Module 1 Introduction To Graph Theory and Probabilistic Graphical Models

The material in this presentation belongs to St. Francis Institute of Technology and is solely for educational purposes.
Distribution and modifications of the content is prohibited.
Probabilistic Graphical Models

CSDLO5011
2024-25
Subject Incharge
Dr. Bidisha Roy
Associate Professor
Room No. 401
email: bidisharoy@sfit.ac.in
St. Francis Institute of Technology PGM

Department of Computer Engineering Dr. Bidisha Roy 1
The material in this presentation belongs to St. Francis Institute of Technology and is solely for educational purposes. Distribution and modifications of the content is prohibited.
Module 1: Part II
Introduction to Graph Theory and PGMs

Topics to be covered
❑ Introduction to Graph Theory
❑Introduction to Probabilistic Graph Models
❑Applications of PGM

Topics to be covered
❑ Introduction to Graph Theory
❑Introduction to Probabilistic Graph Models
❑Applications of PGM

Graph Theory
❑Study of graphs
❑Mathematical structures used to model pairwise
relations between objects
❑Many problems in computer science, physics, chemistry,
biology, genetics, social sciences, sociology, linguistics and
other fields can be represented as graph problems
❑Group theory, matrix theory, probability, topology have
close connection with graph theory

Graphs
❑G=(V, E), a non-empty set V of vertices (or nodes, points)
and a set E of edges (or links, arcs)
❑Vertex/Node
❑Basic element of a graph
❑Drawn as a node/dot
❑Vertex set of G usually denoted as V(G) or V or VG
❑Edge/Links/Arcs
❑A set of two elements
❑Drawn as a line connecting two vertices, called end vertices or
end points
❑Usually denoted by E(G) or E or EG
Graphs
❑Neighborhood
❑The set of nodes its connected to via an edge, represented as N(v)

Types of Graphs
❑Dense
❑Sparse
❑Directed/Directional
❑Undirected (unordered)
❑Weighted

Terminologies
❑Degree of a node:- In an Undirected Graph, it is the number
of edges connected to a node.
deg(A) = 2, deg(B) = 2, deg(C) = 2, deg(D) = 3,

deg(E) = 2, deg (F) = 2, deg(G) = 3
Practice Problems

Graph Terminologies
❑ Degree of a node:- For Directed Graph,
❑ The in-degree of a vertex v is the number of edges incident to
that node i.e. Edges coming to that node.
❑The out-degree of a vertex v is the number of edges incident
from that node i.e. Edges going out from that node.
Node Indegree Outdegree
1 0 3
2 1 2
3 2 1
4 3 3
5 1 2
6 3 0
7 2 1
Graph Terminologies
❑Subgraph:- A graph G’ whose vertex sets and edge sets are subsets
of Graph G
❑Every graph is a subgraph of itself

Graph Terminologies
❑ Path:- A sequence of edges connecting a sequence of distinct
vertices or no repeated vertices.
LENGTH OF THE PATH:- Total edges included in the path.
1–5–2–4–3–6
4–3–2–5–4–6
❑ Trail:- No repeated edges or a path where all edges are distinct
❑ Cycle:- Trail, except that the last vertex is the same as the first vertex
❑ Loop:- An edge that connects a vertex to itself
Real World Applications

❑Social Networks: Modeling relationships and interactions
Centrality, connectivity, clustering coefficient
❑Transportation Networks: Optimizing routes and connections
Route optimization, traffic flow analysis, logistics planning
❑Biological Networks: Understanding protein interactions and
gene regulation
Understanding diseases, drug discovery, systems biology
❑Computer Networks: Designing and analyzing communication
protocols
Network design, protocol analysis, cybersecurity
Graph Terminologies
❑Subgraph:- A graph G’ whose vertex sets and edge sets are subsets
of Graph G
❑Every graph is a subgraph of itself

PGMs
❑Statistical models that encode complex joint multivariate probability
distributions using graphs
❑Help study/observe the conditional independence relations between

random variables
❑Useful in solving tasks such as inference, learning etc.
❑Applications: Used in various fields such as artificial intelligence,

machine learning, bioinformatics, and more
Directed Graphical Models

❑Graph with vertices representing RVs and edges representing
conditional variables
❑ The second graph is a DAG which could represent a Bayesian

Network
❑The direction of edges represents the influence of one RV on
another
Bayesian Networks
❑Vertices/Nodes correspond to RVs
❑ Directed edges represent causal relationships or conditional

probabilities between RVs
❑No directed edge indicates RVs being independent of each other

Bayesian Networks

Markov Model
❑Also known as Markov Chain
❑Used to describe a system that evolves over a set of discrete states
❑Markov Property: A future state of the system depends upon the

current state, not on the sequence of states/events that preceded it.
❑Memory less property, “What happens next depends only on the
state of affairs now”
❑A Markov model consists of:
❑States, State Transition Probabilities and Initial State Distribution
Markov Model

Markov Random Fields (MRFs)

❑A Markov random field is a model used for undirected graphical
models where the joint distribution of a set of random variables has
a Markov property described by an undirected graph
❑MRFs do not have inherent directionality; instead, they represent

mutual dependencies without specifying transition directions
❑Often used in spatial data analysis, image processing, and other

areas where mutual dependency is more important than transition
direction
Hidden Markov Model (HMM)

❑A statistical model where the system being modeled is assumed to
be a Markov process with hidden states
❑HMM has following states:
❑ Hidden states, Observations(states), State transition probabilities
(hidden), observation probabilities, initial state distribution
(hidden)
❑The observed data provides indirect information about the
underlying hidden states
❑Applications: Speech recognition (PoS tagging), bioinformatics,
finance.
Hidden Markov Model (HMM)

Summary
❑Graph Theory
❑PGMs
❑Bayesian Networks
❑Markov Models
❑Hidden Markov Models


Module 1 Introduction To Graph Theory and Probabilistic Graphical Models

Uploaded by

Module 1 Introduction To Graph Theory and Probabilistic Graphical Models

Uploaded by

The material in this presentation belongs to St. Francis Institute of Technology and is solely for educational purposes.

Distribution and modifications of the content is prohibited.

Probabilistic Graphical Models

St. Francis Institute of Technology PGM

St. Francis Institute of Technology PGM

❑ Introduction to Graph Theory

❑Introduction to Probabilistic Graph Models

St. Francis Institute of Technology PGM

❑ Introduction to Graph Theory

❑Introduction to Probabilistic Graph Models

St. Francis Institute of Technology PGM

St. Francis Institute of Technology PGM

St. Francis Institute of Technology PGM

St. Francis Institute of Technology PGM

deg(A) = 2, deg(B) = 2, deg(C) = 2, deg(D) = 3,

St. Francis Institute of Technology PGM

St. Francis Institute of Technology PGM

Real World Applications

St. Francis Institute of Technology PGM

❑Help study/observe the conditional independence relations between

❑Useful in solving tasks such as inference, learning etc.

❑Applications: Used in various fields such as artificial intelligence,

Directed Graphical Models

❑ The second graph is a DAG which could represent a Bayesian

❑ Directed edges represent causal relationships or conditional

❑No directed edge indicates RVs being independent of each other

St. Francis Institute of Technology PGM

St. Francis Institute of Technology PGM

❑Used to describe a system that evolves over a set of discrete states

❑Markov Property: A future state of the system depends upon the

St. Francis Institute of Technology PGM

Markov Random Fields (MRFs)

❑MRFs do not have inherent directionality; instead, they represent

❑Often used in spatial data analysis, image processing, and other

Hidden Markov Model (HMM)

Hidden Markov Model (HMM)

St. Francis Institute of Technology PGM

❑Hidden Markov Models

St. Francis Institute of Technology PGM

You might also like