SlideShare a Scribd company logo
Models of Graphs
Jérôme Kunegis
Oberseminar
2013-08-29
Jérôme Kunegis Models of Graphs 2
Erdős–Rényi
Each edge has
probability p of existing
P(G) = pm
(1 − p)(M − m)
m = #edges
M = max possible #edges
Jérôme Kunegis Models of Graphs 3
Barabási–Albert
An edge appears with probability
proportional to the degree of the
node it connects
P({u, v}) d(u)∼
d(u) = degree of node u
Jérôme Kunegis Models of Graphs 4
What Everybody Thinks
My network model leads to graphs
that have the same properties as
actual social networks
Hmmm...
Jérôme Kunegis Models of Graphs 5
P(G) = pm
(1 − p)(M − m)
P({u, v}) d(u)∼
Why don't you use
the same formalism??
Comparison
Jérôme Kunegis Models of Graphs 6
Formalisms for Graph Models
(1) Specify a graph generation algorithm
(2) Specify a graph growth algorithm
(3) Specify the probability of any graph
(4) Specify the probability of any edge
(5) Specify the probability of any event
(6) Specify a score for node pairs
(7) Matrix model
(8) Graph compression
Jérôme Kunegis Models of Graphs 7
(1) Specify a Graph Generation Algorithm
STEP 1: Specify rules for generating a graph
Take a lattice, and rewire
a certain proportion
of edges randomly
EXAMPLE: small-world model (Watts & Strogatz 1998)
STEP 2: Generate random graph(s)
STEP 3: Compare with actual networks
Hey, a small diameter and
large clustering coefficient!
●
Not generative
●
Not probabilistic
Jérôme Kunegis Models of Graphs 8
(2) Specify a Graph Growth Algorithm
An edge appears with probability proportional
to the degree with probability p and at
random with probability (1 − p)
STEP 1: Specify exact growth rules
STEP 2: Generate random graph(s)
STEP 3: Compare with actual networks
Look, a power law!
EXAMPLE: preferential attachment (Barabási & Albert 1999)
●
No overall probability
Jérôme Kunegis Models of Graphs 9
What We Need: A Probabilistic Model
A probabilistic model assigns a probability to each possible value.
X: set of possible values
x ∈ X: a value
p: A parameter of the model
P(x; p): Probability of x, given p, OR
Likelihood of p, given x
Σx∈X P(x; p) = 1 // Because P is a distribution for a given p
Given a set of values {xi} for i = 1, … N, the best fitting p can be found by
maximum likelihood:
maxp Πi P(xi, p)
So, are “values” whole graphs
or individual edges?
Jérôme Kunegis Models of Graphs 10
(3) Specify the Probability of Any Graph
Each edge has
probability p of existing
STEP 1: Specify the probability of any graph G
●
Not generative
●
Needs multiple graphs for inference
STEP 2: Given a set of graphs with the same number of
nodes, compute the likelihood of any value p
EXAMPLE: (Erdős & Rényi 1959)
Jérôme Kunegis Models of Graphs 11
Example: Extension of Erdős–Rényi using Formalism (3)
Goal: Add a parameter that controls the number of triangles.
Idea: The E–R model with parameter p is an exponential family; the extension
should be too.
P(G) = (1 / C) pm
(1 − p)(M − m)
qt
(1 − q)(T − t)
where t is the #triangles, T is the maximum possible #triangles.
Note: q = 1/2 gives the ordinary E–R model.
Result: exponential random graph models (ERGM) and p* models
The normalization constant C cannot be computed.
It would be necessary to count the number of graphs with
n vertices, m edges and t triangles. This is a hard, open problem.
Gibbs sampling works, however.
Open problem: Use Gibbs sampling to generate mini-models of networks.
Jérôme Kunegis Models of Graphs 12
(4) Specify the Probability of Any Edge
STEP 1: Specify probability for all pairs {u, v}
EXAMPLE: Use a given degree vector d as parameter, and P({u, v}) = du dv
EXAMPLE: The p1 model based on node attributes (Holland & Leinhard 1977)
STEP 2: Compute likelihood of parameters
●
Not generative
Let's model each
edge as an event,
not a full graph
●
Supports multiple edges
Jérôme Kunegis Models of Graphs 13
Preliminary Results for Formalism (4)
The best rank-1 model is given by the preferential attachment model.
Let a graph G be given. Among all models of the form P({u, v}) = x xT
,
the one with maximum likelihood is given by
P({u, v}) = d(u) d(v) / 2m
Proof: By induction over n.
Open problem: define other models using this formalism
Hey, that's different
from minimizing the least
squares distance to the
given adjacency matrix,
where the SVD is best
Jérôme Kunegis Models of Graphs 14
(5) Specify the Probability of Any Event
Let's specify the probability
of an edge addition,
given the current graph
STEP 1: Specify the probability of an edge addition given the current graph
EXAMPLE: P({u, v}) = p / n² + (1 − p) d(u) d(v) / 2m
STEP 2: Compute the likelihood
OTHER EXAMPLE: (Akkermans & al. 2012)
Open problem: Inference of parameters from real networks.
Generalizes naturally to edge removal events.
Jérôme Kunegis Models of Graphs 15
(6) Specify a Score for Node Pairs
Read my paper
STEP 1: Given a graph, specify a score for each node pairs
STEP 2: Evaluate using information retrieval methods
I know, that's link prediction!
●
Not probabilistic
(Liben-Nowell & Kleinberg 2003)
Jérôme Kunegis Models of Graphs 16
(7) Matrix Model
STEP 1: Specify a probability matrix
STEP 2: Map nodes of the graph to rows/columns of the matrix
STEP 3: Compute the likelihood
Let's try the Kronecker product
EXAMPLE: (Leskovec & al. 2005)
●
Not generative
Can I do this with any matrix?
Jérôme Kunegis Models of Graphs 17
(8) Graph Compression
STEP 1: Specify a graph compression algorithm
STEP 2: Check how well it compresses a graph
(Shannon)
More probable values should
have shorter representations
I wonder how the E-R model
can be used here
●
Not generative
Now let's
do some
research!
SUMMARY
(1) Graph generation (e.g., Watts–Strogatz)
(2) Graph growth (e.g., Barabási–Albert)
(3) Graph probability (e.g., Erdős–Rényi)
(4) Edge probability
(5) Event probability
(6) Edge score (link prediction)
(7) Matrix model (e.g., Leskovec & al.)
(8) Graph compression
Inference
Mini-models
Rank-2 model
Spectral model
supercededby
Equivalence

More Related Content

What's hot

I. AO* SEARCH ALGORITHM
I. AO* SEARCH ALGORITHMI. AO* SEARCH ALGORITHM
I. AO* SEARCH ALGORITHM
vikas dhakane
 
Problem reduction AND OR GRAPH & AO* algorithm.ppt
Problem reduction AND OR GRAPH & AO* algorithm.pptProblem reduction AND OR GRAPH & AO* algorithm.ppt
Problem reduction AND OR GRAPH & AO* algorithm.ppt
arunsingh660
 
Permutation graphsandapplications
Permutation graphsandapplicationsPermutation graphsandapplications
Permutation graphsandapplicationsJoe Krall
 
Matlab integration
Matlab integrationMatlab integration
Matlab integration
pramodkumar1804
 
Lesson 28: The Fundamental Theorem of Calculus
Lesson 28: The Fundamental Theorem of CalculusLesson 28: The Fundamental Theorem of Calculus
Lesson 28: The Fundamental Theorem of CalculusMatthew Leingang
 
10_rnn.pdf
10_rnn.pdf10_rnn.pdf
14_autoencoders.pdf
14_autoencoders.pdf14_autoencoders.pdf
14_autoencoders.pdf
KSChidanandKumarJSSS
 
15_representation.pdf
15_representation.pdf15_representation.pdf
15_representation.pdf
KSChidanandKumarJSSS
 
Stratified sampling and resampling for approximate Bayesian computation
Stratified sampling and resampling for approximate Bayesian computationStratified sampling and resampling for approximate Bayesian computation
Stratified sampling and resampling for approximate Bayesian computation
Umberto Picchini
 
And or graph problem reduction using predicate logic
And or graph problem reduction using predicate logicAnd or graph problem reduction using predicate logic
And or graph problem reduction using predicate logic
Mohanlal Sukhadia University (MLSU)
 
hospital management
hospital managementhospital management
hospital management
guestbcbbb5c
 
Deep Learning Opening Workshop - ProxSARAH Algorithms for Stochastic Composit...
Deep Learning Opening Workshop - ProxSARAH Algorithms for Stochastic Composit...Deep Learning Opening Workshop - ProxSARAH Algorithms for Stochastic Composit...
Deep Learning Opening Workshop - ProxSARAH Algorithms for Stochastic Composit...
The Statistical and Applied Mathematical Sciences Institute
 
Dijkstra algorithm a dynammic programming approach
Dijkstra algorithm   a dynammic programming approachDijkstra algorithm   a dynammic programming approach
Dijkstra algorithm a dynammic programming approach
Akash Sethiya
 
Matlab lecture 8 – newton's forward and backword interpolation@taj copy
Matlab lecture 8 – newton's forward and backword interpolation@taj   copyMatlab lecture 8 – newton's forward and backword interpolation@taj   copy
Matlab lecture 8 – newton's forward and backword interpolation@taj copy
Tajim Md. Niamat Ullah Akhund
 
A* Search Algorithm
A* Search AlgorithmA* Search Algorithm
A* Search Algorithm
vikas dhakane
 
Csr2011 june14 11_00_aaronson
Csr2011 june14 11_00_aaronsonCsr2011 june14 11_00_aaronson
Csr2011 june14 11_00_aaronson
CSR2011
 
A* Algorithm
A* AlgorithmA* Algorithm
A* Algorithm
Dr. C.V. Suresh Babu
 
Computer Science Assignment Help
Computer Science Assignment Help Computer Science Assignment Help
Computer Science Assignment Help
Programming Homework Help
 
13_linear_factors.pdf
13_linear_factors.pdf13_linear_factors.pdf
13_linear_factors.pdf
KSChidanandKumarJSSS
 

What's hot (19)

I. AO* SEARCH ALGORITHM
I. AO* SEARCH ALGORITHMI. AO* SEARCH ALGORITHM
I. AO* SEARCH ALGORITHM
 
Problem reduction AND OR GRAPH & AO* algorithm.ppt
Problem reduction AND OR GRAPH & AO* algorithm.pptProblem reduction AND OR GRAPH & AO* algorithm.ppt
Problem reduction AND OR GRAPH & AO* algorithm.ppt
 
Permutation graphsandapplications
Permutation graphsandapplicationsPermutation graphsandapplications
Permutation graphsandapplications
 
Matlab integration
Matlab integrationMatlab integration
Matlab integration
 
Lesson 28: The Fundamental Theorem of Calculus
Lesson 28: The Fundamental Theorem of CalculusLesson 28: The Fundamental Theorem of Calculus
Lesson 28: The Fundamental Theorem of Calculus
 
10_rnn.pdf
10_rnn.pdf10_rnn.pdf
10_rnn.pdf
 
14_autoencoders.pdf
14_autoencoders.pdf14_autoencoders.pdf
14_autoencoders.pdf
 
15_representation.pdf
15_representation.pdf15_representation.pdf
15_representation.pdf
 
Stratified sampling and resampling for approximate Bayesian computation
Stratified sampling and resampling for approximate Bayesian computationStratified sampling and resampling for approximate Bayesian computation
Stratified sampling and resampling for approximate Bayesian computation
 
And or graph problem reduction using predicate logic
And or graph problem reduction using predicate logicAnd or graph problem reduction using predicate logic
And or graph problem reduction using predicate logic
 
hospital management
hospital managementhospital management
hospital management
 
Deep Learning Opening Workshop - ProxSARAH Algorithms for Stochastic Composit...
Deep Learning Opening Workshop - ProxSARAH Algorithms for Stochastic Composit...Deep Learning Opening Workshop - ProxSARAH Algorithms for Stochastic Composit...
Deep Learning Opening Workshop - ProxSARAH Algorithms for Stochastic Composit...
 
Dijkstra algorithm a dynammic programming approach
Dijkstra algorithm   a dynammic programming approachDijkstra algorithm   a dynammic programming approach
Dijkstra algorithm a dynammic programming approach
 
Matlab lecture 8 – newton's forward and backword interpolation@taj copy
Matlab lecture 8 – newton's forward and backword interpolation@taj   copyMatlab lecture 8 – newton's forward and backword interpolation@taj   copy
Matlab lecture 8 – newton's forward and backword interpolation@taj copy
 
A* Search Algorithm
A* Search AlgorithmA* Search Algorithm
A* Search Algorithm
 
Csr2011 june14 11_00_aaronson
Csr2011 june14 11_00_aaronsonCsr2011 june14 11_00_aaronson
Csr2011 june14 11_00_aaronson
 
A* Algorithm
A* AlgorithmA* Algorithm
A* Algorithm
 
Computer Science Assignment Help
Computer Science Assignment Help Computer Science Assignment Help
Computer Science Assignment Help
 
13_linear_factors.pdf
13_linear_factors.pdf13_linear_factors.pdf
13_linear_factors.pdf
 

Viewers also liked

KONECT Cloud – Large Scale Network Mining in the Cloud
KONECT Cloud – Large Scale Network Mining in the CloudKONECT Cloud – Large Scale Network Mining in the Cloud
KONECT Cloud – Large Scale Network Mining in the Cloud
Jérôme KUNEGIS
 
Algebraic Graph-theoretic Measures of Conflict
Algebraic Graph-theoretic Measures of ConflictAlgebraic Graph-theoretic Measures of Conflict
Algebraic Graph-theoretic Measures of Conflict
Jérôme KUNEGIS
 
What Is the Added Value of Negative Links in Online Social Networks?
What Is the Added Value of Negative Links in Online Social Networks?What Is the Added Value of Negative Links in Online Social Networks?
What Is the Added Value of Negative Links in Online Social Networks?
Jérôme KUNEGIS
 
Winning Science Slam by Jérôme Kunegis – First Prize at ICWSM 2016
Winning Science Slam by Jérôme Kunegis – First Prize at ICWSM 2016Winning Science Slam by Jérôme Kunegis – First Prize at ICWSM 2016
Winning Science Slam by Jérôme Kunegis – First Prize at ICWSM 2016
Jérôme KUNEGIS
 
Why Beyoncé Is More Popular Than Me – Fairness, Diversity and Other Measures
Why Beyoncé Is More Popular Than Me – Fairness, Diversity and Other MeasuresWhy Beyoncé Is More Popular Than Me – Fairness, Diversity and Other Measures
Why Beyoncé Is More Popular Than Me – Fairness, Diversity and Other MeasuresJérôme KUNEGIS
 
Learning Spectral Graph Transformations for Link Prediction
Learning Spectral Graph Transformations for Link PredictionLearning Spectral Graph Transformations for Link Prediction
Learning Spectral Graph Transformations for Link Prediction
Jérôme KUNEGIS
 

Viewers also liked (6)

KONECT Cloud – Large Scale Network Mining in the Cloud
KONECT Cloud – Large Scale Network Mining in the CloudKONECT Cloud – Large Scale Network Mining in the Cloud
KONECT Cloud – Large Scale Network Mining in the Cloud
 
Algebraic Graph-theoretic Measures of Conflict
Algebraic Graph-theoretic Measures of ConflictAlgebraic Graph-theoretic Measures of Conflict
Algebraic Graph-theoretic Measures of Conflict
 
What Is the Added Value of Negative Links in Online Social Networks?
What Is the Added Value of Negative Links in Online Social Networks?What Is the Added Value of Negative Links in Online Social Networks?
What Is the Added Value of Negative Links in Online Social Networks?
 
Winning Science Slam by Jérôme Kunegis – First Prize at ICWSM 2016
Winning Science Slam by Jérôme Kunegis – First Prize at ICWSM 2016Winning Science Slam by Jérôme Kunegis – First Prize at ICWSM 2016
Winning Science Slam by Jérôme Kunegis – First Prize at ICWSM 2016
 
Why Beyoncé Is More Popular Than Me – Fairness, Diversity and Other Measures
Why Beyoncé Is More Popular Than Me – Fairness, Diversity and Other MeasuresWhy Beyoncé Is More Popular Than Me – Fairness, Diversity and Other Measures
Why Beyoncé Is More Popular Than Me – Fairness, Diversity and Other Measures
 
Learning Spectral Graph Transformations for Link Prediction
Learning Spectral Graph Transformations for Link PredictionLearning Spectral Graph Transformations for Link Prediction
Learning Spectral Graph Transformations for Link Prediction
 

Similar to Eight Formalisms for Defining Graph Models

An Importance Sampling Approach to Integrate Expert Knowledge When Learning B...
An Importance Sampling Approach to Integrate Expert Knowledge When Learning B...An Importance Sampling Approach to Integrate Expert Knowledge When Learning B...
An Importance Sampling Approach to Integrate Expert Knowledge When Learning B...
NTNU
 
Optimal L-shaped matrix reordering, aka graph's core-periphery
Optimal L-shaped matrix reordering, aka graph's core-peripheryOptimal L-shaped matrix reordering, aka graph's core-periphery
Optimal L-shaped matrix reordering, aka graph's core-periphery
Francesco Tudisco
 
QMC: Transition Workshop - Monte Carlo and (Randomized) Quasi-Monte Carlo Sim...
QMC: Transition Workshop - Monte Carlo and (Randomized) Quasi-Monte Carlo Sim...QMC: Transition Workshop - Monte Carlo and (Randomized) Quasi-Monte Carlo Sim...
QMC: Transition Workshop - Monte Carlo and (Randomized) Quasi-Monte Carlo Sim...
The Statistical and Applied Mathematical Sciences Institute
 
Exhaustive Combinatorial Enumeration
Exhaustive Combinatorial EnumerationExhaustive Combinatorial Enumeration
Exhaustive Combinatorial Enumeration
Mathieu Dutour Sikiric
 
Lego like spheres and tori, enumeration and drawings
Lego like spheres and tori, enumeration and drawingsLego like spheres and tori, enumeration and drawings
Lego like spheres and tori, enumeration and drawings
Mathieu Dutour Sikiric
 
Prim algorithm for the implementation of random mazes in videogames
Prim algorithm for the  implementation of random mazes  in videogamesPrim algorithm for the  implementation of random mazes  in videogames
Prim algorithm for the implementation of random mazes in videogamesFélix Santos
 
Accelerating Metropolis Hastings with Lightweight Inference Compilation
Accelerating Metropolis Hastings with Lightweight Inference CompilationAccelerating Metropolis Hastings with Lightweight Inference Compilation
Accelerating Metropolis Hastings with Lightweight Inference Compilation
Feynman Liang
 
Learning multifractal structure in large networks (KDD 2014)
Learning multifractal structure in large networks (KDD 2014)Learning multifractal structure in large networks (KDD 2014)
Learning multifractal structure in large networks (KDD 2014)
Austin Benson
 
Scalable MAP inference in Bayesian networks based on a Map-Reduce approach (P...
Scalable MAP inference in Bayesian networks based on a Map-Reduce approach (P...Scalable MAP inference in Bayesian networks based on a Map-Reduce approach (P...
Scalable MAP inference in Bayesian networks based on a Map-Reduce approach (P...
AMIDST Toolbox
 
Simulated annealing for MMR-Path
Simulated annealing for MMR-PathSimulated annealing for MMR-Path
Simulated annealing for MMR-PathFrancisco Pérez
 
Spline interpolation numerical methods presentation
Spline interpolation numerical methods presentationSpline interpolation numerical methods presentation
Spline interpolation numerical methods presentation
Shohanur Nishad
 
MUMS: Bayesian, Fiducial, and Frequentist Conference - Model Selection in the...
MUMS: Bayesian, Fiducial, and Frequentist Conference - Model Selection in the...MUMS: Bayesian, Fiducial, and Frequentist Conference - Model Selection in the...
MUMS: Bayesian, Fiducial, and Frequentist Conference - Model Selection in the...
The Statistical and Applied Mathematical Sciences Institute
 
d-VMP: Distributed Variational Message Passing (PGM2016)
d-VMP: Distributed Variational Message Passing (PGM2016)d-VMP: Distributed Variational Message Passing (PGM2016)
d-VMP: Distributed Variational Message Passing (PGM2016)
AMIDST Toolbox
 
MUMS Opening Workshop - An Overview of Reduced-Order Models and Emulators (ED...
MUMS Opening Workshop - An Overview of Reduced-Order Models and Emulators (ED...MUMS Opening Workshop - An Overview of Reduced-Order Models and Emulators (ED...
MUMS Opening Workshop - An Overview of Reduced-Order Models and Emulators (ED...
The Statistical and Applied Mathematical Sciences Institute
 
Mit15 082 jf10_lec01
Mit15 082 jf10_lec01Mit15 082 jf10_lec01
Mit15 082 jf10_lec01Saad Liaqat
 
02 math essentials
02 math essentials02 math essentials
02 math essentials
Poongodi Mano
 
Modelling the Clustering Coefficient of a Random graph
Modelling the Clustering Coefficient of a Random graphModelling the Clustering Coefficient of a Random graph
Modelling the Clustering Coefficient of a Random graph
Graph-TA
 

Similar to Eight Formalisms for Defining Graph Models (20)

An Importance Sampling Approach to Integrate Expert Knowledge When Learning B...
An Importance Sampling Approach to Integrate Expert Knowledge When Learning B...An Importance Sampling Approach to Integrate Expert Knowledge When Learning B...
An Importance Sampling Approach to Integrate Expert Knowledge When Learning B...
 
Optimal L-shaped matrix reordering, aka graph's core-periphery
Optimal L-shaped matrix reordering, aka graph's core-peripheryOptimal L-shaped matrix reordering, aka graph's core-periphery
Optimal L-shaped matrix reordering, aka graph's core-periphery
 
QMC: Transition Workshop - Monte Carlo and (Randomized) Quasi-Monte Carlo Sim...
QMC: Transition Workshop - Monte Carlo and (Randomized) Quasi-Monte Carlo Sim...QMC: Transition Workshop - Monte Carlo and (Randomized) Quasi-Monte Carlo Sim...
QMC: Transition Workshop - Monte Carlo and (Randomized) Quasi-Monte Carlo Sim...
 
Exhaustive Combinatorial Enumeration
Exhaustive Combinatorial EnumerationExhaustive Combinatorial Enumeration
Exhaustive Combinatorial Enumeration
 
Lego like spheres and tori, enumeration and drawings
Lego like spheres and tori, enumeration and drawingsLego like spheres and tori, enumeration and drawings
Lego like spheres and tori, enumeration and drawings
 
Planted Clique Research Paper
Planted Clique Research PaperPlanted Clique Research Paper
Planted Clique Research Paper
 
Prim algorithm for the implementation of random mazes in videogames
Prim algorithm for the  implementation of random mazes  in videogamesPrim algorithm for the  implementation of random mazes  in videogames
Prim algorithm for the implementation of random mazes in videogames
 
Accelerating Metropolis Hastings with Lightweight Inference Compilation
Accelerating Metropolis Hastings with Lightweight Inference CompilationAccelerating Metropolis Hastings with Lightweight Inference Compilation
Accelerating Metropolis Hastings with Lightweight Inference Compilation
 
Learning multifractal structure in large networks (KDD 2014)
Learning multifractal structure in large networks (KDD 2014)Learning multifractal structure in large networks (KDD 2014)
Learning multifractal structure in large networks (KDD 2014)
 
Scalable MAP inference in Bayesian networks based on a Map-Reduce approach (P...
Scalable MAP inference in Bayesian networks based on a Map-Reduce approach (P...Scalable MAP inference in Bayesian networks based on a Map-Reduce approach (P...
Scalable MAP inference in Bayesian networks based on a Map-Reduce approach (P...
 
Simulated annealing for MMR-Path
Simulated annealing for MMR-PathSimulated annealing for MMR-Path
Simulated annealing for MMR-Path
 
Spline interpolation numerical methods presentation
Spline interpolation numerical methods presentationSpline interpolation numerical methods presentation
Spline interpolation numerical methods presentation
 
MUMS: Bayesian, Fiducial, and Frequentist Conference - Model Selection in the...
MUMS: Bayesian, Fiducial, and Frequentist Conference - Model Selection in the...MUMS: Bayesian, Fiducial, and Frequentist Conference - Model Selection in the...
MUMS: Bayesian, Fiducial, and Frequentist Conference - Model Selection in the...
 
Lec 2-2
Lec 2-2Lec 2-2
Lec 2-2
 
algorithm Unit 3
algorithm Unit 3algorithm Unit 3
algorithm Unit 3
 
d-VMP: Distributed Variational Message Passing (PGM2016)
d-VMP: Distributed Variational Message Passing (PGM2016)d-VMP: Distributed Variational Message Passing (PGM2016)
d-VMP: Distributed Variational Message Passing (PGM2016)
 
MUMS Opening Workshop - An Overview of Reduced-Order Models and Emulators (ED...
MUMS Opening Workshop - An Overview of Reduced-Order Models and Emulators (ED...MUMS Opening Workshop - An Overview of Reduced-Order Models and Emulators (ED...
MUMS Opening Workshop - An Overview of Reduced-Order Models and Emulators (ED...
 
Mit15 082 jf10_lec01
Mit15 082 jf10_lec01Mit15 082 jf10_lec01
Mit15 082 jf10_lec01
 
02 math essentials
02 math essentials02 math essentials
02 math essentials
 
Modelling the Clustering Coefficient of a Random graph
Modelling the Clustering Coefficient of a Random graphModelling the Clustering Coefficient of a Random graph
Modelling the Clustering Coefficient of a Random graph
 

More from Jérôme KUNEGIS

Succinct Summarisation of Large Networks via Small Synthetic Representative G...
Succinct Summarisation of Large Networks via Small Synthetic Representative G...Succinct Summarisation of Large Networks via Small Synthetic Representative G...
Succinct Summarisation of Large Networks via Small Synthetic Representative G...
Jérôme KUNEGIS
 
Title: What Is the Difference between a Social and a Hyperlink Network? -- Ho...
Title: What Is the Difference between a Social and a Hyperlink Network? -- Ho...Title: What Is the Difference between a Social and a Hyperlink Network? -- Ho...
Title: What Is the Difference between a Social and a Hyperlink Network? -- Ho...
Jérôme KUNEGIS
 
Measuring the Conflict in a Social Network with Friends and Foes: A Recent Al...
Measuring the Conflict in a Social Network with Friends and Foes: A Recent Al...Measuring the Conflict in a Social Network with Friends and Foes: A Recent Al...
Measuring the Conflict in a Social Network with Friends and Foes: A Recent Al...
Jérôme KUNEGIS
 
Schach und Computer
Schach und ComputerSchach und Computer
Schach und Computer
Jérôme KUNEGIS
 
Generating Networks with Arbitrary Properties
Generating Networks with Arbitrary PropertiesGenerating Networks with Arbitrary Properties
Generating Networks with Arbitrary PropertiesJérôme KUNEGIS
 
Karriere Lounge – INFORMATIK 2013
Karriere Lounge – INFORMATIK 2013Karriere Lounge – INFORMATIK 2013
Karriere Lounge – INFORMATIK 2013
Jérôme KUNEGIS
 
KONECT – The Koblenz Network Collection
KONECT – The Koblenz Network CollectionKONECT – The Koblenz Network Collection
KONECT – The Koblenz Network Collection
Jérôme KUNEGIS
 
Preferential Attachment in Online Networks: Measurement and Explanations
Preferential Attachment in Online Networks:  Measurement and ExplanationsPreferential Attachment in Online Networks:  Measurement and Explanations
Preferential Attachment in Online Networks: Measurement and Explanations
Jérôme KUNEGIS
 
Predicting Directed Links using Nondiagonal Matrix Decompositions
Predicting Directed Links using Nondiagonal Matrix DecompositionsPredicting Directed Links using Nondiagonal Matrix Decompositions
Predicting Directed Links using Nondiagonal Matrix DecompositionsJérôme KUNEGIS
 
Online Dating Recommender Systems: The Split-complex Number Approach
Online Dating Recommender Systems: The Split-complex Number ApproachOnline Dating Recommender Systems: The Split-complex Number Approach
Online Dating Recommender Systems: The Split-complex Number Approach
Jérôme KUNEGIS
 
Fairness on the Web: Alternatives to the Power Law (WebSci 2012)
Fairness on the Web:  Alternatives to the Power Law (WebSci 2012)Fairness on the Web:  Alternatives to the Power Law (WebSci 2012)
Fairness on the Web: Alternatives to the Power Law (WebSci 2012)Jérôme KUNEGIS
 
Fairness on the Web: Alternatives to the Power Law
Fairness on the Web:  Alternatives to the Power LawFairness on the Web:  Alternatives to the Power Law
Fairness on the Web: Alternatives to the Power Law
Jérôme KUNEGIS
 
On the Spectral Evolution of Large Networks (PhD Thesis by Jérôme Kunegis)
On the Spectral Evolution of Large Networks (PhD Thesis by Jérôme Kunegis)On the Spectral Evolution of Large Networks (PhD Thesis by Jérôme Kunegis)
On the Spectral Evolution of Large Networks (PhD Thesis by Jérôme Kunegis)
Jérôme KUNEGIS
 
Searching Microblogs: Coping with Sparsity and Document Quality
Searching Microblogs: Coping with Sparsity and Document QualitySearching Microblogs: Coping with Sparsity and Document Quality
Searching Microblogs: Coping with Sparsity and Document QualityJérôme KUNEGIS
 
Bad News Travel Fast: A Content-based Analysis of Interestingness on Twitter
Bad News Travel Fast: A Content-based Analysis of Interestingness on TwitterBad News Travel Fast: A Content-based Analysis of Interestingness on Twitter
Bad News Travel Fast: A Content-based Analysis of Interestingness on Twitter
Jérôme KUNEGIS
 
On the Scalability of Graph Kernels Applied to Collaborative Recommenders
On the Scalability of Graph Kernels Applied to Collaborative RecommendersOn the Scalability of Graph Kernels Applied to Collaborative Recommenders
On the Scalability of Graph Kernels Applied to Collaborative Recommenders
Jérôme KUNEGIS
 
The Slashdot Zoo: Mining a Social Network with Negative Edges
The Slashdot Zoo:  Mining a Social Network with Negative EdgesThe Slashdot Zoo:  Mining a Social Network with Negative Edges
The Slashdot Zoo: Mining a Social Network with Negative Edges
Jérôme KUNEGIS
 
Spectral Analysis of Signed Graphs for Clustering, Prediction and Visualization
Spectral Analysis of Signed Graphs for Clustering, Prediction and VisualizationSpectral Analysis of Signed Graphs for Clustering, Prediction and Visualization
Spectral Analysis of Signed Graphs for Clustering, Prediction and Visualization
Jérôme KUNEGIS
 
Network Growth and the Spectral Evolution Model
Network Growth and the Spectral Evolution ModelNetwork Growth and the Spectral Evolution Model
Network Growth and the Spectral Evolution Model
Jérôme KUNEGIS
 

More from Jérôme KUNEGIS (19)

Succinct Summarisation of Large Networks via Small Synthetic Representative G...
Succinct Summarisation of Large Networks via Small Synthetic Representative G...Succinct Summarisation of Large Networks via Small Synthetic Representative G...
Succinct Summarisation of Large Networks via Small Synthetic Representative G...
 
Title: What Is the Difference between a Social and a Hyperlink Network? -- Ho...
Title: What Is the Difference between a Social and a Hyperlink Network? -- Ho...Title: What Is the Difference between a Social and a Hyperlink Network? -- Ho...
Title: What Is the Difference between a Social and a Hyperlink Network? -- Ho...
 
Measuring the Conflict in a Social Network with Friends and Foes: A Recent Al...
Measuring the Conflict in a Social Network with Friends and Foes: A Recent Al...Measuring the Conflict in a Social Network with Friends and Foes: A Recent Al...
Measuring the Conflict in a Social Network with Friends and Foes: A Recent Al...
 
Schach und Computer
Schach und ComputerSchach und Computer
Schach und Computer
 
Generating Networks with Arbitrary Properties
Generating Networks with Arbitrary PropertiesGenerating Networks with Arbitrary Properties
Generating Networks with Arbitrary Properties
 
Karriere Lounge – INFORMATIK 2013
Karriere Lounge – INFORMATIK 2013Karriere Lounge – INFORMATIK 2013
Karriere Lounge – INFORMATIK 2013
 
KONECT – The Koblenz Network Collection
KONECT – The Koblenz Network CollectionKONECT – The Koblenz Network Collection
KONECT – The Koblenz Network Collection
 
Preferential Attachment in Online Networks: Measurement and Explanations
Preferential Attachment in Online Networks:  Measurement and ExplanationsPreferential Attachment in Online Networks:  Measurement and Explanations
Preferential Attachment in Online Networks: Measurement and Explanations
 
Predicting Directed Links using Nondiagonal Matrix Decompositions
Predicting Directed Links using Nondiagonal Matrix DecompositionsPredicting Directed Links using Nondiagonal Matrix Decompositions
Predicting Directed Links using Nondiagonal Matrix Decompositions
 
Online Dating Recommender Systems: The Split-complex Number Approach
Online Dating Recommender Systems: The Split-complex Number ApproachOnline Dating Recommender Systems: The Split-complex Number Approach
Online Dating Recommender Systems: The Split-complex Number Approach
 
Fairness on the Web: Alternatives to the Power Law (WebSci 2012)
Fairness on the Web:  Alternatives to the Power Law (WebSci 2012)Fairness on the Web:  Alternatives to the Power Law (WebSci 2012)
Fairness on the Web: Alternatives to the Power Law (WebSci 2012)
 
Fairness on the Web: Alternatives to the Power Law
Fairness on the Web:  Alternatives to the Power LawFairness on the Web:  Alternatives to the Power Law
Fairness on the Web: Alternatives to the Power Law
 
On the Spectral Evolution of Large Networks (PhD Thesis by Jérôme Kunegis)
On the Spectral Evolution of Large Networks (PhD Thesis by Jérôme Kunegis)On the Spectral Evolution of Large Networks (PhD Thesis by Jérôme Kunegis)
On the Spectral Evolution of Large Networks (PhD Thesis by Jérôme Kunegis)
 
Searching Microblogs: Coping with Sparsity and Document Quality
Searching Microblogs: Coping with Sparsity and Document QualitySearching Microblogs: Coping with Sparsity and Document Quality
Searching Microblogs: Coping with Sparsity and Document Quality
 
Bad News Travel Fast: A Content-based Analysis of Interestingness on Twitter
Bad News Travel Fast: A Content-based Analysis of Interestingness on TwitterBad News Travel Fast: A Content-based Analysis of Interestingness on Twitter
Bad News Travel Fast: A Content-based Analysis of Interestingness on Twitter
 
On the Scalability of Graph Kernels Applied to Collaborative Recommenders
On the Scalability of Graph Kernels Applied to Collaborative RecommendersOn the Scalability of Graph Kernels Applied to Collaborative Recommenders
On the Scalability of Graph Kernels Applied to Collaborative Recommenders
 
The Slashdot Zoo: Mining a Social Network with Negative Edges
The Slashdot Zoo:  Mining a Social Network with Negative EdgesThe Slashdot Zoo:  Mining a Social Network with Negative Edges
The Slashdot Zoo: Mining a Social Network with Negative Edges
 
Spectral Analysis of Signed Graphs for Clustering, Prediction and Visualization
Spectral Analysis of Signed Graphs for Clustering, Prediction and VisualizationSpectral Analysis of Signed Graphs for Clustering, Prediction and Visualization
Spectral Analysis of Signed Graphs for Clustering, Prediction and Visualization
 
Network Growth and the Spectral Evolution Model
Network Growth and the Spectral Evolution ModelNetwork Growth and the Spectral Evolution Model
Network Growth and the Spectral Evolution Model
 

Recently uploaded

The French Revolution Class 9 Study Material pdf free download
The French Revolution Class 9 Study Material pdf free downloadThe French Revolution Class 9 Study Material pdf free download
The French Revolution Class 9 Study Material pdf free download
Vivekanand Anglo Vedic Academy
 
Welcome to TechSoup New Member Orientation and Q&A (May 2024).pdf
Welcome to TechSoup   New Member Orientation and Q&A (May 2024).pdfWelcome to TechSoup   New Member Orientation and Q&A (May 2024).pdf
Welcome to TechSoup New Member Orientation and Q&A (May 2024).pdf
TechSoup
 
The geography of Taylor Swift - some ideas
The geography of Taylor Swift - some ideasThe geography of Taylor Swift - some ideas
The geography of Taylor Swift - some ideas
GeoBlogs
 
Unit 8 - Information and Communication Technology (Paper I).pdf
Unit 8 - Information and Communication Technology (Paper I).pdfUnit 8 - Information and Communication Technology (Paper I).pdf
Unit 8 - Information and Communication Technology (Paper I).pdf
Thiyagu K
 
"Protectable subject matters, Protection in biotechnology, Protection of othe...
"Protectable subject matters, Protection in biotechnology, Protection of othe..."Protectable subject matters, Protection in biotechnology, Protection of othe...
"Protectable subject matters, Protection in biotechnology, Protection of othe...
SACHIN R KONDAGURI
 
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
siemaillard
 
2024.06.01 Introducing a competency framework for languag learning materials ...
2024.06.01 Introducing a competency framework for languag learning materials ...2024.06.01 Introducing a competency framework for languag learning materials ...
2024.06.01 Introducing a competency framework for languag learning materials ...
Sandy Millin
 
The approach at University of Liverpool.pptx
The approach at University of Liverpool.pptxThe approach at University of Liverpool.pptx
The approach at University of Liverpool.pptx
Jisc
 
Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46
Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46
Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46
MysoreMuleSoftMeetup
 
678020731-Sumas-y-Restas-Para-Colorear.pdf
678020731-Sumas-y-Restas-Para-Colorear.pdf678020731-Sumas-y-Restas-Para-Colorear.pdf
678020731-Sumas-y-Restas-Para-Colorear.pdf
CarlosHernanMontoyab2
 
Home assignment II on Spectroscopy 2024 Answers.pdf
Home assignment II on Spectroscopy 2024 Answers.pdfHome assignment II on Spectroscopy 2024 Answers.pdf
Home assignment II on Spectroscopy 2024 Answers.pdf
Tamralipta Mahavidyalaya
 
How to Make a Field invisible in Odoo 17
How to Make a Field invisible in Odoo 17How to Make a Field invisible in Odoo 17
How to Make a Field invisible in Odoo 17
Celine George
 
Introduction to AI for Nonprofits with Tapp Network
Introduction to AI for Nonprofits with Tapp NetworkIntroduction to AI for Nonprofits with Tapp Network
Introduction to AI for Nonprofits with Tapp Network
TechSoup
 
Guidance_and_Counselling.pdf B.Ed. 4th Semester
Guidance_and_Counselling.pdf B.Ed. 4th SemesterGuidance_and_Counselling.pdf B.Ed. 4th Semester
Guidance_and_Counselling.pdf B.Ed. 4th Semester
Atul Kumar Singh
 
The basics of sentences session 5pptx.pptx
The basics of sentences session 5pptx.pptxThe basics of sentences session 5pptx.pptx
The basics of sentences session 5pptx.pptx
heathfieldcps1
 
A Strategic Approach: GenAI in Education
A Strategic Approach: GenAI in EducationA Strategic Approach: GenAI in Education
A Strategic Approach: GenAI in Education
Peter Windle
 
Francesca Gottschalk - How can education support child empowerment.pptx
Francesca Gottschalk - How can education support child empowerment.pptxFrancesca Gottschalk - How can education support child empowerment.pptx
Francesca Gottschalk - How can education support child empowerment.pptx
EduSkills OECD
 
The Challenger.pdf DNHS Official Publication
The Challenger.pdf DNHS Official PublicationThe Challenger.pdf DNHS Official Publication
The Challenger.pdf DNHS Official Publication
Delapenabediema
 
CLASS 11 CBSE B.St Project AIDS TO TRADE - INSURANCE
CLASS 11 CBSE B.St Project AIDS TO TRADE - INSURANCECLASS 11 CBSE B.St Project AIDS TO TRADE - INSURANCE
CLASS 11 CBSE B.St Project AIDS TO TRADE - INSURANCE
BhavyaRajput3
 
The Accursed House by Émile Gaboriau.pptx
The Accursed House by Émile Gaboriau.pptxThe Accursed House by Émile Gaboriau.pptx
The Accursed House by Émile Gaboriau.pptx
DhatriParmar
 

Recently uploaded (20)

The French Revolution Class 9 Study Material pdf free download
The French Revolution Class 9 Study Material pdf free downloadThe French Revolution Class 9 Study Material pdf free download
The French Revolution Class 9 Study Material pdf free download
 
Welcome to TechSoup New Member Orientation and Q&A (May 2024).pdf
Welcome to TechSoup   New Member Orientation and Q&A (May 2024).pdfWelcome to TechSoup   New Member Orientation and Q&A (May 2024).pdf
Welcome to TechSoup New Member Orientation and Q&A (May 2024).pdf
 
The geography of Taylor Swift - some ideas
The geography of Taylor Swift - some ideasThe geography of Taylor Swift - some ideas
The geography of Taylor Swift - some ideas
 
Unit 8 - Information and Communication Technology (Paper I).pdf
Unit 8 - Information and Communication Technology (Paper I).pdfUnit 8 - Information and Communication Technology (Paper I).pdf
Unit 8 - Information and Communication Technology (Paper I).pdf
 
"Protectable subject matters, Protection in biotechnology, Protection of othe...
"Protectable subject matters, Protection in biotechnology, Protection of othe..."Protectable subject matters, Protection in biotechnology, Protection of othe...
"Protectable subject matters, Protection in biotechnology, Protection of othe...
 
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
 
2024.06.01 Introducing a competency framework for languag learning materials ...
2024.06.01 Introducing a competency framework for languag learning materials ...2024.06.01 Introducing a competency framework for languag learning materials ...
2024.06.01 Introducing a competency framework for languag learning materials ...
 
The approach at University of Liverpool.pptx
The approach at University of Liverpool.pptxThe approach at University of Liverpool.pptx
The approach at University of Liverpool.pptx
 
Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46
Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46
Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46
 
678020731-Sumas-y-Restas-Para-Colorear.pdf
678020731-Sumas-y-Restas-Para-Colorear.pdf678020731-Sumas-y-Restas-Para-Colorear.pdf
678020731-Sumas-y-Restas-Para-Colorear.pdf
 
Home assignment II on Spectroscopy 2024 Answers.pdf
Home assignment II on Spectroscopy 2024 Answers.pdfHome assignment II on Spectroscopy 2024 Answers.pdf
Home assignment II on Spectroscopy 2024 Answers.pdf
 
How to Make a Field invisible in Odoo 17
How to Make a Field invisible in Odoo 17How to Make a Field invisible in Odoo 17
How to Make a Field invisible in Odoo 17
 
Introduction to AI for Nonprofits with Tapp Network
Introduction to AI for Nonprofits with Tapp NetworkIntroduction to AI for Nonprofits with Tapp Network
Introduction to AI for Nonprofits with Tapp Network
 
Guidance_and_Counselling.pdf B.Ed. 4th Semester
Guidance_and_Counselling.pdf B.Ed. 4th SemesterGuidance_and_Counselling.pdf B.Ed. 4th Semester
Guidance_and_Counselling.pdf B.Ed. 4th Semester
 
The basics of sentences session 5pptx.pptx
The basics of sentences session 5pptx.pptxThe basics of sentences session 5pptx.pptx
The basics of sentences session 5pptx.pptx
 
A Strategic Approach: GenAI in Education
A Strategic Approach: GenAI in EducationA Strategic Approach: GenAI in Education
A Strategic Approach: GenAI in Education
 
Francesca Gottschalk - How can education support child empowerment.pptx
Francesca Gottschalk - How can education support child empowerment.pptxFrancesca Gottschalk - How can education support child empowerment.pptx
Francesca Gottschalk - How can education support child empowerment.pptx
 
The Challenger.pdf DNHS Official Publication
The Challenger.pdf DNHS Official PublicationThe Challenger.pdf DNHS Official Publication
The Challenger.pdf DNHS Official Publication
 
CLASS 11 CBSE B.St Project AIDS TO TRADE - INSURANCE
CLASS 11 CBSE B.St Project AIDS TO TRADE - INSURANCECLASS 11 CBSE B.St Project AIDS TO TRADE - INSURANCE
CLASS 11 CBSE B.St Project AIDS TO TRADE - INSURANCE
 
The Accursed House by Émile Gaboriau.pptx
The Accursed House by Émile Gaboriau.pptxThe Accursed House by Émile Gaboriau.pptx
The Accursed House by Émile Gaboriau.pptx
 

Eight Formalisms for Defining Graph Models

  • 1. Models of Graphs Jérôme Kunegis Oberseminar 2013-08-29
  • 2. Jérôme Kunegis Models of Graphs 2 Erdős–Rényi Each edge has probability p of existing P(G) = pm (1 − p)(M − m) m = #edges M = max possible #edges
  • 3. Jérôme Kunegis Models of Graphs 3 Barabási–Albert An edge appears with probability proportional to the degree of the node it connects P({u, v}) d(u)∼ d(u) = degree of node u
  • 4. Jérôme Kunegis Models of Graphs 4 What Everybody Thinks My network model leads to graphs that have the same properties as actual social networks Hmmm...
  • 5. Jérôme Kunegis Models of Graphs 5 P(G) = pm (1 − p)(M − m) P({u, v}) d(u)∼ Why don't you use the same formalism?? Comparison
  • 6. Jérôme Kunegis Models of Graphs 6 Formalisms for Graph Models (1) Specify a graph generation algorithm (2) Specify a graph growth algorithm (3) Specify the probability of any graph (4) Specify the probability of any edge (5) Specify the probability of any event (6) Specify a score for node pairs (7) Matrix model (8) Graph compression
  • 7. Jérôme Kunegis Models of Graphs 7 (1) Specify a Graph Generation Algorithm STEP 1: Specify rules for generating a graph Take a lattice, and rewire a certain proportion of edges randomly EXAMPLE: small-world model (Watts & Strogatz 1998) STEP 2: Generate random graph(s) STEP 3: Compare with actual networks Hey, a small diameter and large clustering coefficient! ● Not generative ● Not probabilistic
  • 8. Jérôme Kunegis Models of Graphs 8 (2) Specify a Graph Growth Algorithm An edge appears with probability proportional to the degree with probability p and at random with probability (1 − p) STEP 1: Specify exact growth rules STEP 2: Generate random graph(s) STEP 3: Compare with actual networks Look, a power law! EXAMPLE: preferential attachment (Barabási & Albert 1999) ● No overall probability
  • 9. Jérôme Kunegis Models of Graphs 9 What We Need: A Probabilistic Model A probabilistic model assigns a probability to each possible value. X: set of possible values x ∈ X: a value p: A parameter of the model P(x; p): Probability of x, given p, OR Likelihood of p, given x Σx∈X P(x; p) = 1 // Because P is a distribution for a given p Given a set of values {xi} for i = 1, … N, the best fitting p can be found by maximum likelihood: maxp Πi P(xi, p) So, are “values” whole graphs or individual edges?
  • 10. Jérôme Kunegis Models of Graphs 10 (3) Specify the Probability of Any Graph Each edge has probability p of existing STEP 1: Specify the probability of any graph G ● Not generative ● Needs multiple graphs for inference STEP 2: Given a set of graphs with the same number of nodes, compute the likelihood of any value p EXAMPLE: (Erdős & Rényi 1959)
  • 11. Jérôme Kunegis Models of Graphs 11 Example: Extension of Erdős–Rényi using Formalism (3) Goal: Add a parameter that controls the number of triangles. Idea: The E–R model with parameter p is an exponential family; the extension should be too. P(G) = (1 / C) pm (1 − p)(M − m) qt (1 − q)(T − t) where t is the #triangles, T is the maximum possible #triangles. Note: q = 1/2 gives the ordinary E–R model. Result: exponential random graph models (ERGM) and p* models The normalization constant C cannot be computed. It would be necessary to count the number of graphs with n vertices, m edges and t triangles. This is a hard, open problem. Gibbs sampling works, however. Open problem: Use Gibbs sampling to generate mini-models of networks.
  • 12. Jérôme Kunegis Models of Graphs 12 (4) Specify the Probability of Any Edge STEP 1: Specify probability for all pairs {u, v} EXAMPLE: Use a given degree vector d as parameter, and P({u, v}) = du dv EXAMPLE: The p1 model based on node attributes (Holland & Leinhard 1977) STEP 2: Compute likelihood of parameters ● Not generative Let's model each edge as an event, not a full graph ● Supports multiple edges
  • 13. Jérôme Kunegis Models of Graphs 13 Preliminary Results for Formalism (4) The best rank-1 model is given by the preferential attachment model. Let a graph G be given. Among all models of the form P({u, v}) = x xT , the one with maximum likelihood is given by P({u, v}) = d(u) d(v) / 2m Proof: By induction over n. Open problem: define other models using this formalism Hey, that's different from minimizing the least squares distance to the given adjacency matrix, where the SVD is best
  • 14. Jérôme Kunegis Models of Graphs 14 (5) Specify the Probability of Any Event Let's specify the probability of an edge addition, given the current graph STEP 1: Specify the probability of an edge addition given the current graph EXAMPLE: P({u, v}) = p / n² + (1 − p) d(u) d(v) / 2m STEP 2: Compute the likelihood OTHER EXAMPLE: (Akkermans & al. 2012) Open problem: Inference of parameters from real networks. Generalizes naturally to edge removal events.
  • 15. Jérôme Kunegis Models of Graphs 15 (6) Specify a Score for Node Pairs Read my paper STEP 1: Given a graph, specify a score for each node pairs STEP 2: Evaluate using information retrieval methods I know, that's link prediction! ● Not probabilistic (Liben-Nowell & Kleinberg 2003)
  • 16. Jérôme Kunegis Models of Graphs 16 (7) Matrix Model STEP 1: Specify a probability matrix STEP 2: Map nodes of the graph to rows/columns of the matrix STEP 3: Compute the likelihood Let's try the Kronecker product EXAMPLE: (Leskovec & al. 2005) ● Not generative Can I do this with any matrix?
  • 17. Jérôme Kunegis Models of Graphs 17 (8) Graph Compression STEP 1: Specify a graph compression algorithm STEP 2: Check how well it compresses a graph (Shannon) More probable values should have shorter representations I wonder how the E-R model can be used here ● Not generative
  • 18. Now let's do some research! SUMMARY (1) Graph generation (e.g., Watts–Strogatz) (2) Graph growth (e.g., Barabási–Albert) (3) Graph probability (e.g., Erdős–Rényi) (4) Edge probability (5) Event probability (6) Edge score (link prediction) (7) Matrix model (e.g., Leskovec & al.) (8) Graph compression Inference Mini-models Rank-2 model Spectral model supercededby Equivalence