UNIT III UNCERTAINTY

Non monotonic reasoning-Fuzzy Logic-Fuzzy rules-fuzzy inference-Temporal Logic-Temporal

Reasoning-Neural Networks-Neuro-fuzzy Inference.

NON MONOTONIC REASONING

Reasoning is the act of deriving a conclusion from certain premises using a given methodology.

Uncertainty in reasoning
The world is an uncertain place; often the knowledge is imperfect which causes uncertainty.
Thus reasoning must be able to operate under uncertainty. AI systems must have ability to reason under
conditions of uncertainty. Uncertainties arise when there is
 Incompleteness in knowledge
 Inconsistencies in knowledge
 Changing knowledge

Monotonic Reasoning
In monotonic reasoning all conclusions are still valid after adding more information to the existing
information.
Example
All humans are mortal
Socrates is a human
Conclusion: Socrates is mortal
In monotonic reasoning, if we enlarge a set of axioms we cannot retract any existing assertions.

Non- Monotonic Reasoning

A logic is non monotonic if the truth of the proposition may change when new information is
added.
Example
Birds typically fly.
Tweety is a bird.
Conclusion: Tweety flies.
If Tweety is a Penguin, it is incorrect to conclude that Tweety flies. Conclusion of non-
monotonic argument may not be correct. Non-monotonic system deals with incomplete and uncertain
models. It is used to reason when a complete, consistent model of the world is not available.

The types of non-monotonic reasoning

i. Default reasoning
ii. Circumscription
iii. Truth Maintenance System

i. Default reasoning
This is very common form of non-monotonic reasoning. The conclusions are drawn based
on what is most likely to be true. There are two approaches of default reasoning.
 Non-monotonic logic
 Default logic

Non-monotonic logic
Non-monotonic logic s a predicate logic with one extension called model operator M which means
“consistent with everything we know”. The purpose of M is to allow consistency. A way to define
consistency with prolog notation is
To show that fact p is true, we attempt to prove ┐p.
If we fail, we may say that p is consistent since ┐p is false.
Example
∀x:plays_instrument(x)  manage (x)jazz_musician(x)
 States that for all x, that x plays an instrument and if the fact that x can manage is consistent
with all other manage is consistent with all other knowledge then we can conclude that x is a
jazz_musician.

Default Logic
𝐴:𝐵
Default logic initiates new inference rule: 𝐶
where, A is kown as pre-requisite
B is the justification
C is the consequent
“If A, and if it is consistent with the rest of what is known to assume that B, then conclude that C”
The rule says that given the pre-requisite, the consequent inferred provided it is consistent with
the rest of the data.
Example
Rule that “Birds typically fly” would be represented as
𝑏𝑖𝑟𝑑(𝑥): 𝑓𝑙𝑖𝑒𝑠(𝑥)
𝑓𝑙𝑖𝑒𝑠(𝑥)
If x is a bird and the claim that x flies is consistent with what we know, then infer that x flies.
The idea behind non-monotonic reasoning is to reason with first order logic, and if an inference
can not be obtained then use the set of default rules available within the first order formulation.
Applying default rules
By applying default rules, it is necessary to check their justification for consistency, not only
with initial data, but also with the consequent of any other default rules that may be applied. The
application of one rule may thus block the application of another. To solve this problem, a concept of
default theory was introduced.
Default theory
It consist of set of premises W and a set of default rules D. An extension for the default theory
is a set of sentences E which can be derived from W by applying as many rules of D as possible without
generating inconsistency. D is the set of default rules has a unique syntax of the form
∝ (𝑥⃗): 𝐸𝛽(𝑥⃗)
ƞ(𝑥⃗)
∝ (𝑥⃗) is the pre-requisite of the default rule
𝐸𝛽(𝑥⃗) is the consistency test of the default rule
ƞ(𝑥⃗) is the consequent of the default rule
This rule can be read as, for all individual x1,x2,…,xn , if ∝ (𝑥⃗) is believed and if each of 𝐸𝛽(𝑥⃗)
is consistent with our beliefs, then ƞ(𝑥⃗) may be believed.
Example
A default rule says “Typically an American adult owns a car”
𝐴𝑚𝑒𝑟𝑖𝑐𝑎𝑛𝑠(𝑥)Adult(x): M((∃y). car(y)owns(x, y))
(((∃y). car(y)owns(x, y))
The rule is only accessed if we wish to know whether or not John owns a car then an answer
can be deduced from our current beliefs. This default rule is applicable if we can prove from
our beliefs that John is an American and an adult, and believing that there is some car that is
owned by John does not lead to an inconsistency. If these two sets of premises are satisfied,
then the rule states that we can conclude that John owns a car.
ii) Circumscription
Circumscription is a non-monotonic logic to formalize the common sense assumption. Circumscription
is a formalized rule of guess that can be used along with the rules of inference of first order logic.
Circumscription involves formulating rules of thumb with “abnormality” predicates and then restricting
the extension of these predicates, circumscribing them, so that they apply to only those things to which
they are currently known.
Example
The rule of thumb is that “birds typically fly” is conditional. The predicate “Abnormal” signifies
abnormality with respect to flying ability.
Observe that the rule
∀x:birds(x) ┐Abnormal(x)flies(x)
Does not allow as to infer that “Tweety flies”, since we do not know that he is abnormal with respect
to flying ability. But if we add axioms which circumscribes the abnormality predicate to which they are
known say “Bird Tweety” then the inference can be drawn.

iii) Truth Maintenance System

The truth maintenance system maintains the consistency of a knowledge base as soon as new knowledge
is added. It considers only one state at a time so it is not possible to manipulate environment.
The functions of truth maintenance system are
 Provides justification for conclusions
When a problem solving system gives an answer to a user’s query, an explanation of that answer
is required.
Eg. An advice to a stockbroker is supported by an explanation of the reasons for that advice.
This is constructed by the inference engine by tracing the justification of the assertion.
 Recognize inconsistencies
The inference may tell the TMS that some sentences are contradictory. Then, TMS may find
that all those sentences are believed true, and reports to the inference engine which can
eliminate the inconsistencies by determining the assumptions used and changing them
appropriately.
Eg. A statement that either Abbott, or Babbitt, or Cabot ids guilty together with other statements
that Abbott is not guilty, Babbitt is not guilty, and Cabot is not guilty, forms a contradiction.
 Support default reasoning
In the absence of any firm knowledge, in many situations we want to reason from default
assumptions.
Eg. If “Tweety is a bird”, then until told otherwise, assume that “Tweety flies” and for
justification use the fact that “Tweety is a bird” and the assumption that “birs fly”.
Implementation issues
The issues related to implementation of non-monotonic reasoning in problem solving are..
 How to derive exactly those non-monotonic conclusions that are relevant to solving the
problem at hand while not wasting time on those that are not necessary.
 How to update our knowledge incrementally as problem solving progresses.
 How to overcome the problem where more than one interpretation of the known facts is
qualified or approved by the available inference rules.
 In general the theories are not computationally effective, decidable or semi decidable.
The solutions offered, considering the reasoning processes in to two parts
i) A problem solver that uses whatever mechanism it happens to have to draw conclusions as
necessary.
 ii) A truth maintenance system whose job is to maintain consistency in knowledge
representation of a knowledge base.

FUZZY REASONING
Fuzzy Logic (FL) is a method of reasoning that resembles human reasoning. The approach of
FL imitates the way of decision making in humans that involves all intermediate possibilities between
digital values YES and NO. In fuzzy logic, the degree of truth is between 0 and 1.
Example: William is smart (0.8 truth)
The fuzzy logic works on the levels of possibilities of input to achieve the definite output.
Fuzzy logic is useful for commercial and practical purposes.
 It can control machines and consumer products.
 It may not give accurate reasoning, but acceptable reasoning.
 Fuzzy logic helps to deal with the uncertainty in engineering.
Fuzzy Inference Systems Architecture
It has four main parts as shown −
 Fuzzification Module − It transforms the system inputs, which are crisp numbers, into fuzzy
sets. It splits the input signal into five steps such as −
LP x is Large Positive

MP x is Medium Positive

S x is Small

MN x is Medium Negative

LN x is Large Negative
 Knowledge Base − It stores IF-THEN rules provided by experts.
 Inference Engine − It simulates the human reasoning process by making fuzzy inference on
the inputs and IF-THEN rules.
 Defuzzification Module − It transforms the fuzzy set obtained by the inference engine into a
crisp value.

A membership function for a fuzzy set A on the universe of discourse X is defined as

µA:X→[0,1].Here, each element of X is mapped to a value between 0 and 1. It is called membership
value or degree of membership.
All membership functions for LP, MP, S, MN, and LN are shown as below −
Here, the input to 5-level fuzzifier varies from -10 volts to +10 volts. Hence the corresponding output
also changes.

Example of a Fuzzy Logic System

Let us consider an air conditioning system with 5-level fuzzy logic system. This system adjusts the
temperature of air conditioner by comparing the room temperature and the target temperature value.

Algorithm
 Define linguistic variables and terms.
 Construct membership functions for them.
 Construct knowledge base of rules.
 Convert crisp data into fuzzy data sets using membership functions. (fuzzification)
 Evaluate rules in the rule base. (interface engine)
 Combine results from each rule. (interface engine)
 Convert output data into non-fuzzy values. (defuzzification)
Logic Development
Step 1: Define linguistic variables and terms
Linguistic variables are input and output variables in the form of simple words or sentences. For room
temperature, cold, warm, hot, etc., are linguistic terms.
Temperature (t) = {very-cold, cold, warm, hot, Very-hot}
Step 2: Construct membership functions
A membership function (MF) is a curve that defines how each point in the input space is mapped to
a membership value (or degree of membership) between 0 and 1.
The membership functions of for Air conditioning system (temperature) variable are as shown −

Step3: Construct knowledge base rules

Create a matrix of room temperature values versus target temperature values that an air conditioning
system is expected to provide.
RoomTemp. Very-Cold Cold Warm Hot Very-Hot
/Target

Very_Cold No_Change Heat Heat Heat Heat

Cold Cool No_Change Heat Heat Heat

Warm Cool Cool No_Change Heat Heat

Hot Cool Cool Cool No_Change Heat

Very-Hot Cool Cool Cool Cool No_Change

Build a set of rules into the knowledge base in the form of IF-THEN-ELSE structures.
Sr. Condition Action
No.

1 IF temperature=(Cold OR Very_Cold) AND target=Warm THEN Heat

2 IF temperature=(Hot OR Very_Hot) AND target=Warm THEN Cool

3 IF (temperature=Warm) AND (target=Warm) THEN No_Change

Step 4: Obtain fuzzy value
Fuzzy set operations perform evaluation of rules. The operations used for OR and AND are Max and
Min respectively. Combine all results of evaluation to form a final result. This result is a fuzzy value.
For Example:
Room Temperature: 21
Target Temperature: 20
The Room temperature 21 falls in the fuzzy region warm and hot.
After plotting, it will give the membership value 0.1
Hence it is decided that 0.1 of hot region and 0.9(1-0.1) of warm region.
Now the fuzzy set is {0.1, 0.9}

Step 5: Perform defuzzification

Defuzzification is the process of producing a quantifiable result in Crisp logic, given fuzzy sets and
corresponding membership degrees. It is the process that maps a fuzzy set to a crisp set. It is typically
needed in fuzzy control systems. Defuzzification is then performed according to membership function
for output variable.

Application Areas of Fuzzy Logic

Automotive Systems
 Automatic Gearboxes
 Four-Wheel Steering
 Vehicle environment control
Consumer Electronic Goods
 Hi-Fi Systems
 Photocopiers
 Still and Video Cameras
 Television
Domestic Goods
 Microwave Ovens
 Refrigerators
 Toasters
 Vacuum Cleaners
 Washing Machines
Environment Control
 Air Conditioners/Dryers/Heaters
 Humidifiers
Advantages of Fuzzy Logic System
 Mathematical concepts within fuzzy reasoning are very simple.
 Able to modify Fuzzy Logic System by just adding or deleting rules due to flexibility of fuzzy
logic.
 Fuzzy logic Systems can take imprecise, distorted, noisy input information.
 FLSs are easy to construct and understand.
 Fuzzy logic is a solution to complex problems in all fields of life, including medicine, as it
resembles human reasoning and decision making.
Disadvantages of Fuzzy Logic System
 There is no systematic approach to fuzzy system designing.
 They are understandable only when simple.
 They are suitable for the problems which do not need high accuracy.
 TEMPORAL LOGIC
Temporal logic is representation and reasoning about time and temporal information with in a logical
frame work. In Temporal logics, evaluation takes place within a set of worlds. Thus, “it is Monday”
may be satisfied in some worlds but not in others.
Time Models
1. Instant-based model time model
In instance-based models, the primitive temporal entities are time instances and the basic relationships
between them is precedence. Thus the flow of time is represented as a set of time instances with a
binary relation of precedence on it, (T, <). Some of the basic properties of the instances-based model
may or may not be expressed with FOL sentences are as follows

 Reflexivity: ∀x(x<x)
 Respectively irreflexivity: ∀x(x<x)
 Transitivity: ∀x∀y∀z(x<yy<zx<z)
 Anti-symmetry: ∀x∀y(x<yy<xx=y)
 Trichotomy: ∀x∀y(x=yx<yy<x)
 Density: ∀x∀y(x<y∃z(x<zz<y))
 No beginning: ∀x∃y(y<x)
 No end: ∀x∃y(x<y)
 Every instance has an immediate successor: ∀x∃y(x<y∀z(x<zy≤z))
 Every instance has an immediate predessor: ∀x∃y(y<x∀z(z<xz≤y))
 This model is suitable for reasoning about real-world events with duration.

2. Interval/Period based time model

Interval based temporal models are ontologically much richer than instance-based ones, as there are
many more possible relationships between time intervals that between time instance. It include
relation precedence <, inclusion ⊆ and overlap O.
Some basic properties

 Reflexivity of ⊆: ∀x(x⊆x)
 Anti-symmetry of ⊆: ∀x∀y(x⊆yy⊆xx=y)
 Atomicity of ⊆:∀x∃y(y⊆x∀z(z⊆yz=y))
 Symmetry of O: ∀x∀y(xOyyOx)

Properties of TL
1. Safety Property
It implies that “something bad must not happen”. For example, system should not crash. Finite
length error trace should be there.
2. Liveness Property
It implies that “something good must happen”. For example, every packet sent must be received
at its destination. So there can be infinite-length error trace.
Semantics of TL
A temporal frame T=<T,<> defines the flow of time over which the meaning of the tense operators
are to be defined.
Operators in Temporal Logic
1. The temporal operators provide syntactic approach to linear temporal logic. These temporal
operators include always, never, next, until and before, amongst others.
2. The always operator holds if its operands holds in every single cycle, whilst the never
operator holds if its operand fails to hold in every single cycle. The next operator holds if
its operand holds in the cycle that immediately follows.
3. The next operator can take a number of cycles as an argument, enclosed in square brackets,
as in:
assert always req->next[2] (grant);
This means that whatever req is true, grant must be true two cycles later.
4. The meaning of the until operator
assert always req-> next(ack until grant);
This asserts that whenever req is true, ack is true in the following cycle and ack remains
true until the first subsequent cycle in which grant is true.
5. The before operator
assert req -> next (ack before grant);
This asserts that whenever req is true, ack must be true at least once in the period starting
in the following cycle ending in the last cycle before grant is true.
Linear Temporal Logic
It is a model logic for reasoning about dynamic scenarios similar to temporal modal logic. It contains

 More complex modal operators

 Compact, graphical representations of sets.
Modal operators are
G globally: true now and forever
F eventually (finally): true now or sometime in the future
X next: true in the next state
Uuntil: something true until something else happens
Applications of Temporal Logic

 Formalism for clarifying philosophical issues about time

 Framework within which to define the semantics of temporal expressions in natural language
 Language for encoding temporal knowledge in artificial intelligence
 Tool for specification, formal analysis and verification of the executions of computer
programs and systems.
NEURAL NETWORKS
A neural network consists of inter connected processing elements called neurons that work together
to produce an output function. The output of a neural network relies on the cooperation of the
individual neurons within the network to operate. Well-designed neural networks are trainable systems
that can often “learn” to solve complex problems from a set of exemplars and generalize the “acquired
knowledge” to solve unforeseen problems, i.e. they are self-adaptive systems. A neural network is
used to refer to a network of biological neurons. A neural network consists of a set of highly
interconnected entities called nodes or units. Each unit accepts a weighted set of inputs and responds
with an output.
 Mathematically let I = (I1, I2, … … In) represent the set of inputs presented to the unit U. Each
input has an associated weight that represents the strength of that particular
connection. Let W = (W1, W2, … … Wn) represent the weight vector corresponding to the input
vector X. By applying to
V, these weighted inputs produce a net sum at U given by

S = SUM (Wi × Ii)

FEATURES OF ARTIFICIAL NETWORK (ANN)

ANN is just a parallel computational system consisting of many simple processing elements
connected together in a specific way in order to perform a particular task. There are some important
features of artificial networks as follows.

(1) Artificial neural networks are extremely powerful computational devices (Universal
computers).
(2) ANNs are modeled on the basis of current brain theories, in which information is represented
by weights.
(3) ANNs have massive parallelism which makes them very e f f i c i e n t .
(4) They can learn and generalize from training data so there is no need for enormous feats of
programming.
(5) Storage is fault tolerant i.e. some portions of the neural net can be removed and there will be
only a small degradation in the quality of stored data.
(6) They are particularly fault tolerant which is equivalent to the “graceful degradation” found in
biological systems.
(7) Data are naturally stored in the form of associative memory which contrasts with conventional
memory, in which data are recalled by specifying address of that data.
(8) They are very noise tolerant, so they can cope with situations where normal symbolic systems
would have difficulty.
(9) In practice, they can do anything a symbolic/ logic system can do and more.
(10) Neural networks can extrapolate and intrapolate from their stored information. The neural
networks can also be trained. Special training teaches the net to look for significant features or
relationships of data.
TYPES OF NEURAL NETWORKS
Single Layer Network
A single layer neural network consists of a set of units organized in a layer. Each unit U n
receives a weighted input Ijwith weight Wjn. Figure shows a single layer neural network with j
inputs and outputs.

Single Layer neural Network

Multilayer Network
A multilayer network has two or more layers of units, with the output from one layer serving
as input to the next. Generally in a multilayer network there are 3 layers present like, input layer,
output layer and hidden layer. The layer with no external output connections are referred to as hidden
layers. A multilayer neural network structure is given in figure.

A multilayer neural network

Feed Forward neural network

In this network, the information moves in only one direction, forward from the input nodes,
through the hidden nodes and to the output nodes. There are no cycles or loops in the network. In other
way we can say the feed forward neural network is one that does not have any connections from output
to input. All inputs with variable weights are connected with every other node. A single layer feed
forward network has one layer of nodes, whereas a multilayer feed forward network has multiple layers
of nodes. The structure of a feed forward multilayer network is given in figure.

Multilayer Feed Forward Neural Network

Back Propagation neural network

Multilayer neural networks use a most common technique from a variety of learning technique, called
the back propagation algorithm. In back propagation neural network, the output values are compared
with the correct answer to compute the value of some predefined error function. By various techniques
the error is then fed back through the network. Using this information, the algorithms adjust the
weights of each connection in order to reduce the value of the error function by some small amount.
After repeating this process for a sufficiently large number of training cycles the network will usually
converge to some state where the error of the calculation is small.

The goal of back propagation, as with most training algorithms, is to iteratively adjust the
weights in the network to produce the desired output by minimizing the output error. The algorithm’s
goal is to solve credit assignment problem. Back propagation is a gradient-descent approach in that it
uses the minimization of first-order derivatives to find an optimal solution. The standard back
propagation algorithm is given below.

Step1: Build a network with the choosen number of input, hidden and output u n i t s .
Step2: Initialize all the weights to low random values.
Step3: Randomly, choose a single training pair.
Step4: Copy the input pattern to the input layer.
Step5: Cycle the network so that the activation from the inputs generates the activations in the
hidden and output layers.
Step6: Calculate the error derivative between the output activation and the final o u t p u t .
Step7: Apply the method of back propagation to the summed products of the weights and errors in
the output layer in order to calculate the error in the hidden units.
Step8:Update the weights attached the each unit according to the error in that unit, the output
from the unit below it and the learning parameters, until the error is sufficiently low.

MACHINE LEARNING
Machine learning is the systematic study of algorithms and systems that improve their
knowledge or performance (learn a model for accomplishing a task) with experience (from available
data /examples)
Examples:
 Given an URL decide whether it is a Sports website or not
 Given that a buyer is buying a book at online store, suggest some related products for
that buyer
 Given an ultrasound image of abdomen scan of a pregnant lady, predict the weight of
the baby
 Given a CT scan image set, decide whether there is stenosis or not
  Given marks of all the students in a class, assign relative grades based on statistical
distribution
 Given a mail received, check whether it is a SPAM
 Given a speech recording, identify the emotion of the speaker
 Given a DNA sequence, predict the promoter regions in that sequence

These are some examples of “Intelligent tasks” — tasks that are “easy” for humans but
“extremely difficult” for a machine to achieve Artificial Intelligence is about building systems that can
efficiently perform such “intelligent tasks”
One of the important aspects that enable humans to perform such intelligent tasks is their ability
to learn from experiences (either supervised or unsupervised). Machine learning tasks are typically
classified into three broad categories, depending on the nature of the learning "signal" or "feedback"
available to a learning system.

(i) Supervised learning

The computer is presented with example inputs and their desired outputs, given by a "teacher",
and the goal is to learn a general rule that maps inputs to outputs.

o Prediction
o Classification (discrete labels),
o Regression (real values)

Prediction
Example: Price of a used car
x : car attributes
y : price
y = g (x | θ )
θ parameters
Classification

Example 1

Suppose you have a basket and it is filled with different kinds of fruits. Your task is to arrange them
as groups. For understanding let me clear the names of the fruits in our basket.
You already learn from your previous work about the physical characters of fruits. So arranging the
same type of fruits at one place is easy now. Your previous work is called as training data in data mining.
You already learn the things from your train data; this is because of response variable. Response variable
means just a decision variable.

Suppose you have taken a new fruit from the basket then you will see the size, color and shape of

No. SIZE COLOR SHAPE FRUIT NAME

Rounded shape with a depression at the

1 Big Red top Apple

2 Small Red Heart-shaped to nearly globular Cherry

3 Big Green Long curving cylinder Banana

4 Small Green Round to oval, Bunch shape Cylindrical Grape

that particular fruit. If size is Big, color is Red, shape is rounded shape with a depression at the top, you
will conform the fruit name as apple and you will put in apple group.
If you learn the thing before from training data and then applying that knowledge to the test data
(for new fruit), this type of learning is called as Supervised Learning.

Regression

Given example pairs of heights and weights of a set of people, find a model to predict the weight
of a person from her height
(ii) Unsupervised learning

Unsupervised learning, no labels are given to the learning algorithm, leaving it on its own to
find structure in its input. Unsupervised learning can be a goal in itself (discovering hidden patterns in
data) or a means towards an end.

o Clustering

In clustering, a set of inputs is to be divided into groups. Unlike in

classification, the groups are not known beforehand, making this typically an
unsupervised task.
o Probability distribution estimation
o Finding association (in features)
o Dimension reduction
 Dimensionality reduction simplifies inputs by mapping them into a lower-
dimensional space. Topic modeling is a related problem, where a program is
given a list of human language documents and is tasked to find out which
documents cover similar topics.

Example
 Suppose you have a basket and it is filled with some different types fruits, your task is to arrange
them as groups.
 This time you don’t know anything about the fruits, honestly saying this is the first time you
have seen them. You have no clue about those.
 So, how will you arrange them? What will you do first???
 You will take a fruit and you will arrange them by considering physical character of that
particular fruit.
 Suppose you have considered color.
 Then you will arrange them on considering base condition as color.
 Then the groups will be something like this.
 RED COLOR GROUP: apples & cherry fruits.
 GREEN COLOR GROUP: bananas & grapes.
 So now you will take another physical character such as size.
 RED COLOR AND BIG SIZE: apple.
 RED COLOR AND SMALL SIZE: cherry fruits.
 GREEN COLOR AND BIG SIZE: bananas.
  GREEN COLOR AND SMALL SIZE: grapes.
 Job done happy ending.
 Here you did not learn anything before, means no train data and no response variable.
 This type of learning is known as unsupervised learning.
 Clustering comes under unsupervised learning.

(iii) Reinforcement learning

In reinforcement learning, a computer program interacts with a dynamic environment in which

it must perform a certain goal (such as driving a vehicle), without a teacher explicitly telling it whether
it has come close to its goal or not. Another example is learning to play a game by playing against an
opponent.

o Decision making (robot, chess machine)

Heuristic Reinforcement

NEURO-FUZZY INFERENCE
Neuro fuzzy system is the combination of fuzzy logic and neural network. Fuzzy logic and
neural networks are natural complementary tools in building intelligent systems. While neural networks
are low-level computational structures that perform well when dealing with raw data, fuzzy logic deals
with reasoning on a higher level, using linguistic information acquired from domain experts. However,
fuzzy systems lack the ability to learn and cannot adjust themselves to a new environment. On the other
hand, although neural networks can learn, they are opaque to the user.
Integrated neuro-fuzzy systems can combine the parallel computation and learning abilities of
neural networks with the human-like knowledge representation and explanation abilities of fuzzy
systems. As a result, neural networks become more transparent, while fuzzy systems become capable
of learning.
A neuro-fuzzy system is a neural network which is functionally equivalent to a fuzzy inference
model. It can be trained to develop IF-THEN fuzzy rules and determine membership functions for input
and output variables of the system. Expert knowledge can be incorporated into the structure of the
neuro-fuzzy system. At the same time, the connectionist structure avoids fuzzy inference, which entails
a substantial computational burden.
Neuro fuzzy system were created to solve the trade-off between:
– The mapping precision & automation of Neural Networks
– The interpretability of Fuzzy Systems

Steps in Development of NFS

• Development of Fuzzy Neural Models [Neurons]
• Development of synaptic connection models which incorporate fuzziness into Neural Network
[Weights]
• Development of Learning Algorithms [Method of adjusting weights]

Models of NFS
• Model 1: Fuzzy System → Neural Network
• Model 2: Neural Network → Fuzzy Systems
Model 1: Fuzzy System → Neural Network

Neural Network → Fuzzy System

The structure of a neuro-fuzzy system is similar to a multi-layer neural network. In general, a

neuro-fuzzy system has input and output layers, and three hidden layers that represent membership
functions and fuzzy rules.
Each layer in the neuro-fuzzy system is associated with a particular step in the fuzzy inference
process.
Layer 1 is the input layer. Each neuron in this layer transmits external crisp signals directly to the next
layer.

Layer 2 is the fuzzification layer. Neurons in this layer represent fuzzy sets used in the antecedents of
fuzzy rules. A fuzzification neuron receives a crisp input and determines the degree to which this input
belongs to the neuron’s fuzzy set.
The activation function of a membership neuron is set to the function that specifies the neuron’s fuzzy
set. A triangular membership function can be specified by two parameters {a, b} as follows:

Layer 3 is the fuzzy rule layer. Each neuron in this layer corresponds to a single fuzzy rule. A fuzzy
rule neuron receives inputs from the fuzzification neurons that represent fuzzy sets in the rule
antecedents. For instance, neuron R1, which corresponds to Rule 1, receives inputs from neurons A1
and B1. In a neuro-fuzzy system, intersection can be implemented by the product operator.

Layer 4 is the output membership layer. Neurons in this layer represent fuzzy sets used in the
consequent of fuzzy rules. An output membership neuron combines all its inputs by using the fuzzy
operation union.

Layer 5 is the defuzzification layer. Each neuron in this layer represents a single output of the neuro-
fuzzy system. It takes the output fuzzy sets clipped by the respective integrated firing strengths and
combines them into a single fuzzy set. Neuro-fuzzy systems can apply standard defuzzification
methods, including the centroid technique. The sum-product composition calculates the crisp output as
the weighted average of the centroids of all output membership functions. For example, the weighted
average of the centroids of the clipped fuzzy sets C1 and C2 is calculated as,
A neuro-fuzzy system is essentially a multi-layer neural network, and thus it can apply standard learning
algorithms developed for neural networks, including the back-propagation algorithm.

Applications of NFS
• Measuring opacity/transparency of water in washing machine – Hitachi, Japan
• Improving the rating of convertible bonds – Nikko Securities, Japan
• Adjusting exposure in photocopy machines – Sanyo, Japan
• Electric fan that rotates towards the user – Sanyo, Japan

Comparison of fuzzy logic and neural network

Point Fuzzy Systems Neural Network
Knowledge Source Human Experts Sample Sets
Learning Mechanism Induction Adjusting Weights
Reasoning Mechanism Heuristic Search Parallel Computation
Learning Speed High Low
Reasoning Speed Low High
Fault Tolerance Low Very High
Implementation Explicit Implicit
Flexibility Low High