You are on page 1of 14

1.

Choose the correct alternative from the following

51. Unsupervised learning is


learning without computers problem based learning
learning from environment learning from teachers
52.
Perceptron is

General class of approaches to a Performing several computations


problem. simultaneously

Structures in a database those Simple forerunner of modern neural


are statistically relevant networks, without hidden layers.

53. In supervised learning


classes are not predefined classes are predefined
classes are not required classification is not done
54. What is composed of large number of highly interconnected processing elements
(neurons) working in unison to solve problems.
ANN Perceptron
neuron None of these
55. Artificial neural network used for
Pattern Recognition Classification
Clustering All of these
56. A Neural Network can answer
For Loop questions what-if questions
IF-Then-Else Analysis None of these
Questions
57. Ability to learn how to do tasks based on the data given for training or initial
experience
Self-Organization Adaptive Learning
Fault tolerance Robustness
58. When the cell is said to be fired
if potential of body reaches a if there is impulse reaction
steady threshold values
during upbeat of heart None of these
59. In artificial Neural Network interconnected processing elements are called
nodes or neurons weights
axons Soma
60. Each connection link in ANN is associated with ________ which has information
about the input signal.
neurons weights
bias activation function
61. What was the 2nd stage in perceptron model called?
sensory units summing unit
association unit output unit
62.
which is called Internal state of neuron that is the function of the inputs the
neurons receives

Weight activation or activity level of neuron


Bias None of these
63. How many signal(s) can Neuron send at a time?
multiple 1
2 None of these
64.
Artificial neural network is based on the concept

gradient descent technique random descent technique

biological neuron (a) and (c).

65. Kohonen network / adaptive resonance theory model is


purely unsupervised supervised

reinforcement none of these.

66. Classification of XOR input pattern is


separable by perceptron model separable by Kohonen model
separable by multilayer all of these.
feedback network

67. In ANN model


learning constant should be should be constant throughout the
small epoch

should be ‘one’ Should be small but adaptive and


remain stable to irrelevant input.

68.
Acquired knowledge is stored in the ANN with the help of

activation function local induced field

synaptic weight input signal.

69. Expert systems


Combining different types of Approach to the design of learning
method or information algorithms that is structured along the
lines of the theory of evolution
an information base filled with None of these
the knowledge of an expert
formulated in terms of if-then
rules
70. Falsification is
Modular design of a software Showing a universal law or rule to be
application that facilitates the invalid by providing a counter example
integration of new modules
A set of attributes in a database None of these
table that refers to data in
another table
71. Evolutionary computation is
Combining different types of Approach to the design of learning
method or information algorithms that is structured along the
lines of the theory of evolution.
Decision support systems that None of these
contain an information base
filled with the knowledge of an
expert formulated in terms of
if-then rules.
72. Extendible architecture is
Modular design of a software Showing a universal law or rule to be
application that facilitates the invalid by providing a counter example
integration of new modules
A set of attributes in a database None of these
table that refers to data in
another table
73. Massively parallel machine is
A programming language based A computer where each processor has
on logic its own operating system, its own
memory, and its own hard disk
Describes the structure of the None of these
contents of a database.
74. Search space
The large set of candidate The information stored in a database
solutions possible for a problem that can be retrieved with a single
query.
Worth of the output of a None of these
machine learning program that
makes it understandable for
humans
75. Shallow knowledge
The large set of candidate The information stored in a database
solutions possible for a problem that can be retrieved with a single query
Worth of the output of a All of these
machine learning program that
makes it understandable for
humans
76. Quantitative attributes are
A reference to the speed of an Attributes of a database table that can
algorithm, which is take only numerical values
quadratically dependent on the
size of the data
Tools designed to query a All of these
database
77. Subject orientation
The science of collecting, Measure of the probability that a
organizing, and applying certain hypothesis is incorrect given
numerical facts certain observations.
One of the defining aspects of a None of these
data warehouse, which is
specially built around all the
existing applications of the
operational data
78. Vector
It does not need the control of An arrow in a multi-dimensional space.
the human operator during It is a quantity usually characterized by
their execution an ordered set of scalars
The validation of a theory on All of these
the basis of a finite number of
examples
79. Transparency
The large set of candidate The information stored in a database
solutions possible for a problem that can be retrieved with a single query
Worth of the output of a None of these
machine learning program that
makes it understandable for
humans
80. An auto-associative network is
a neural network that contains a neural network that contains feedback
no loops
a neural network that has only a single layer feed-forward neural
one loop network with pre-processing
81. Which of the following is true?
(i) On average, neural networks have higher computational rates than
conventional computers.
(ii) Neural networks learn by example.
(iii) Neural networks mimic the way the human brain works.
All of these ii and iii
i, ii and iii None of these
82. Which of the following is true for neural networks?
(i) The training time depends on the size of the network.
(ii) Neural networks can be simulated on a conventional computer.
(iii) Artificial neurons are identical in operation to biological ones.
All of these ii and iii
i and ii None of these
83. What are the advantages of neural networks over conventional computers?
(i) They have the ability to learn by example
(ii) They are more fault tolerant
(iii) They are more suited for real time operation due to their high
‘computational’ rates
All of these ii and iii
i, ii None of these
84 Which of the following is true?
Single layer associative neural networks do not have the ability to:
(i) perform pattern recognition
(ii) find the parity of a picture
(iii)determine whether two or more shapes in a picture are connected or not
All of these ii and iii
i, ii None of these
85 Factors which affect the performance of learner system does not include
Representation scheme used Training scenario
Type of feedback Good data structures
86 Different learning methods does not include
Memorization Analogy
Deduction Introduction
87 In language understanding, the levels of knowledge that does not include
Phonological Syntactic
Empirical Logical
88 A model of language consists of the categories which does not include
Language units Role structure of units

System constraints Structural units


89 What is a top-down parser?
Begins by hypothesizing a Begins by hypothesizing a sentence (the
sentence (the symbol S) and symbol S) and successively predicting
successively predicting lower upper level constituents until individual
level constituents until preterminal symbols are written
individual preterminal symbols
are written
Begins by hypothesizing lower Begins by hypothesizing upper level
level constituents and constituents and successively predicting
successively predicting a a sentence (the symbol S)
sentence (the symbol S)
90 Among the following which is not a horn clause?
p Øp V q
p→q p → Øq
91 Which instruments are used for perceiving and acting upon the environment?
Sensors and Actuators Sensors
Perceiver None of these
92 What is meant by agent’s percept sequence?
Used to perceive the Complete history of actuator
environment
Complete history of perceived None of these
things
93 Single layer perceptron is used for
linear seperability error minimization
non-linear separability annealing.
94 Which mode of training, all neurons in Hopfield networks fire at random.
Stable status Output
Synchronous Asynchronous
95 X-OR problem can be solved by
single layer perceptron Bayes theorem
multi-layer perceptron single layer perception and Bayes
theorem
96 For a network with inputs [ x1, x2, x3 ]= [ 0 3,0.5,0.6] and weights [w1,w2 ,w3 ] = [
0 2, 0 1, -0 3] , the net output to the output M-P neuron is
0·07 – 0·07
0·7 – 0·7.
97 In fuzzy logic controller the rule base is updated by the help of
Fuzzy logic ANN and GA
Hebb's rule none of these.
98 If a problem is linearly separable, then we can find the weight vector Wi, such
that WiT . X > 0 for each X, with desired output '+1' and WiT . X < 0 , for each X,
with
desired output '–1'.

True False
Can't say None of these.
99 For a 3-input neuron representing a perceptron, where [x1 ,x2 ,x3 ] = [ 0 8, 0 6, 0
4 ] and weights [ w1,w2 ,w3 ]= [ 0 1, 0 3, 0 2 ] and bias b = 0·35, the output of the
neuron using bipolar sigmoid activation function is
0·625 0·764
0·259 0·346.
100 Which one is unsupervised learning rule?
Error-correction Delta
Widrow-Hoff Hebbian.
101 In simple perceptron learning rule change of synaptic weight is proportional to
product of error and i/p signal product of i/p and o/p signals
product of error and o/p signal Gradient of cost function.
102 Which is the most direct application of neural networks?
vector quantization pattern mapping
pattern classification control applications
103 What are pros of neural networks over computers?
they have ability to learn b they have real time high computational
examples rates
they have more tolerance All of these
104 What is the rule of simple reflex agent
Simple-action rule Condition-action rule
Simple & Condition-action rule None of these
105 Neuro software is:
A software used to analyze It is powerful and easy neural network
neurons
Designed to aid experts in real It is software used by Neurosurgeon
world
106 A 3-input neuron is trained to output a zero when the input is 110 and a one when
the input is 111. After generalization, the output will be zero when and only when
the input is:
000 or 110 or 011 or 101 010 or 100 or 110 or 101
000 or 010 or 110 or 100 100 or 111 or 101 or 001
107 A 4-input neuron has weights 1, 2, 3 and 4. The transfer function is linear with the
constant of proportionality being equal to 2. The inputs are 4, 10, 5 and 20
respectively. The output will be:
238 76
119 123
108 What happens in upper subnet of the hamming network?
classification storage
output None of these
109 The competition in upper subnet of hamming network continues till?
only one unit remains negative all units are destroyed
output of only one unit remains None of these
positive
110 Which is true for neural networks?
It has set of nodes and Each node computes it’s weighted input
connections
Node could be in excited state All of these
or non-excited state
111 What is shape of dendrites like
oval round
tree rectangular
112 What are the issues on which biological networks proves to be superior than AI
networks?
robustness & fault tolerance flexibility
collective computation All of these
113 What are dendrites?
fibers of nerves nuclear projections
other name for nucleus None of these
114 Feature of ANN in which ANN creates its own organization or representation of
information it receives during learning time is
Adaptive Learning Self-Organization
What-If Analysis Supervised Learning
115 Signal transmission at synapse is a?
physical process chemical process
physical & chemical both None of these
116 What is purpose of Axon?
receptors transmitter
transmission None of these
117 What is approx size of neuron body(in micrometer)
below 5 5-10
10-80 above 100
118 Where does the chemical reaction take place in neuron?
dendrites axon
synapses nucleus
119 Function of dendrites is?
Receptors transmitter
both receptor & transmitter None of these
120 Prolog is
A programming language based A computer where each processor has
on logic its own operating system, its own
memory, and its own hard disk
Describes the structure of the None of these
contents of a database
121 Back propagation is a learning technique that adjusts weights in the neural
network by propagating weight changes.
Forward from source to sink Backward from sink to source
Forward from source to hidden Backward from sink to hidden node
nodes
122 What is the gap at synapses(in nanometer)
50 100
150 200
123 An artificial neuron receives n inputs x1, x2, x3............xn with weights w1, w2,
..........wn 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.
Σ wi Σ xi
Σ wi* xi None of these
124 How does the transmission/pulse acknowledged
by lowering electric potential of by raising electric potential of neuron
neuron body body
both by lowering & raising None of these
electric potential
125 Slots and facets are used in
Semantic Networks Frames
Rules All of these
126 A neuron with 3 inputs has the weight vector [0.2 -0.1 0.1] and a bias θ = 0. If the
input vector is X = [0.2 0.4 0.2] then the total input to the neuron is:
0.20 1.0
0.02 -1.0
127. Which of the following neural networks uses supervised learning?
(A) Multilayer perceptron
(B) Self organizing feature map
(C) Hopfield network
(A) only (B) only
(A) and (B) only (A) and (C) only
128. What is charge at protoplasm in state of inactivity
positive negative
neutral may be positive or negative
129. What is the main constituent of neural liquid?
sodium potassium
Iron None of these
130. What is average potential of neural liquid in inactive state
+70mv +35mv
-35mv -70mv
131. What is effect on neuron as a whole when its potential get raised to -60mv
it get fired no effect
it get compressed it expands
132. What is the function of neurotransmitter?
they transmit data directly at they modify conductance of post
synapse to other neuron synaptic membrane for certain ions
cause polarization or both polarization & modify conductance
depolarization of membrane
133. The cell body of neuron can be analogous to what mathematical operation?
Summing Differentiator
integrator None of these
134. What is the critical threshold voltage values at which neuron get fired?
30mv 20mv
25mv 10mv
135. Why can’t we design a perfect neural network?
full operation is still not known number of neuron is itself not precisely
of biological neurons known
number of interconnection is All of these
very large & is very complex
136. How many synaptic connections are there in human brain?
1010 1015
1020 105
137. Operations in the neural networks can perform what kind of operations?
serial parallel
serial or parallel None of these
138. Which action is faster pattern classification or adjustment of weights in neural
nets?
pattern classification adjustment of weights
equal either of them can be fast, depending on
conditions
139. What is the feature of ANNs due to which they can deal with noisy, fuzzy,
inconsistent data?
associative nature of networks distributive nature of networks
both associative & distributive None of these
140. What was the name of the first model which can perform weighted sum of inputs?
McCulloch-pitts neuron model Marvin Minsky neuron model
Hopfield model of neuron None of these
141. Who developed the first learning machine in which connection strengths could be
adapted automatically?
McCulloch-pitts Marvin Minsky
Hopfield None of these
142. Who proposed the first perceptron model in 1958?
McCulloch-pitts Marvin Minsky
Hopfield Rosenblatt
143. John Hopfield was credited for what important aspect of neuron?
learning algorithms adaptive signal processing
energy analysis None of these
144. What is the contribution of Ackley, Hinton in neural?
Perceptron Boltzmann machine
learning algorithms None of these
145. What is ART in neural networks
automatic resonance theory artificial resonance theory

adaptive resonance theory None of these


146. What is an activation value?
weighted sum of inputs threshold value
main input to neuron None of these
147. Positive sign of weight indicates?
excitatory input inhibitory input
can be either excitatory or None of these
inhibitory as such
148. Negative sign of weight indicates?
excitatory input inhibitory input
excitatory output inhibitory output
149. The amount of output of one unit received by another unit depends on what?
output unit input unit
activation value weight
150. The process of adjusting the weight is known as?
activation synchronization
learning None of these
151. The procedure to incrementally update each of weights in neural is referred to
as?
synchronisation learning law
learning algorithm both learning algorithm & law
152. In what ways can output be determined from activation value?
deterministically stochastically
both deterministically & None of these
stochastically
153. How can output be updated in neural network?
synchronously asynchronously
both synchronously & None of these
asynchronously
154. What is asynchronous update in neural networks?
output units are updated output units are updated in parallel
sequentially fashion
can be either sequentially or in None of these
parallel fashion

You might also like