Professional Documents
Culture Documents
Uncertainty
Bayesian Theory and Fuzzy Sets
Introduction
With knowledge representation, we might write A→B, which means if A is true then B is true, but
consider a situation where we are not sure about whether A is true or not then we cannot
express this statement, this situation is called uncertainty
Uncertainty is the reported value that lies within the range of values within which the true
value is asserted to lie in.
In AI, there are numerous sources of uncertainty, including variation in specific data values and
the sample of data collected from the domain.
There are different methods to manage uncertainty in AI and expert system: non-monotonic
logic, probability reasoning and fuzzy logic, Bayesian theory
Contd…..
Agents may need to handle uncertainty, whether due to partial observability, no determinism or a
combination of both
Agent may not know for certain what state it is in or where it will end up after a sequence of actions
4 types of uncertainty in decision making process as Data, Action, Prediction and Judgment
A purely logical approach has two main problems as RISK or Wastage of resources
eg. if you want to reach at airport 2:00 pm and distance of airport from resident is approximate 25 km.
How much buffer time will you keep to reach at airport without delay or earlier?
Tool of Degree of Belief = Probability Theory
Ontological Commitments of logical and probability theory are same that world is composed of facts that
do or do not hold in any particular case
Epistemological Commitments are different that logical agent believes each sentence to be true, false or
not, whereas probabilistic agent may have a numerical degree of belief between 0 and 1.
Causes of Uncertainty
Laziness
Theoretical Ignorance
Practical Ignorance
Information occurred from unreliable sources
Experimental Errors
Climate Change
Equipment fault
How to handle Uncertainty
1. Implicit Methods
Ignore uncertainty as much as possible
Build procedures that are robust to uncertainty
This is a approach used in the planning methods
2. Explicit Methods
Build a model of the world that describes the uncertainty ( sensors, model, states, dynamics)
Reason about the effect of actions given the model
Probabilistic Reasoning – something is true or false
Fuzzy Logic – handles degree of TRUTH, not uncertainty
Default or non monotonic logic – make assumptions unless contradicted by evidence.
Rule based approaches
Fuzzy Logic
fuzzy logic was first used with 1965 by Lotfi Zadeh a professor of UC Berkeley in California
It was designed to allow the computer to determine the distinctions among data which is neither true nor
false
Represent uncertainty with degree
Represent the belongingness of member of crisp set (0,1) to fuzzy set
It is a method of reasoning that resembles the human reasoning
The approach of FL initiates the way of decision making in human that involves all intermediate possibilities
between digital values {0,1}
It is an alternative more institutive approach to reasoning with imprecise knowledge.
It does not claim to have the mathematical precision of probability theory, but rather to capture common
sense notions.
Contd….
• Support: { (x2, 0.2), (x3, 0.5), (x4, 1), (x5, 1), (x6, 1),
(x7, 0.5), (x8, 0.2) }
• Core: { (x4, 1), (x5, 1), (x6, 1) }
• Crossover Points: { (x3, 0.5), (x7, 0.5) }
• Alpha Cut0.2: { (x2, 0.2), (x3, 0.5), (x4, 1), (x5, 1), (x6,
1), (x7, 0.5), (x8, 0.2) }
• Strong Alpha Cut0.2+: { (x3, 0.5), (x4, 1), (x5, 1), (x6, 1),
(x7, 0.5) }
• Boundary: { (x2, 0.2), (x3, 0.5), (x7, 0.5), (x8, 0.2) }
• Bandwidth : x7 – x3
• Normality = True
Fuzzy Set Operations
9. De Morgan's Law:
Defuzzification Methods
Where, M ={ xi | μC(xi) = h(C) }, Or M is the set of points having highest membership value
Contd….
Find defuzzified value for given aggregated fuzzy output set using
weighted average method from given figure
• Find crisp value corresponding to following fuzzy output sets using various defuzzification methods.
• The fundamental task of any FIS is to apply the if-then rules on fuzzy input and produce the
corresponding fuzzy output.
• The whole process is based on the computer paradigm including fuzzy set theory, if-then rules and the
fuzzy reasoning process.
• Fuzzy inference (reasoning) is the actual process of mapping from a given input to an output using fuzzy
logic
• FIS have been successfully applied in fields such as automatic control, data classification, decision
analysis, expert systems , controller designing and many more
Fuzzy Inference Method (FIM)
• Fuzzy expert systems are different than the conventional expert systems which relies on the symbolic
logic. Where as, fuzzy expert system is inclined towards numeric computing.
• Fuzzy expert system perform reasoning from data using membership functions and fuzzy rules
• Consider the design of a fuzzy controller for a steam turbine. Assume the input of the fuzzy controller as
temperature and pressure. The output will be the throttle setting of a steam turbine. Use 3 descriptors
for input and 5 descriptors for output variables. Derive the set of rules for controller action and get the
defuzzified values.
• Assume that the current temperature is 30% and pressure is 40% and we have to determine the throttle
position of the turbine for this particular condition
Solution
Step 1: Identification of variables
• Inputs: Temperature and pressure
• Output: Throttle setting of steam turbine
Contd….
Step 5: Fuzzification
Contd….
• Conditional probability is a probability of occurring an event when another event has already
happened.
• The probability of A under the conditions of B
Example :
In a class, there are 70% of the students who like English and 40% of the students who likes English and
Mathematics. What is the percent of students who like English also like mathematics?
• A is an event that a student likes Mathematics
• B is an event that a student likes English.
Hence, 57% are the students who like English also like Mathematics.
Bayes Theorem
• In probability theory, it relates the conditional probability and marginal probabilities of two random
events.
• Bayes theorem can be derived using product rule and conditional probability of event A with known
event B
• As from product rule we can write
• Suppose 15 men out of 300 men and 25 women out of 1000 are good orators. An orator is chosen at
random. Find the probability that a male person is selected. Assume that there are equal numbers of
men and women?
Let there be 1000 men and 1000 women.
Let E1 and E2 be the events of choosing a man and a woman respectively.
P(E1) = 1000/2000 = 1/2 , and P(E2) = 1000/2000 = ½
Probability of selecting a male person ,given that the person selected is a good orator
P(E1|E) = P(E|E1) * P(E1)/ P(E|E1) * P(E1) + P(E|E2) * P(E2)
= (1/2 * 1/20) /{(1/2 * 1/20) + (1/2 * 1/40)}
= 2/3
Hence the required probability is 2/3.
Example 2
• A man is known to speak to throw die 1 out of 4 times. He throws a die and reports that it is a six. Find
the probability that is actually a six?
E1 = event of getting a six,
E2 = event of not getting a six and
E = event that the man reports that it is a six.
P(E1) = 1/6, and P(E2) = (1 – 1/6) = 5/6
P(E|E1) = probability that the man speaks the truth
= 3/4
P(E|E2) = probability that the man does not speak the truth
= (1 – 3/4) = 1/4
Probability of getting a six ,given that the man reports it to be six
P(E1|E) = P(E|E1) * P(E1)/P(E|E1) * P(E1) + P(E|E2) * P(E2)
= (3/4* 1/6)/ (3/4*1/6)+ (1/4 * 5/6)
= 3/8
Bayesian Belief Network
• A Bayesian network is a probabilistic graphical model which represents a set of variables and
their conditional dependencies using a directed acyclic graph.
• It is also called a Bayes network, belief network, decision network, or Bayesian model.
• It can also be used in various tasks including optimized search engine, diagnosis of different
diseases, filtering spam emails, prediction and decision making under uncertainty.
• Bayesian Network can be used for building models from data and experts opinions and it
consists of two parts:
1. Directed Acyclic Graph
2. Table of conditional probabilities
Contd….
• Harry installed a new burglar alarm at his home to detect burglary. The alarm reliably
responds at detecting a burglary but also responds for minor earthquakes. Harry has two
neighbors David and Sophia, who have taken a responsibility to inform Harry at work when
they hear the alarm. David always calls Harry when he hears the alarm, but sometimes he got
confused with the phone ringing and calls at that time too. On the other hand, Sophia likes to
listen to high music, so sometimes she misses to hear the alarm. Here we would like to
compute the probability of Burglary Alarm.
Problem:
• Calculate the probability that alarm has sounded, but there is neither a burglary nor an
earthquake occurred and David and Sophia both called the Harry.
Contd….
Example
Contd….