Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

17
Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi

Transcript of Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Page 1: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Belief Networks

Done by:Amna Omar Abdullah

Fatima Mahmood Al-HajjatNajwa Ahmad Al-Zubi

Page 2: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Definition of Belief Networks:

A belief network: is a directed acyclic graph (DAG) which uses a graphical representation to represent probabilistic structure. It consists of vertices (or nodes) that are connected by causal links - represented by arrows - which point from parent nodes (causes) to child nodes (effects).

Page 3: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Each node is used to represent a random variable, and each directed edge represents an immediate dependence or direct influence. How to design a Belief Network?

• Explore the causal relations

Page 4: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Example

Holmes and Watson have moved to IA. He wakes up to find his lawn wet. He wonders if it has rained or if he left his sprinkler on. He looks at his neighbor Watson’s lawn and he sees it is wet too. So, he concludes it must have rained.

Rain Sprinklers

Holme’s lawn wet

Watson’s lawn wet

Defining casual components

Page 5: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Cont’d

Holmes and Watson have moved to IA. He wakes up to find his lawn wet. He wonders if it has rained or if he left his sprinkler on. He looks at his neighbor Watson’s lawn and he sees it is wet too. So, he concludes it must have rained.

Rain Sprinklers

Holme’s lawn wet

Watson’s lawn wet

Define the probabilities

Page 6: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Cont’d

Holmes and Watson have moved to IA. He wakes up to find his lawn wet. He wonders if it has rained or if he left his sprinkler on. He looks at his neighbor Watson’s lawn and he sees it is wet too. So, he concludes it must have rained.

Rain Sprinklers

Holme’s lawn wet

Watson’s lawn wet

Given w, P(R) goes up and P(S) goes down

Page 7: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Joint Probability Distribution:

Belief network is related to the joint probability distribution that:

1- Representing information in a more understandable and logical manner.

2- making construction and interpretation much simpler, It does not fully specify it. What it does is give the conditional dependencies in the probability distribution.

On top of these direct dependence relationships, we need to give some numbers (or functions) which define how the variables relate to one another.

Page 8: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Joint distribution:• P(X1, X2, …, Xn)• Probability assignment to all combinations

of values of random variables and provide complete information about the probabilities of its random variables.

• A JPD table for n random variables, each ranging over k distinct values, has kⁿ entries!

Toothache :Toothache

Cavity .04 .06:Cavity .01 .89

Page 9: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Using the joint distribution

A joint probability distribution P(A,B,C,D,E) over a number of variables can be written in the following form using the chain rule of probability:

P(A,B,C,D,E) = P(A) P(B|A) P(C|B,A) P(D|C,B,A) P(E|D,C,B,A).

Page 10: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Probability theory

• Conditioning

• P(A) = P(A | B) P(B) + P(A | :B) P(:B)

• A and B are independent iff

• P(A | B) = P(A)

• P(B | A) = P(B)

• A and B are conditionally independent given C iff

• P(A | B, C) = P(A | C)

• P(B | A, C) = P(B | C)

• Bayes’ Rule

• P(A | B) = P(B | A) P(A) / P(B)

• P(A | B, C) = P(B | A, C) P(A | C) / P(B | C)

Page 11: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Cont’d

Variables: V1, …, Vn• Values: v1, …, vn• P(V1=v1, V2=v2, …, Vn=vn) = Πi

P(Vi=vi | parents(Vi))

P(ABCD) = P(A=true, B=true,C=true, D=true)

P(ABCD) =P(D|ABC)P(ABC)

A B

C

D

P(A) P(B)

P(C|A,B)

P(D|C)

Page 12: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

P(ABCD) = P(D|ABC)P(ABC) = P(D|C) P(ABC) =

A independent from D given CB independent from D given C

A B

C

D

P(A)

P(C|A,B)

P(D|C)

P(B)

Page 13: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

P(ABCD) = P(D|ABC)P(ABC) = P(D|C) P(ABC) = P(D|C) P(C|AB) P(AB) =

A independent from D given CB independent from D given C

A B

C

D

P(A)

P(C|A,B)

P(D|C)

P(B)

Page 14: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Cont’d

P(ABCD) =P(D|ABC)P(ABC) =P(D|C) P(ABC) =P(D|C) P(C|AB) P(AB) =P(D|C) P(C|AB) P(A)P(B)

A independent from B

A B

C

D

P(A)

P(C|A,B)

P(D|C)

P(B)

Page 15: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Probability that Watson Crashes

Icy

Holmes Crash

Watson Crash

P(I)=0.7

P(W) = P(W| I) P(I) + P(W|-I) P(-I)

= 0.8*0.7 + 0.1*0.3

= 0.56 + 0.03

= 0.59

P(H|I)

I 0.8

-I 0.1

P(W|I)

I 0.8

-I 0.1

Page 16: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Cont’d

Icy

Holmes Crash

Watson Crash

P(I)=0.7

P(H|I)

I 0.8

-I 0.1

P(W|I)

I 0.8

-I 0.1

P(I | W) = P(W | I) P(I) / P(W)

= 0.8*0.7 / 0.59

= 0.95

We started with P(I) = 0.7; knowing that Watson crashed raised the probability to 0.95

Page 17: Belief Networks Done by: Amna Omar Abdullah Fatima Mahmood Al-Hajjat Najwa Ahmad Al-Zubi.

Thank you for listening, if

you have any questions

don’t hesitate to ask Mr.

Ashraf.