one consider quasi-pmp actions?

Theorem (Ratio Pointwise Ergodic, Hopf 1937)

Let Z yα (X , µ) be an ergodic pmp action. For each f , g ∈ L1(X , µ) withg > 0 and for a.e. x ∈ X ,

limn→∞

Af [Fn ·α x ]

Ag [Fn ·α x ]=

∫X f dµ∫X gdµ

.

I Replacing the measure µ with dµg..= gdµ makes the action Γ y (X , µg )

quasi-pmp and turns the ratio ergodic theorem into a usual one.I A quasi-pmp ergodic theorem for Zd was proven by Feldman (2007) and

generalized by Hochman (2010) and Dooley–Jarrett (2016).

I Also known for lattice actions on homogeneous spaces (Nevo and co.), aweaker form for groups of polynomial growth (Hochman 2013), anindirect version for free groups (Bowen–Nevo 2014).

Page 48: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Pointwise ergodic theorems for quasi-pmp actions

I An action Γ y (X , µ) is quasi-pmp if every γ ∈ Γ is nonsingular, i.e.maps nonnull sets to nonnull sets (possibly changing their measure).

I Why would one consider quasi-pmp actions?

Theorem (Ratio Pointwise Ergodic, Hopf 1937)

Let Z yα (X , µ) be an ergodic pmp action. For each f , g ∈ L1(X , µ) withg > 0 and for a.e. x ∈ X ,

limn→∞

Af [Fn ·α x ]

Ag [Fn ·α x ]=

∫X f dµ∫X gdµ

.

I Replacing the measure µ with dµg..= gdµ makes the action Γ y (X , µg )

quasi-pmp and turns the ratio ergodic theorem into a usual one.I A quasi-pmp ergodic theorem for Zd was proven by Feldman (2007) and

generalized by Hochman (2010) and Dooley–Jarrett (2016).I Also known for lattice actions on homogeneous spaces (Nevo and co.), a

weaker form for groups of polynomial growth (Hochman 2013), anindirect version for free groups (Bowen–Nevo 2014).

Page 49: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Failures of pointwise ergodic for some quasi-pmp actions

I However, Hochman also showed (2012) that the ratio ergodic theorem isfalse for Γ ..=

⊕n∈N Z along any sequence (Fn) of window frames Fn ⊆ Γ.

I Moreover, it is also false for the nonabelian free groups along anysubsequence of balls.

I Thus, the pointwise ergodic theorem is false for the quasi-pmp actions of⊕n∈N Z and the nonablian free groups.

I Still, is there any pointwise ergodic statement for these actions?

I Let’s abandon the action and look at the induced Cayley–Schreier graph.

Page 50: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Failures of pointwise ergodic for some quasi-pmp actions

I However, Hochman also showed (2012) that the ratio ergodic theorem isfalse for Γ ..=

⊕n∈N Z along any sequence (Fn) of window frames Fn ⊆ Γ.

I Moreover, it is also false for the nonabelian free groups along anysubsequence of balls.

I Thus, the pointwise ergodic theorem is false for the quasi-pmp actions of⊕n∈N Z and the nonablian free groups.

I Still, is there any pointwise ergodic statement for these actions?

I Let’s abandon the action and look at the induced Cayley–Schreier graph.

Page 51: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Failures of pointwise ergodic for some quasi-pmp actions

I However, Hochman also showed (2012) that the ratio ergodic theorem isfalse for Γ ..=

⊕n∈N Z along any sequence (Fn) of window frames Fn ⊆ Γ.

I Moreover, it is also false for the nonabelian free groups along anysubsequence of balls.

I Thus, the pointwise ergodic theorem is false for the quasi-pmp actions of⊕n∈N Z and the nonablian free groups.

I Still, is there any pointwise ergodic statement for these actions?

I Let’s abandon the action and look at the induced Cayley–Schreier graph.

Page 52: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Failures of pointwise ergodic for some quasi-pmp actions

I However, Hochman also showed (2012) that the ratio ergodic theorem isfalse for Γ ..=

⊕n∈N Z along any sequence (Fn) of window frames Fn ⊆ Γ.

I Moreover, it is also false for the nonabelian free groups along anysubsequence of balls.

I Thus, the pointwise ergodic theorem is false for the quasi-pmp actions of⊕n∈N Z and the nonablian free groups.

I Still, is there any pointwise ergodic statement for these actions?

I Let’s abandon the action and look at the induced Cayley–Schreier graph.

Page 53: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Failures of pointwise ergodic for some quasi-pmp actions

I However, Hochman also showed (2012) that the ratio ergodic theorem isfalse for Γ ..=

⊕n∈N Z along any sequence (Fn) of window frames Fn ⊆ Γ.

I Moreover, it is also false for the nonabelian free groups along anysubsequence of balls.

I Thus, the pointwise ergodic theorem is false for the quasi-pmp actions of⊕n∈N Z and the nonablian free groups.

I Still, is there any pointwise ergodic statement for these actions?

I Let’s abandon the action and look at the induced Cayley–Schreier graph.

Page 54: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

The Cayley–Schreier graph and uniform test windows

I A Borel action of a countable group Γ = 〈S〉 on (X , µ) induces a locallycountable Borel graph GS on X — its Cayley–Schreier graph:

xGSy ..⇔ σ · x = y for some σ ∈ S .

I Pointwise ergodic theorems extract a sequence (Fn) of subsets of Γ(window frames) and for a.e. x ∈ X , assert that the local averagesAf [Fn · x ] of an f ∈ L1(X , µ) over the test windows Fn · x converge tothe global average

∫X f dµ.

I Let’s consider graphs in general.

Page 55: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

The Cayley–Schreier graph and uniform test windows

I A Borel action of a countable group Γ = 〈S〉 on (X , µ) induces a locallycountable Borel graph GS on X — its Cayley–Schreier graph:

xGSy ..⇔ σ · x = y for some σ ∈ S .

I Pointwise ergodic theorems extract a sequence (Fn) of subsets of Γ(window frames)

and for a.e. x ∈ X , assert that the local averagesAf [Fn · x ] of an f ∈ L1(X , µ) over the test windows Fn · x converge tothe global average

∫X f dµ.

I Let’s consider graphs in general.

Page 56: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

The Cayley–Schreier graph and uniform test windows

I A Borel action of a countable group Γ = 〈S〉 on (X , µ) induces a locallycountable Borel graph GS on X — its Cayley–Schreier graph:

xGSy ..⇔ σ · x = y for some σ ∈ S .

I Pointwise ergodic theorems extract a sequence (Fn) of subsets of Γ(window frames) and for a.e. x ∈ X , assert that the local averagesAf [Fn · x ] of an f ∈ L1(X , µ) over the test windows Fn · x converge tothe global average

∫X f dµ.

I Let’s consider graphs in general.

Page 57: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

The Cayley–Schreier graph and uniform test windows

I A Borel action of a countable group Γ = 〈S〉 on (X , µ) induces a locallycountable Borel graph GS on X — its Cayley–Schreier graph:

xGSy ..⇔ σ · x = y for some σ ∈ S .

I Pointwise ergodic theorems extract a sequence (Fn) of subsets of Γ(window frames) and for a.e. x ∈ X , assert that the local averagesAf [Fn · x ] of an f ∈ L1(X , µ) over the test windows Fn · x converge tothe global average

∫X f dµ.

I Let’s consider graphs in general.

Page 58: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

More generally: pmp graphs

I G — a locally countable Borel graph on (X , µ).

I EG — the connectedness equivalence relation of G .

I G is ergodic if each EG -invariant measurable subset is null or conull.

I G is pmp if every Borel automorphism γ of X that permutes everyG -connected component is measure preserving.

I In other words, the points in the same G -connected component haveequal mass.

Page 59: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

More generally: pmp graphs

I G — a locally countable Borel graph on (X , µ).

I EG — the connectedness equivalence relation of G .

I G is ergodic if each EG -invariant measurable subset is null or conull.

I G is pmp if every Borel automorphism γ of X that permutes everyG -connected component is measure preserving.

I In other words, the points in the same G -connected component haveequal mass.

Page 60: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

More generally: pmp graphs

I G — a locally countable Borel graph on (X , µ).

I EG — the connectedness equivalence relation of G .

I G is ergodic if each EG -invariant measurable subset is null or conull.

I G is pmp if every Borel automorphism γ of X that permutes everyG -connected component is measure preserving.

I In other words, the points in the same G -connected component haveequal mass.

Page 61: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

More generally: pmp graphs

I G — a locally countable Borel graph on (X , µ).

I EG — the connectedness equivalence relation of G .

I G is ergodic if each EG -invariant measurable subset is null or conull.

I G is pmp if every Borel automorphism γ of X that permutes everyG -connected component is measure preserving.

I In other words, the points in the same G -connected component haveequal mass.

Page 62: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

More generally: pmp graphs

I G — a locally countable Borel graph on (X , µ).

I EG — the connectedness equivalence relation of G .

I G is ergodic if each EG -invariant measurable subset is null or conull.

I G is pmp if every Borel automorphism γ of X that permutes everyG -connected component is measure preserving.

I In other words, the points in the same G -connected component haveequal mass.

Page 63: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Even more generally: quasi-pmp graphs

I G — a locally countable Borel graph on (X , µ).

I G is quasi-pmp if every Borel automorphism γ of X that permutes everyG -connected component is nonsingular (i.e. null preserving).

I Think: the points in the same G -connected component have possiblydifferent nonzero masses.

I This difference is given by a Borel cocycle ρ : EG → R+, which satisfies

ρ(x , y)ρ(y , z) = ρ(x , z).

I Thinking of ρ(x , y) as mass(x)/mass(y), write ρy (x) instead.I It makes µ ρ-invariant, i.e. for any γ as above,

µ(γB) =

∫Bρx(γx) dµ(x).

I The ρ-weighted average of a function f over a finite G -connected U ⊆ X :

Aρf [U] ..=

∑u∈U f (u)ρx(u)∑

u∈U ρx(u),

where x is any/some point in the G -connected component of U.

Page 64: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Even more generally: quasi-pmp graphs

I G — a locally countable Borel graph on (X , µ).I G is quasi-pmp if every Borel automorphism γ of X that permutes every

G -connected component is nonsingular (i.e. null preserving).

I Think: the points in the same G -connected component have possiblydifferent nonzero masses.

I This difference is given by a Borel cocycle ρ : EG → R+, which satisfies

ρ(x , y)ρ(y , z) = ρ(x , z).

I Thinking of ρ(x , y) as mass(x)/mass(y), write ρy (x) instead.I It makes µ ρ-invariant, i.e. for any γ as above,

µ(γB) =

∫Bρx(γx) dµ(x).

I The ρ-weighted average of a function f over a finite G -connected U ⊆ X :

Aρf [U] ..=

∑u∈U f (u)ρx(u)∑

u∈U ρx(u),

where x is any/some point in the G -connected component of U.

Page 65: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Even more generally: quasi-pmp graphs

I G — a locally countable Borel graph on (X , µ).I G is quasi-pmp if every Borel automorphism γ of X that permutes every

G -connected component is nonsingular (i.e. null preserving).I Think: the points in the same G -connected component have possibly

different nonzero masses.

I This difference is given by a Borel cocycle ρ : EG → R+, which satisfies

ρ(x , y)ρ(y , z) = ρ(x , z).

I Thinking of ρ(x , y) as mass(x)/mass(y), write ρy (x) instead.I It makes µ ρ-invariant, i.e. for any γ as above,

µ(γB) =

∫Bρx(γx) dµ(x).

I The ρ-weighted average of a function f over a finite G -connected U ⊆ X :

Aρf [U] ..=

∑u∈U f (u)ρx(u)∑

u∈U ρx(u),

where x is any/some point in the G -connected component of U.

Page 66: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Even more generally: quasi-pmp graphs

I G — a locally countable Borel graph on (X , µ).I G is quasi-pmp if every Borel automorphism γ of X that permutes every

G -connected component is nonsingular (i.e. null preserving).I Think: the points in the same G -connected component have possibly

different nonzero masses.I This difference is given by a Borel cocycle ρ : EG → R+, which satisfies

ρ(x , y)ρ(y , z) = ρ(x , z).

I Thinking of ρ(x , y) as mass(x)/mass(y), write ρy (x) instead.I It makes µ ρ-invariant, i.e. for any γ as above,

µ(γB) =

∫Bρx(γx) dµ(x).

I The ρ-weighted average of a function f over a finite G -connected U ⊆ X :

Aρf [U] ..=

∑u∈U f (u)ρx(u)∑

u∈U ρx(u),

where x is any/some point in the G -connected component of U.

Page 67: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Even more generally: quasi-pmp graphs

I G — a locally countable Borel graph on (X , µ).I G is quasi-pmp if every Borel automorphism γ of X that permutes every

G -connected component is nonsingular (i.e. null preserving).I Think: the points in the same G -connected component have possibly

different nonzero masses.I This difference is given by a Borel cocycle ρ : EG → R+, which satisfies

ρ(x , y)ρ(y , z) = ρ(x , z).

I Thinking of ρ(x , y) as mass(x)/mass(y), write ρy (x) instead.

I It makes µ ρ-invariant, i.e. for any γ as above,

µ(γB) =

∫Bρx(γx) dµ(x).

I The ρ-weighted average of a function f over a finite G -connected U ⊆ X :

Aρf [U] ..=

∑u∈U f (u)ρx(u)∑

u∈U ρx(u),

where x is any/some point in the G -connected component of U.

Page 68: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Even more generally: quasi-pmp graphs

I G — a locally countable Borel graph on (X , µ).I G is quasi-pmp if every Borel automorphism γ of X that permutes every

G -connected component is nonsingular (i.e. null preserving).I Think: the points in the same G -connected component have possibly

different nonzero masses.I This difference is given by a Borel cocycle ρ : EG → R+, which satisfies

ρ(x , y)ρ(y , z) = ρ(x , z).

I Thinking of ρ(x , y) as mass(x)/mass(y), write ρy (x) instead.I It makes µ ρ-invariant,

i.e. for any γ as above,

µ(γB) =

∫Bρx(γx) dµ(x).

I The ρ-weighted average of a function f over a finite G -connected U ⊆ X :

Aρf [U] ..=

∑u∈U f (u)ρx(u)∑

u∈U ρx(u),

where x is any/some point in the G -connected component of U.

Page 69: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Even more generally: quasi-pmp graphs

I G — a locally countable Borel graph on (X , µ).I G is quasi-pmp if every Borel automorphism γ of X that permutes every

G -connected component is nonsingular (i.e. null preserving).I Think: the points in the same G -connected component have possibly

different nonzero masses.I This difference is given by a Borel cocycle ρ : EG → R+, which satisfies

ρ(x , y)ρ(y , z) = ρ(x , z).

I Thinking of ρ(x , y) as mass(x)/mass(y), write ρy (x) instead.I It makes µ ρ-invariant, i.e. for any γ as above,

µ(γB) =

∫Bρx(γx) dµ(x).

I The ρ-weighted average of a function f over a finite G -connected U ⊆ X :

Aρf [U] ..=

∑u∈U f (u)ρx(u)∑

u∈U ρx(u),

where x is any/some point in the G -connected component of U.

Page 70: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Even more generally: quasi-pmp graphs

I G — a locally countable Borel graph on (X , µ).I G is quasi-pmp if every Borel automorphism γ of X that permutes every

G -connected component is nonsingular (i.e. null preserving).I Think: the points in the same G -connected component have possibly

different nonzero masses.I This difference is given by a Borel cocycle ρ : EG → R+, which satisfies

ρ(x , y)ρ(y , z) = ρ(x , z).

I Thinking of ρ(x , y) as mass(x)/mass(y), write ρy (x) instead.I It makes µ ρ-invariant, i.e. for any γ as above,

µ(γB) =

∫Bρx(γx) dµ(x).

I The ρ-weighted average of a function f over a finite G -connected U ⊆ X :

Aρf [U] ..=

∑u∈U f (u)ρx(u)∑

u∈U ρx(u),

where x is any/some point in the G -connected component of U.

Page 71: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

A pointwise ergodic theorem for quasi-pmp graphs

I Let G be a locally countable quasi-pmp ergodic graph on (X , µ) and letρ : EG → R+ be the corresponding cocycle.

I Even if G came from a group action (it always does), Hochman’s resultsshow that uniform/deterministic window frames taken from the groupmay not yield correct averages.

I Nevertheless, we can obtain a Borel pairwise disjoint collection of suchwindows in X (i.e., a finite Borel equivalence relation on X ) ensuring thateach window is G -connected — the main challenge. In other words,

Theorem (Ծ. 2018)

There is an increasing sequence (Fn) of G -connected finite Borelequivalence relations on X such that for every f ∈ L1(X , µ),

limn→∞

Aρf [x ]Fn =

∫Xf dµ, for a.e. x ∈ X .

Here, Aρf [x ]Fn is the ρ-weighted average of f over the Fn-class [x ]Fn of x .

Page 72: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

A pointwise ergodic theorem for quasi-pmp graphs

I Let G be a locally countable quasi-pmp ergodic graph on (X , µ) and letρ : EG → R+ be the corresponding cocycle.

I Even if G came from a group action (it always does), Hochman’s resultsshow that uniform/deterministic window frames taken from the groupmay not yield correct averages.

I Nevertheless, we can obtain a Borel pairwise disjoint collection of suchwindows in X (i.e., a finite Borel equivalence relation on X ) ensuring thateach window is G -connected — the main challenge. In other words,

Theorem (Ծ. 2018)

There is an increasing sequence (Fn) of G -connected finite Borelequivalence relations on X such that for every f ∈ L1(X , µ),

limn→∞

Aρf [x ]Fn =

∫Xf dµ, for a.e. x ∈ X .

Here, Aρf [x ]Fn is the ρ-weighted average of f over the Fn-class [x ]Fn of x .

Page 73: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

A pointwise ergodic theorem for quasi-pmp graphs

I Let G be a locally countable quasi-pmp ergodic graph on (X , µ) and letρ : EG → R+ be the corresponding cocycle.

I Even if G came from a group action (it always does), Hochman’s resultsshow that uniform/deterministic window frames taken from the groupmay not yield correct averages.

I Nevertheless, we can obtain a Borel pairwise disjoint collection of suchwindows in X (i.e., a finite Borel equivalence relation on X ) ensuring thateach window is G -connected — the main challenge.

In other words,

Theorem (Ծ. 2018)

There is an increasing sequence (Fn) of G -connected finite Borelequivalence relations on X such that for every f ∈ L1(X , µ),

limn→∞

Aρf [x ]Fn =

∫Xf dµ, for a.e. x ∈ X .

Here, Aρf [x ]Fn is the ρ-weighted average of f over the Fn-class [x ]Fn of x .

Page 74: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

A pointwise ergodic theorem for quasi-pmp graphs

I Let G be a locally countable quasi-pmp ergodic graph on (X , µ) and letρ : EG → R+ be the corresponding cocycle.

I Even if G came from a group action (it always does), Hochman’s resultsshow that uniform/deterministic window frames taken from the groupmay not yield correct averages.

I Nevertheless, we can obtain a Borel pairwise disjoint collection of suchwindows in X (i.e., a finite Borel equivalence relation on X ) ensuring thateach window is G -connected — the main challenge. In other words,

Theorem (Ծ. 2018)

There is an increasing sequence (Fn) of G -connected finite Borelequivalence relations on X such that for every f ∈ L1(X , µ),

limn→∞

Aρf [x ]Fn =

∫Xf dµ, for a.e. x ∈ X .

Here, Aρf [x ]Fn is the ρ-weighted average of f over the Fn-class [x ]Fn of x .

Page 75: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Deterministic vs. random test windows

Page 76: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Credits and applications

The main theorem in the pmp case was first proven by Tucker-Drob:

Theorem (Tucker-Drob 2016)

The pointwise ergodic theorem is true for locally countable ergodic pmpgraphs.

Applications

Answer to Bowen’s question: Every pmp ergodic countable treeable Borelequivalence relation admits an ergodic hyperfinite free factor.

Ergodic 1% lemma (Tucker-Drob and Conley–Gaboriau–Marks–Tucker-Drob):Every locally countable ergodic nonhyperfinite pmp Borel graph G admitsBorel sets A ⊆ X of arbitrarily small measure such that G |A is stillergodic and nonhyperfinite.

Could have been used in Ergodic Hjorth’s lemma for cost attained (Miller–Ծ.2017): If a countable pmp ergodic Borel equivalence relation E is treeableand has cost n ∈ N∪ {∞}, then it is induced by an a.e. free action of Fn

such that each of the n standard generators of Fn alone acts ergodically.

Page 77: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Credits and applications

The main theorem in the pmp case was first proven by Tucker-Drob:

Theorem (Tucker-Drob 2016)

The pointwise ergodic theorem is true for locally countable ergodic pmpgraphs.

Applications

Answer to Bowen’s question: Every pmp ergodic countable treeable Borelequivalence relation admits an ergodic hyperfinite free factor.

Ergodic 1% lemma (Tucker-Drob and Conley–Gaboriau–Marks–Tucker-Drob):Every locally countable ergodic nonhyperfinite pmp Borel graph G admitsBorel sets A ⊆ X of arbitrarily small measure such that G |A is stillergodic and nonhyperfinite.

Could have been used in Ergodic Hjorth’s lemma for cost attained (Miller–Ծ.2017): If a countable pmp ergodic Borel equivalence relation E is treeableand has cost n ∈ N∪ {∞}, then it is induced by an a.e. free action of Fn

such that each of the n standard generators of Fn alone acts ergodically.

Page 78: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Credits and applications

The main theorem in the pmp case was first proven by Tucker-Drob:

Theorem (Tucker-Drob 2016)

The pointwise ergodic theorem is true for locally countable ergodic pmpgraphs.

Applications

Answer to Bowen’s question: Every ���XXXpmp ergodic countable treeable Borelequivalence relation admits an ergodic hyperfinite free factor.

Ergodic 1% lemma (Tucker-Drob and Conley–Gaboriau–Marks–Tucker-Drob):Every locally countable ergodic nonhyperfinite pmp Borel graph G admitsBorel sets A ⊆ X of arbitrarily small measure such that G |A is stillergodic and nonhyperfinite.

Could have been used in Ergodic Hjorth’s lemma for cost attained (Miller–Ծ.2017): If a countable pmp ergodic Borel equivalence relation E is treeableand has cost n ∈ N∪ {∞}, then it is induced by an a.e. free action of Fn

such that each of the n standard generators of Fn alone acts ergodically.

Page 79: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Two different proofs

I Tucker-Drob’s proof of his theorem is based on a rather deep result inpercolation theory by Hutchcroft and Nachmias (2015) onindistinguishability of the Wired Uniform Spanning Forest.

I Furthermore, to derive his theorem from this, Tucker-Drob makes use ofother probabilistic techniques:

(Gaboriau–Lyons) Wilson’s algorithm rooted at infinity

(essentially Hatami–Lovasz–Szegedy) an analogue for graphs of theAbert–Weiss theorem.

I All these techniques are inherently pmp.

I A year later, I found a combinatorial/purely descriptive set theoretic proofof Tucker-Drob’s theorem (for pmp graphs).

I Another half a year later, the argument became applicable to quasi-pmpgraphs, yielding a generalization.

Page 80: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Two different proofs

I Tucker-Drob’s proof of his theorem is based on a rather deep result inpercolation theory by Hutchcroft and Nachmias (2015) onindistinguishability of the Wired Uniform Spanning Forest.

I Furthermore, to derive his theorem from this, Tucker-Drob makes use ofother probabilistic techniques:

(Gaboriau–Lyons) Wilson’s algorithm rooted at infinity

(essentially Hatami–Lovasz–Szegedy) an analogue for graphs of theAbert–Weiss theorem.

I All these techniques are inherently pmp.

I A year later, I found a combinatorial/purely descriptive set theoretic proofof Tucker-Drob’s theorem (for pmp graphs).

I Another half a year later, the argument became applicable to quasi-pmpgraphs, yielding a generalization.

Page 81: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Two different proofs

I Tucker-Drob’s proof of his theorem is based on a rather deep result inpercolation theory by Hutchcroft and Nachmias (2015) onindistinguishability of the Wired Uniform Spanning Forest.

I Furthermore, to derive his theorem from this, Tucker-Drob makes use ofother probabilistic techniques:

(Gaboriau–Lyons) Wilson’s algorithm rooted at infinity

(essentially Hatami–Lovasz–Szegedy) an analogue for graphs of theAbert–Weiss theorem.

I All these techniques are inherently pmp.

I A year later, I found a combinatorial/purely descriptive set theoretic proofof Tucker-Drob’s theorem (for pmp graphs).

I Another half a year later, the argument became applicable to quasi-pmpgraphs, yielding a generalization.

Page 82: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Two different proofs

I Tucker-Drob’s proof of his theorem is based on a rather deep result inpercolation theory by Hutchcroft and Nachmias (2015) onindistinguishability of the Wired Uniform Spanning Forest.

I Furthermore, to derive his theorem from this, Tucker-Drob makes use ofother probabilistic techniques:

(Gaboriau–Lyons) Wilson’s algorithm rooted at infinity

(essentially Hatami–Lovasz–Szegedy) an analogue for graphs of theAbert–Weiss theorem.

I All these techniques are inherently pmp.

I A year later, I found a combinatorial/purely descriptive set theoretic proofof Tucker-Drob’s theorem (for pmp graphs).

I Another half a year later, the argument became applicable to quasi-pmpgraphs, yielding a generalization.

Page 83: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Two different proofs

I Tucker-Drob’s proof of his theorem is based on a rather deep result inpercolation theory by Hutchcroft and Nachmias (2015) onindistinguishability of the Wired Uniform Spanning Forest.

I Furthermore, to derive his theorem from this, Tucker-Drob makes use ofother probabilistic techniques:

(Gaboriau–Lyons) Wilson’s algorithm rooted at infinity

(essentially Hatami–Lovasz–Szegedy) an analogue for graphs of theAbert–Weiss theorem.

I All these techniques are inherently pmp.

I A year later, I found a combinatorial/purely descriptive set theoretic proofof Tucker-Drob’s theorem (for pmp graphs).

I Another half a year later, the argument became applicable to quasi-pmpgraphs, yielding a generalization.

Page 84: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

The main theorem (again) and what’s involved

Diagonalizing through a dense sequence in L1(X , µ) reduces the maintheorem to:

Theorem

Let G be a locally countable quasi-pmp ergodic Borel graph on (X , µ) andlet ρ : EG → R+ be the cocycle. For every f ∈ L1(X , µ) and ε > 0, there isa G -connected finite Borel equivalence relation F on X such that

Aρf [x ]F ≈ε

∫Xf dµ

for all but ε-measure-many x ∈ X .

The proof required new tools:

asymptotic averages along a graph

finitizing vertex-cuts (exploits nonhyperfiniteness)

saturated and packed prepartitions

an iteration technique via measure-compactness

for a cocycle ρ on EG , notions of ρ-ratio and (G , ρ)-visibility.

In the remaining time, I’ll discuss the difficulty of getting each F -classG -connected.

Page 85: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

The main theorem (again) and what’s involved

Diagonalizing through a dense sequence in L1(X , µ) reduces the maintheorem to:

Theorem

Let G be a locally countable quasi-pmp ergodic Borel graph on (X , µ) andlet ρ : EG → R+ be the cocycle. For every f ∈ L1(X , µ) and ε > 0, there isa G -connected finite Borel equivalence relation F on X such that

Aρf [x ]F ≈ε

∫Xf dµ

for all but ε-measure-many x ∈ X .

The proof required new tools:

asymptotic averages along a graph

finitizing vertex-cuts (exploits nonhyperfiniteness)

saturated and packed prepartitions

an iteration technique via measure-compactness

for a cocycle ρ on EG , notions of ρ-ratio and (G , ρ)-visibility.

In the remaining time, I’ll discuss the difficulty of getting each F -classG -connected.

Page 86: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

The main theorem (again) and what’s involved

Diagonalizing through a dense sequence in L1(X , µ) reduces the maintheorem to:

Theorem

Let G be a locally countable quasi-pmp ergodic Borel graph on (X , µ) andlet ρ : EG → R+ be the cocycle. For every f ∈ L1(X , µ) and ε > 0, there isa G -connected finite Borel equivalence relation F on X such that

Aρf [x ]F ≈ε

∫Xf dµ

for all but ε-measure-many x ∈ X .

The proof required new tools:

asymptotic averages along a graph

finitizing vertex-cuts (exploits nonhyperfiniteness)

saturated and packed prepartitions

an iteration technique via measure-compactness

for a cocycle ρ on EG , notions of ρ-ratio and (G , ρ)-visibility.

In the remaining time, I’ll discuss the difficulty of getting each F -classG -connected.

Page 87: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Borel G -prepartitions

I For the sake of the talk, assume G is pmp.

I It suffices to fix an f ∈ L∞(X , µ) and assume∫X f dµ = 0.

I Building a G -connected finite Borel equivalence relation amounts toconstructing a Borel collection P of pairwise disjoint finite G -connectedsubsets of X — call it a G -prepartition.

— We do not require the domain dom(P) ..=⋃P to be all of X .

I To prove the theorem, we need to build a Borel G -prepartition P with1− ε domain whose cells U ∈ P satisfy Af [U] ≈ε 0 — call these U good.

I By Kechris–Miller, there is always a Borel maximal such prepartition P .

I How much land does this maximal P conquer? That is: what’s themeasure of dom(P)?

I The first step in the proof is showing that dom(P) meets everyG -connected component.

I Thus, dom(P) has positive measure, but this measure could be arbitrarilysmall.

Page 88: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Borel G -prepartitions

I For the sake of the talk, assume G is pmp.

I It suffices to fix an f ∈ L∞(X , µ) and assume∫X f dµ = 0.

I Building a G -connected finite Borel equivalence relation amounts toconstructing a Borel collection P of pairwise disjoint finite G -connectedsubsets of X — call it a G -prepartition.

— We do not require the domain dom(P) ..=⋃P to be all of X .

I To prove the theorem, we need to build a Borel G -prepartition P with1− ε domain whose cells U ∈ P satisfy Af [U] ≈ε 0 — call these U good.

I By Kechris–Miller, there is always a Borel maximal such prepartition P .

I How much land does this maximal P conquer? That is: what’s themeasure of dom(P)?

I The first step in the proof is showing that dom(P) meets everyG -connected component.

I Thus, dom(P) has positive measure, but this measure could be arbitrarilysmall.

Page 89: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Borel G -prepartitions

I For the sake of the talk, assume G is pmp.

I It suffices to fix an f ∈ L∞(X , µ) and assume∫X f dµ = 0.

I Building a G -connected finite Borel equivalence relation amounts toconstructing a Borel collection P of pairwise disjoint finite G -connectedsubsets of X — call it a G -prepartition.

— We do not require the domain dom(P) ..=⋃P to be all of X .

I To prove the theorem, we need to build a Borel G -prepartition P with1− ε domain whose cells U ∈ P satisfy Af [U] ≈ε 0 — call these U good.

I By Kechris–Miller, there is always a Borel maximal such prepartition P .

I How much land does this maximal P conquer? That is: what’s themeasure of dom(P)?

I The first step in the proof is showing that dom(P) meets everyG -connected component.

I Thus, dom(P) has positive measure, but this measure could be arbitrarilysmall.

Page 90: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Borel G -prepartitions

I For the sake of the talk, assume G is pmp.

I It suffices to fix an f ∈ L∞(X , µ) and assume∫X f dµ = 0.

I Building a G -connected finite Borel equivalence relation amounts toconstructing a Borel collection P of pairwise disjoint finite G -connectedsubsets of X — call it a G -prepartition.

— We do not require the domain dom(P) ..=⋃P to be all of X .

I To prove the theorem, we need to build a Borel G -prepartition P with1− ε domain whose cells U ∈ P satisfy Af [U] ≈ε 0 — call these U good.

I By Kechris–Miller, there is always a Borel maximal such prepartition P .

I How much land does this maximal P conquer? That is: what’s themeasure of dom(P)?

I The first step in the proof is showing that dom(P) meets everyG -connected component.

I Thus, dom(P) has positive measure, but this measure could be arbitrarilysmall.

Page 91: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Borel G -prepartitions

I For the sake of the talk, assume G is pmp.

I It suffices to fix an f ∈ L∞(X , µ) and assume∫X f dµ = 0.

I Building a G -connected finite Borel equivalence relation amounts toconstructing a Borel collection P of pairwise disjoint finite G -connectedsubsets of X — call it a G -prepartition.

— We do not require the domain dom(P) ..=⋃P to be all of X .

I To prove the theorem, we need to build a Borel G -prepartition P with1− ε domain whose cells U ∈ P satisfy Af [U] ≈ε 0 — call these U good.

I By Kechris–Miller, there is always a Borel maximal such prepartition P .

I How much land does this maximal P conquer? That is: what’s themeasure of dom(P)?

I The first step in the proof is showing that dom(P) meets everyG -connected component.

I Thus, dom(P) has positive measure, but this measure could be arbitrarilysmall.

Page 92: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Borel G -prepartitions

I For the sake of the talk, assume G is pmp.

I It suffices to fix an f ∈ L∞(X , µ) and assume∫X f dµ = 0.

I Building a G -connected finite Borel equivalence relation amounts toconstructing a Borel collection P of pairwise disjoint finite G -connectedsubsets of X — call it a G -prepartition.

— We do not require the domain dom(P) ..=⋃P to be all of X .

I To prove the theorem, we need to build a Borel G -prepartition P with1− ε domain whose cells U ∈ P satisfy Af [U] ≈ε 0 — call these U good.

I By Kechris–Miller, there is always a Borel maximal such prepartition P .

I How much land does this maximal P conquer? That is: what’s themeasure of dom(P)?

I The first step in the proof is showing that dom(P) meets everyG -connected component.

I Thus, dom(P) has positive measure, but this measure could be arbitrarilysmall.

Page 93: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Borel G -prepartitions

I For the sake of the talk, assume G is pmp.

I It suffices to fix an f ∈ L∞(X , µ) and assume∫X f dµ = 0.

I Building a G -connected finite Borel equivalence relation amounts toconstructing a Borel collection P of pairwise disjoint finite G -connectedsubsets of X — call it a G -prepartition.

— We do not require the domain dom(P) ..=⋃P to be all of X .

I To prove the theorem, we need to build a Borel G -prepartition P with1− ε domain whose cells U ∈ P satisfy Af [U] ≈ε 0 — call these U good.

I By Kechris–Miller, there is always a Borel maximal such prepartition P .

I How much land does this maximal P conquer? That is: what’s themeasure of dom(P)?

I The first step in the proof is showing that dom(P) meets everyG -connected component.

I Thus, dom(P) has positive measure, but this measure could be arbitrarilysmall.

Page 94: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Borel G -prepartitions

I For the sake of the talk, assume G is pmp.

I It suffices to fix an f ∈ L∞(X , µ) and assume∫X f dµ = 0.

I Building a G -connected finite Borel equivalence relation amounts toconstructing a Borel collection P of pairwise disjoint finite G -connectedsubsets of X — call it a G -prepartition.

— We do not require the domain dom(P) ..=⋃P to be all of X .

I To prove the theorem, we need to build a Borel G -prepartition P with1− ε domain whose cells U ∈ P satisfy Af [U] ≈ε 0 — call these U good.

I By Kechris–Miller, there is always a Borel maximal such prepartition P .

I How much land does this maximal P conquer? That is: what’s themeasure of dom(P)?

I The first step in the proof is showing that dom(P) meets everyG -connected component.

I Thus, dom(P) has positive measure, but this measure could be arbitrarilysmall.

Page 95: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Maximal prepartition — not good enough

P — a Borel maximal G -prepartition into good cells U, i.e. Af [U] ≈ε 0.

I To prove the theorem, we need the measure of dom(P) to be 1− ε.

I However, dom(P) may leave out infinite G -connected components.

I How to get rid of these infinite clusters?

I Need a stronger notion of maximality — packed prepartitions!

I Going into these however is too much for a morning talk :)

Page 96: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Maximal prepartition — not good enough

P — a Borel maximal G -prepartition into good cells U, i.e. Af [U] ≈ε 0.

I To prove the theorem, we need the measure of dom(P) to be 1− ε.

I However, dom(P) may leave out infinite G -connected components.

I How to get rid of these infinite clusters?

I Need a stronger notion of maximality — packed prepartitions!

I Going into these however is too much for a morning talk :)

Page 97: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Maximal prepartition — not good enough

P — a Borel maximal G -prepartition into good cells U, i.e. Af [U] ≈ε 0.

I To prove the theorem, we need the measure of dom(P) to be 1− ε.

I However, dom(P) may leave out infinite G -connected components.

I How to get rid of these infinite clusters?

I Need a stronger notion of maximality — packed prepartitions!

I Going into these however is too much for a morning talk :)

Page 98: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Maximal prepartition — not good enough

P — a Borel maximal G -prepartition into good cells U, i.e. Af [U] ≈ε 0.

I To prove the theorem, we need the measure of dom(P) to be 1− ε.

I However, dom(P) may leave out infinite G -connected components.

I How to get rid of these infinite clusters?

I Need a stronger notion of maximality — packed prepartitions!

I Going into these however is too much for a morning talk :)

Page 99: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Maximal prepartition — not good enough

P — a Borel maximal G -prepartition into good cells U, i.e. Af [U] ≈ε 0.

I To prove the theorem, we need the measure of dom(P) to be 1− ε.

I However, dom(P) may leave out infinite G -connected components.

I How to get rid of these infinite clusters?

I Need a stronger notion of maximality — packed prepartitions!

I Going into these however is too much for a morning talk :)

Page 100: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Maximal prepartition — not good enough

P — a Borel maximal G -prepartition into good cells U, i.e. Af [U] ≈ε 0.

I To prove the theorem, we need the measure of dom(P) to be 1− ε.

I However, dom(P) may leave out infinite G -connected components.

I How to get rid of these infinite clusters?

I Need a stronger notion of maximality — packed prepartitions!

I Going into these however is too much for a morning talk :)

Page 101: Descriptive combinatorics and ergodic theoremsAnush Tserunyan University of Illinois at Urbana-Champaign “ Professor Kechris, descriptive set theory sure is powerful, and beautiful

Thanks!