0% found this document useful (0 votes)

78 views58 pages

Probabilistic Reasoning

This document provides an overview of probabilistic reasoning and Bayesian networks. It discusses the syntax and semantics of Bayesian networks, including how they can be used to efficiently represent conditional distributions through the use of conditional probability tables. It also describes various techniques for performing exact and approximate probabilistic inference on Bayesian networks, including enumeration, variable elimination, stochastic simulation, and Markov chain Monte Carlo methods. Finally, it provides examples of how Bayesian networks can be used to model problems involving diagnosis, alarms, and other scenarios involving probabilistic relationships between variables.

Uploaded by

Nguyễn Thị Mỹ Hân

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

78 views58 pages

Probabilistic Reasoning

Uploaded by

Nguyễn Thị Mỹ Hân

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 58

Probabilistic Reasoning

LESSON 12
Reading
Chapter 14

3/21/2016 503043 - PROBABILISTIC REASONING 2

Outline

 Syntax of Bayesian networks

 Semantics of Bayesian networks

 Efficient representation of conditional distributions

 Exact inference by enumeration

 Exact inference by variable elimination

 Approximate inference by stochastic simulation*

 Approximate inference by Markov Chain Monte Carlo*

3/21/2016
3
503043 - PROBABILISTIC REASONING
Motivations

 Full joint probability distribution can answer any

question but can become intractably large as
number of variable increases
 Specifying probabilities for atomic events can be
difficult, e.g., large set of data, statistical estimates,
etc.
 Independence and conditional independence reduce
the probabilities needed for full joint probability
distribution.
3/21/2016 503043 - PROBABILISTIC REASONING 4
Bayesian networks
 A directed, acyclic graph (DAG)
 A set of nodes, one per variable (discrete or continuous)
 A set of directed links (arrows) connects pairs of nodes. X is a parent
of Y if there is an arrow (direct influence) from node X to node Y.
 Each node X i has a conditional probability distribution
that quantifies the effect of the parents on the node.
 Combinations of the topology and the conditional distributions
specify (implicitly) the full joint distribution for all the variables.

P X i | Parents ( X i ) 

3/21/2016
5
503043 - PROBABILISTIC REASONING
Bayesian networks
Example 1： The Teeth Disease Bayesian

3/21/2016
6
503043 - PROBABILISTIC REASONING
Example: Burglar alarm system

 I have a burglar alarm installed at home

 It is fairly reliable at detecting a burglary, but also responds on
occasion to minor earth quakes.
 I also have two neighbors, John and Mary
 They have promised to call me at work when they hear the alarm

 John always calls when he hears the alarm, but sometimes confuses
the telephone ringing with the alarm and calls then, too.
 Mary likes rather loud music and sometimes misses the alarm
altogether.

 Bayesian networks variables:

 Burglar, Earthquake, Alarm, JohnCalls, MaryCalls

3/21/2016 7
Example: Burglar alarm system
 Network topology reflects “causal” knowledge:
 A burglar can set the alarm off
 An earthquake can set the alarm off
 The alarm can cause Mary to call
 The alarm can cause John to call

conditional probability table (CPT):

each row contains the conditional
probability of each node value for a
conditioning case (a possible combination
of values for the parent nodes).

3/21/2016 8
Compactness of Bayesian networks

A CPT for Boolean Xi with k Boolean parents has 2^k rows for the
combinations of parent values. Each row requires one number p for
Xi =true (the number for Xi =false is just 1-p)

If each variable has no more than k parents, The complete network

requires O(n. 2^k) numbers i.e., grows linearly with n, vs. O(2^k) for
the full joint distribution

For Burglary net,

1+1+4+2+2=10 numbers
(vs. 2^5-1=31)

3/21/2016 503043 - PROBABILISTIC REASONING 9

Global semantics of Bayesian networks

Global semantics defines the full

joint distribution as the product of
the local conditional distributions

, X n )   i 1 p( X i | Parents( X i ))
n
p( X 1 ,
e. g.
p ( j  m  a  b  e)
 p( j | a ) p(m | a) p( a | b, e) p(b) p(e)
 0.90  0.70  0.001  0.999  0.998
3/21/2016
 0.00062
Local semantics of Bayesian network

e.g., JohnCalls is independent of Burglary and Earthquake,

given the value of Alarm.

3/21/2016 11
Markov blanket

e.g., Burglary is independent of JohnCalls and MaryCalls ,

given the value of Alarm and Earthquake.

3/21/2016
Constructing Bayesian networks
Need a method such that a series of locally testable assertions of
conditional independence guarantees the required global semantics.

The correct order in which to add nodes is to add the “root causes”
first, then the variables they influence, and so on.
What happens if we choose the wrong order?

3/21/2016 13
Example

3/21/2016 14
Example

3/21/2016 15
Example

3/21/2016 16
Example

3/21/2016 17
Example

3/21/2016 18
Example

 Deciding conditional independence is hard in noncausal directions.

 Assessing conditional probabilities is hard in noncausal directions
 Network is less compact: 1 + 2 + 4 + 2 + 4 = 13 numbers needed

3/21/2016 19
Example: Car diagnosis
 Initial evidence: car won’t start
 Testable variables (green), “broken, so fix it” variables (orange)
 Hidden variables (gray) ensure sparse structure, reduce parameters

3/21/2016 20
Example: Car insurance

3/21/2016 21
Class Exercise
1. Using the axioms of probability, prove
2. Given Bayesian Network
p( P | P  Q )  1
◦ Calculate
◦ In what condition,
P p(P)

p( P  Q ) Q p(Q|P), p(Q| ﹁ P)

p( P  Q )  p(Q P )

3/21/2016
Efficient representation of conditional
distributions

 CPT grows exponentially with number of parents O(2k)

 CPT becomes infinite with continuous-valued parent or child

 Solution: canonical distribution that can be specified by a few parameters

 Simplest example: deterministic node whose value specified exactly by the values of its
parents, with no uncertainty

3/21/2016 503043 - PROBABILISTIC REASONING 23

Efficient representation of conditional
distributions

 “Noisy” logic relationships: uncertain relationships

 noisy-OR model allows for uncertainty about the ability of each parent to cause
the child to be true, but the causal relationship between parent and child maybe
inhibited.
 E.g.: Fever is caused by Cold, Flu, or Malaria, but a patient could have a
cold, but not exhibit a fever.
 Two assumptions of noisy-OR
 parents include all the possible causes (can add leak node that covers
“miscellaneous causes.”)
 inhibition of each parent is independent of inhibition of any other parents,
e.g., whatever inhibits Malaria from causing a fever is independent of whatever
inhibits Flu from causing a fever

Pfever | cold , flu, malaria  0.6

Pfever | cold , flu, malaria  0.2 Cold Flu Malaria

Pfever | cold , flu, malaria  0.1

Cold

3/21/2016
24
503043 - PROBABILISTIC REASONING
Efficient representation of conditional distributions

 Other probabilities can be calculated from the product of the

inhibition probabilities for each parent

 Number of parameters linear in number of parents

O(2k)O(k)

3/21/2016
25
503043 - PROBABILISTIC REASONING
Bayesian nets with continuous variables
 Hybrid Bayesian network: discrete variables + continuous variables
 Discrete (Subsidy? and Buys?); continuous (Harvest and Cost)

 Two options
 discretization – possibly large errors, large CPTs
 finitely parameterized canonical families(E.g.: Gaussian Distribution )

 Two kinds of conditional distributions for hybrid Bayesian

network
 continuous variable given discrete or continuous parents (e.g., Cost)
 discrete variable given continuous parents (e.g., Buys?)
3/21/2016
26
503043 - PROBABILISTIC REASONING
Continuous child variables
 Need one conditional density function for child variable given
continuous parents, for each possible assignment to discrete parents
 Most common is the linear Gaussian model, e.g.,:

 Mean Cost varies linearly with Harvest, variance is fixed

 the linear model is reasonable only if the harvest size is limited to a
narrow range

3/21/2016
27
503043 - PROBABILISTIC REASONING
Continuous child variables

 Discrete + continuous linear Gaussian network is a conditional

Gaussian network, i.e., a multivariate Gaussian distribution over
all continuous variables for each combination of discrete variable
values.

 A multivariate Gaussian distribution is a surface in more than one

dimension that has a peak at the mean and drops off on all sides

3/21/2016
28
503043 - PROBABILISTIC REASONING
Discrete variable with continuous parents

 Probability of Buys?
 given Cost should

be a “soft” threshold”

 Probit distribution uses

integral of Gaussian:

3/21/2016 503043 - PROBABILISTIC REASONING 29

Discrete variable with continuous parents

 Sigmoid (or logit) distribution also used in neural networks:

 Sigmoid has similar shape to probit but much longer tails

3/21/2016
30
503043 - PROBABILISTIC REASONING
Exact inference by enumeration
 A query can be answered using a Bayesian network by computing
sums of products of conditional probabilities from the network.

sum over hidden variables: earthquake and alarm

d = 2 when we have n Boolean variables

3/21/2016 503043 - PROBABILISTIC REASONING 31

Evaluation tree

3/21/2016 503043 - PROBABILISTIC REASONING 32

Exact inference by variable elimination

 Do the calculation once and save the results for later use – idea of
dynamic programming
 Variable elimination: carry out summations right-to-left, storing
intermediate results (factors) to avoid re-computation

3/21/2016 503043 - PROBABILISTIC REASONING 33

Variable elimination – basic operations

3/21/2016 503043 - PROBABILISTIC REASONING 34

Variable elimination – irrelevant variables

The complexity of variable elimination

 Single connected networks (or polytrees)
 any two nodes are connected by at most one (undirected) path
 time and space cost of variable elimination are O(d k n) , d = 2 (n Boolean variables)
i.e., linear in the number of variables (nodes) if the number of parents of each node is
bounded by a constant
 Multiply connected network
 variable elimination can have exponential time and space complexity even the number
of parents per node is bounded.

3/21/2016 503043 - PROBABILISTIC REASONING 35

Approximate inference by stochastic
simulation*

 Direct sampling
 Generate events from (an empty) network that has no associated evidence
 Rejection sampling: reject samples disagreeing with evidence
 Likelihood weighting: use evidence to weight samples