/
Graph limit theory: an overview Graph limit theory: an overview

Graph limit theory: an overview - PowerPoint Presentation

faustina-dinatale
faustina-dinatale . @faustina-dinatale
Follow
471 views
Uploaded On 2016-05-16

Graph limit theory: an overview - PPT Presentation

L á szl ó Lov á sz Eötvös Lor ánd University Budapest IAS Princeton June 2011 1 June 2011 Limit theories of discrete structures trees graphs digraphs ID: 322555

june 2011 graphs graph 2011 june graph graphs convergent graphing limit bounded inv convergence dense graphings degree distribution local

Share:

Link:

Embed:

Download Presentation from below link

Download Presentation The PPT/PDF document "Graph limit theory: an overview" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.


Presentation Transcript

Slide1

Graph limit theory: an overview

László Lovász Eötvös Loránd University, BudapestIAS, Princeton

June 2011

1Slide2

June 2011

Limit theories of discrete structurestreesgraphsdigraphshypergraphspermutationsposets

abelian

groups

metric

spaces

rational numbers

Aldous, Elek-TardosDiaconis-JansonElek-SzegedyKohayakawaJansonSzegedyGromov Elek

2Slide3

June 2011

Common elements in limit theoriessamplingsampling distancelimiting sample

distributions

combined

limiting

sample

distributionslimit object

overlay distanceregularity lemmaapplications3treesgraphsdigraphshypergraphs

permutations

posets

abelian

groups

metric

spacesSlide4

June 2011

Limit theories for graphsDense graphs: Borgs-Chayes-L-Sós-Vesztergombi

L-Szegedy

Bounded

degree

graphs:

Benjamini-Schramm, Elek

Inbetween: distances Bollobás-Riordan regularity lemma Kohayakawa-Rödl, Scott Laplacian Chung

4Slide5

June 2011

Left and right datavery large graph

counting

edges

,

triangles,...

spectra,...

counting colorations,stable sets,statistical physics,maximum cut,...5Slide6

June 2011

distribution of k-samplesis convergent for every kt(F,G): Probability that random map

V(

F

)

V(G)

preserves edges(

G1,G2,…) convergent: F t(F,Gn) is convergent

Dense

graphs

:

convergence

6Slide7

June 2011

W0 = {W: [0,1]2 [0,1], symmetric, measurable}GnW :

F

:

t(F,

Gn)  t(

F,W)"graphon"Dense graphs: limit objects7Slide8

G

0 0 1 0 0 1 1 0 0 0 1 0 0 1

0 0 1 0 1 0 1 0 0 0 0 0 1 0

1 1 0 1 0 1 1 1 1 0 1 0 1 1

0 0 1 0 1 0 1 0 1 0 1 1 0 0

0 1 0 1 0 1 1 0 0 0 1 0 0 1

1 0 1 0 1 0 1 1 0 1 1 1 0 1

1 1 1 1 1 1 0 1 0 1 1 1 1 0

0 0 1 0 0 1 1 0 1 0 1 0 1 1

0 0 1 1 0 0 0 1 1 1 0 1 0 0

0 0 0 0 0 1 1 0 1 0 1 0 1 0

1 0 1 1 1 1 1 1 0 1 0 1 1 1

0 0 0 1 0 1 1 0 1 0 1 0 1 0

0 1 1 0 0 0 1 1 0 1 1 1 0 1

1 0 1 0 1 1 0 1 0 0 1 0 1 0

A

G

W

G

Graphs

to

graphons

May 2012

8Slide9

June 2011

Dense graphs: basic factsFor every convergent graph sequence (Gn)there is a W

W

0

such that

GnW

.W is essentially unique (up to measure-preserving transformation).Conversely,

W

(

G

n

)

such

that

G

n

W

.

Is

this

the

only

useful

notion

of

convergence

of

dense

graphs

?

9Slide10

June 2011

Bounded degree: convergenceLocal : neighborhood sampling Benjamini-SchrammGlobal : metric

space

Gromov

Local-global

: Hatami-L-Szegedy

Right-convergence,… Borgs-Chayes-Gamarnik10Slide11

June 2011

GraphingsGraphing: bounded degree graph G on [0,1] such that:

 E

(

G

) is a

Borel set

in [0,1]2

measure preserving: 01

deg

B

(x)=2

A

B

11Slide12

June 2011

GraphingsEvery Borel subgraph of a graphing is a graphing.Every

graph

you

ever

want to

construct from a graphing

is a graphingD=1: graphing  measure preserving involutionG is a graphing

G

=

G

1

… 

G

k

measure

preserving

involutions

(

k

2

D

-1)

12Slide13

June 2011

Graphings: examplesE(G)

= {chords

with

angle }

x

x-x+

V

(

G

)

=

circle

13Slide14

June 2011

Graphings: examples

V

(

G

)

= {

rooted

2-colored

grids

}

E

(

G

)

= {shift

the

root

}

14Slide15

June 2011

Graphings: examplesx

x

-

x

+

xx-x+

bipartite

?

disconnected

?

15Slide16

June 2011

Graphings and involution-invariant distributionsGx is a random connected graph with bounded degree

x

:

random

point

of [0,1]

Gx: connected

component of G containing xThis distribution is "invariant" under shifting the

root

.

Every

involution-invariant

distribution

can

be

represented

by

a

graphing

.

Elek

16Slide17

June 2011

Graph limits and involution-invariant distributionsgraphs, graphings,or

inv-inv

distributions

(

Gn

) locally convergent:

Cauchy in dGn  G: d (Gn

,

G

)

 0 (

n

 )

inv-inv

distribution

17Slide18

June 2011

Graph limits and involution-invariant distributionsEvery locally convergent sequence

of bounded-degree

graphs

has a limiting

inv-inv distribution

.Benjamini-Schramm

Is every inv-inv distribution the limit of a locally convergent graph sequence?

Aldous-Lyons

18Slide19

June 2011

Local-global convergence(Gn) locally-globally convergent: Cauchy in

d

k

G

n

 G:

dk(Gn,G)  0 (n  )graphing19Slide20

June 2011

Local-global graph limitsEvery locally-globally convergent sequence of bounded-degree

graphs

has a limit

graphing

.Hatami-L-Szegedy

20Slide21

June 2011

Convergence: examplesGn: random 3-regular graphFn: random 3-regular bipartite graphH

n:

G

n

Gn

Large

girth graphs

Expander

graphs

21Slide22

June 2011

Convergence: examplesLocal limit: Gn, Fn, Hn  rooted

3-regular tree

T

22

Conjecture

:

(

G

n

)

,

(

F

n

) and

(

H

n

)

are

locally-globally

convergent

.

Contains

recent

result

that

independence

ratio

is

convergent

.

Bayati-Gamarnik-TetaliSlide23

June 2011

Convergence: examplesLocal-global limit: Gn, Fn, Hn tend

to

different

graphings

Conjecture

:

G

n

T

{0,1},

where

V

(

T

)

= {

rooted

2-colored

trees

}

E

(

G

)

= {shift

the

root

}

23Slide24

June 2011

Local-global convergence: dense caseEvery convergent

sequence

of

graphs

is Cauchy

in d

kL-Vesztergombi

24Slide25

June 2011

Regularity lemmaGiven an arbitrarily large graph G and an >0,decompose

G

into

f

()

"homogeneous" parts.

(,)-homogeneous graph: SE(G), |S|<|V(

G

)|,

all

connected

components

of

G

-

S

with

> |

V

(

G

)|

nodes

have

the

same

neighborhood

distribution

(

up

to

).

25Slide26

June 2011

Regularity

lemma

n

x

n

grid is (, 

2/18)-homogeneous.>0 >0 bounded-deg G S 

E

(

G

), |

S

|<|

V

(

G

)|,

st

.

all

components

of

G

-

S

are

(,)

-homogeneous

.

Angel-Szegedy

,

Elek-Lippner

26Slide27

June 2011

Regularity lemmaGiven an arbitrarily large graph G and an >0,find

a graph

H

of

size

at most f() such

thatG and H are -close in sampling distance.

Frieze-Kannan

"

Weak

"

Regularity

Lemma

suffices

in

the

dense

case

.

f

(

)

exists

in

the

bounded

degree

case

.

Alon

27Slide28

June 2011

Extremal graph theory It is undecidable whether

holds

for

every graph

G.

Hatami-NorinIt is undecidable whether there is a graphing

with

almost

all

r

-neighborhoods

in

a

given

family

F

.

Csóka

28Slide29

1

1

0

Kruskal-Katona

Bollob

ás

1/2

2/3

3/4

Razborov 2006

Mantel-Tur

án

Goodman

Fisher

Lov

ász-Simonovits

June 2011

29

Extremal

graph

theory

:

dense

graphs

Slide30

D

3/8D2/60June 201130Extremal

graph

theory

:

D-regular

Harangi