Q.

No Question Body

1 Why do we need biological neural networks?

2 What is unsupervised learning?

3 The fundamental unit of network is

4 How many types of Artificial Neural Networks?
5 Neural Networks are complex ______________ with many parameters.
6 The output at each node is called_____.
Feature of ANN in which ANN creates its own organization or representation of information
7
it receives during learning time is
The network that involves backward links from output to the input and hidden layers is called
8
as____.
9 Automated vehicle is an example of ______
Identify the following activation function :
10 φ(V) = Z + (1/ 1 + exp (– x * V + Y) ),
Z, X, Y are parameters

An artificial neuron receives n inputs x1, x2, x3............xn with weights w1, w2, ..........wn
11 attached to the input links. The weighted sum_________________ is computed to be passed
on to a non-linear filter Φ called activation function to release the output.

Which of the following neural networks uses supervised learning?

(A) Multilayer perceptron
12
(B) Self organizing feature map
(C) Hopfield network

13 A perceptron is:

14 Back Propagation is _____

15 What are general limitations of back propagation rule?

16 How can learning process be stopped in backpropagation rule?

17 What is synchronous update in hopfield model?

18 An auto – associative network is?

19 What type of inputs does ART – 1 receives?

20 Which is the most direct application of neural networks?

21 What is an auto-associative network?

22 What is Neuro software?

 A 4-input neuron has weights 1, 2, 3 and 4. The transfer function is linear with the constant of
23 proportionality being equal to 2. The inputs are 4, 10, 5 and 20 respectively. What will be the
output?

24 Why are linearly separable problems of interest of neural network researchers?

25 How many output Fuzzy Logic produce?

26 Fuzzy Set theory defines fuzzy operators. Choose the fuzzy operators from the following.

The room temperature is hot. Here the hot (use of linguistic variable is used) can be
27
represented by _______
28 What action to take when IF (temperature=Warm) AND (target=Warm) THEN?

29 The values of the set membership is represented by

30 The truth values of traditional set theory is ____________ and that of fuzzy set is __________

31 What is Fuzzy Inference Systems?

32 What is another name for Fuzzy Inference Systems?

33 Mamdani's Fuzzy Inference Method was designed to attempt what?

34 Which chaining rule has inductive approch?

35 A rule-based system can be simply created by using?
36 What is the form of Knowledge representation?
37 How many level of fuzzifier is there?
The fuzzy results generated can not be used in an application, where
38
decision has to be taken only on____________________.
39 Which one of the following is not a defuzzification methods?
40 How many steps involved in designing a fuzzy logic controller?

41 The rules in the fuzzy rule base are

42 The decision-making model is characterized by

43 The principle of mass and velocity
44 Centroid method/s is /are

45 An algorithm is complete if ____________

Knowledge may be
I. Declarative.
46
II. Procedural.
III. Non-procedural.
47 What are the following sequence of steps taken in designing a fuzzy logic machine?

48 Where is the minimum criterion used ?

49 The word "genetics" is derived from the____________.
50 The basic terminologies in Genetic algorithm is /are
51 Genetic Alogorithm handles _______________ of possible solutions.
The limitations of Genetic algorithm are
I. The problem of Identifying fitness function
52
II. Definition of reoresentationfor the problem
III. Premature covergence occurs
The PSO parameters can also be tuned by using another overlaying optimizer, a concept
53
known as

_____________________ is a computational method that optimizes a problem by iteratively

54
trying to improve a candidate solution with regard to a given measure of quality.

55 A neuro-fuzzy hybrid system was proposed by

56 Neuro-Fuzzy Hybrid Systems can be classified into

57 The weight can be detemined in genetic algorithm based on back-propagation network is

58 GFRBSs stands for

59 Which are the genetic learning of rules bases approach?

60 A neuro-fuzzy hybrid system also called as
Option 1 Option 2
to solve tasks like machine vision & to apply heuristic search methods to
natural language processing find solutions of problem

features of group explicitly stated number of groups may be known

brain nucleus
2 3
Nonlinear Functions Discrete Functions
node value Weight

Adaptive Learning Self Organization

Self organizing maps Perceptrons

Supervised learning Unsupervised learning

Step function Ramp function

Σ wi Σ xi

(A) only (B) only

a single layer feed-forward neural

an auto-associative neural network
network with pre-processing

the transmission of error back

another name given to the curvy
through the network to adjust the
function in the perceptron
inputs

local minima problem slow convergence

there is convergence involved no heuristic criteria exist
a unit is selected at random and its
all units are updated simultaneously
new state is computed
network in neural which contains network in neural which contains
feedback loops
bipolar binary
vector quantization pattern mapping
a neural network that contains no a neural network that contains
loops feedback
It is powerful and easy neural
A software used to analyze neurons
network
 238 76

Because they are the only class of Because they are the only class of
problem that network can solve problem that Perceptron can solve
successfully successfully
2 3

AND OR

Fuzzy Set Crisp Set

Heat No_Change

Discrete Set Degree of truth

Either 0 or 1, between 0 & 1 Between 0 & 1, either 0 or 1

The process of formulating the Changing the output value to match

mapping from a given input to an the input value to give it an equal
output using fuzzy logic balance
Fuzzy Expert System Fuzzy Modelling

Control any two combinations of any Control a television and remote

two products by synthesising a set of combination by synthesising a set of
linguistic control rules obtained from linguistic control rules obtained from
experienced human operations. experienced human operations

Forward Chaining Backward chaining

assertions rules
IF-THEN IF-THEN-ELSE
4 5

crisp values fuzzy values

Lambda-cut method Centroid methods

7 8
R1: If x is A then y is C R1: If x is A then y is B
R2: If x is B then y is D R2: If x is B then y is A
set of non possible actions set of possible actions
a=mv a=m/v
Center of gravity method Center of sum method (CoS)
It terminates with a solution when
It starts with a solution
one exists

Only (I) Only (II)

Fuzzy Sets->Defuzzification->Rule Rule evaluation->Fuzzification-
evaluation >Defuzzification
When there is an OR operation When there is an AND operation
Spanish word German word
Indiviuals Genes
selection evaluation

Only (I) Only (II)

meta-optimization optimization scenarios

Adaptive mechanisms metaheuristic

J.S.R. Jang Lecun

2 3

(n/m)l n*m*l

Genetic Fuzzy Reasoning Based

Genetic Fuzzy Rules Based Systems
Systems
Pittsburgh Michigan
fuzzy null hybrid fuzzy neural hybrid
Option 3 Option 4
to make smart human interactive &
all of the mentioned
user friendly system
neither feature & nor number of
none of the mentioned
groups is known
neuron axon
4 5
Linear Functions Exponential Functions
neurons axons

What-If Analysis Supervised Learniing

Recurrent neural network Multi layered perceptron

Active learning Reinforcement learning

Sigmoid function Gaussian function

Σ wi + Σ xi Σ wi* xi

(A) and (B) only (A) and (C) only

a double layer auto-associative neural a neural network that contains

network feedback
the transmission of error back
through the network to allow weights
None of these
to be adjusted so that the network can
learn.
scaling all of the mentioned
on basis of average gradient value none of the mentioned
a predefined unit is selected and its
none of the mentioned
new state is computed

network in neural which no loops none of the mentioned

both bipolar and binary none of the mentiobned

pattern classification control applications
a neural network that has only one a single layer feed-forward neural
loop network with pre-processing

Designed to aid experts in real world It is software used by Neurosurgeon

 119 123

Because they are the only

Because they are the only
mathematical functions that are
mathematical functions you can draw
continue
4 5

NOT All of the above

Both Fuzzy Set & Crisp Set None of the Answers

Cool None of the Above

Probabilities Both Degree of truth & Probabilities

Between 0 & 1, between 0 & 1 Either 0 or 1, either 0 or 1

Having a smaller output than the

Having a larger output than the input
input

Fuzzy Logic Controller All of the above

Control a steam engine and a boiler Control a air craft and feul level
combination by synthesising a set of combination by synthesising a set of
linguistic control rules obtained from linguistic control rules obtained from
experienced human operations. experienced human operations.

Spanning Chaining All of the answers

set of assertions All of the answers
IF-ELSE All of the above
6 7

both crisp value & fuzzy values none of the answers

first in last out methods Maxima methods

9 10
R1: If x is A then y is A R1: If x is A then y is A
R2: If x is B then y is B R2: If x is B then y is B
set of non-indiviual set of indiviual
a=(m/v)2 a-=Sqr root of mv
Center of area method (CoA) All the answers

It does not terminate with a solution It has a loop

Only (III) Both (I) and (II)

Fuzzification->Rule evaluation- Defuzzification->Rule evaluation-
>Defuzzification >Fuzzification
In De Morgan's theorem None of the answers
Greek word Hindi word
Fitness All the answers
population generation

Only (III) (I), (II), and (III)

convergence to the optimum empirical

Particle Swarm Optimization gradient

Parker Rumelhart
4 5

(n + m) l (n/m)2l

Genetic Fuzzufucation Rules Based

Genetic Fuzzy Rules Biased Systems
Systems
Oiterative rule learning All the answers
fuzzy network hybrid fuzzy neural hopfield