Id 1

Question Back propagation is a learning technique that adjusts weights in the neural network by
propagating weight changes.
A Forward from source to sink
B Backward from sink to source
C Forward from source to hidden nodes
D Backward from sink to hidden nodes
Answer
Marks 1.5
Unit I
Id 2
Question Identify the following activation function :
φ(V) = Z + (1/ 1 + exp (– x * V + Y) ),
Z, X, Y are parameters
A Step function
B Ramp function
C Sigmoid function
D Gaussian function
Answer
Marks 1.5
Unit II
Id 3
Question A neuron with 3 inputs has the weight vector [0.2 -0.1 0.1]^T and a bias θ = 0. If the input
vector is X = [0.2 0.4 0.2]^T then the total input to the neuron is:
A 0.2
B 1
C 0.02
D -1.0
Answer
Marks 1.5
Unit I
Id 4
Question Which of the following neural networks uses supervised learning?
(A) Multilayer perceptron
(B) Self organizing feature map
(C) Hopfield network
A (A) only
B (B) only
C (A) and (B) only
D (A) and (C) only
Answer
Marks 1.5
Unit II
Id 5
Question In Delta Rule for error minimization
A weights are adjusted w.r.to change in the output
B weights are adjusted w.r.to difference between desired output and actual output
C weights are adjusted w.r.to difference between input and output
D none of the above
Answer
Marks 1.5
Unit I
Id 6
Question Which of the following can be used for clustering of data ?
A Single layer perception
B Multilayer perception
C Self organizing map
D Radial basis function
Answer
Marks 1.5
Unit I
Id 7
Question A network is created when we multiple neurons stack together. Let us take an
example of a neural network simulating an XNOR function.

You can see that the last neuron takes input from two neurons before it. The
activation function for all the neurons is given by:

Suppose X1 is 0 and X2 is 1, what will be the output for the above neural
network?

A 0
B 1
C 2
D 3
Answer
Marks 1.5
Unit I
Id 8
Question In a neural network, knowing the weight and bias of each neuron is the most important
step. If you can somehow get the correct value of weight and bias for each neuron, you
can approximate any function. What would be the best way to approach this?
A Assign random values
B Search every possible combination of weights and biases till you get the best value
C Iteratively check that after assigning a value how far you are from the best values, and
slightly change the assigned values values to make them better
D None of these
Answer
Marks 1.5
Unit I
Id 9
Question What are the steps for using a gradient descent algorithm?
1. Calculate error between the actual value and the predicted value
2. Reiterate until you find the best weights of network
3. Pass an input through the network and get values from output layer
4. Initialize random weight and bias
5. Go to each neurons which contributes to the error and change its respective values to
reduce the error
A 1, 2, 3, 4, 5
B 5, 4, 3, 2, 1
C 3, 2, 1, 5, 4
D 4, 3, 1, 5, 2
Answer
Marks 1.5
Unit II
Id 10
Question Suppose you have inputs as x, y, and z with values -2, 5, and -4 respectively. You
have a neuron ‘q’ and neuron ‘f’ with functions:
q=x+y

f=q*z

Graphical representation of the functions is as follows:

What is the gradient of F with respect to x, y, and z?

A (-3,4,4)
B (4,4,3)
C (-4,-4,3)
D (3,-4,-4)
Answer
Marks 1.5
Unit I
Id 11
Question Which of the following techniques perform similar operations as dropout in a neural
network?
A Bagging
B Boosting
C Stacking
D None of these
Answer
Marks 1.5
Unit II
Id 12
Question Which of the following gives non-linearity to a neural network?
A Stochastic Gradient Descent
B Rectified Linear Unit
C Convolution function
D None of the above
Answer
Marks 1.5
Unit III
Id 13
Question In training a neural network, you notice that the loss does not decrease in the
few starting epochs.

The reasons for this could be:

1.The learning is rate is low

2.Regularization parameter is high

3.Stuck at local minima

What according to you are the probable reasons?

A 1 and 2
B 2 and 3
C 1 and 3
D Any of these
Answer
Marks 1.5
Unit I
Id 14
Question You are building a neural network where it gets input from the previous layer as
well as from itself.

Which of the following architecture has feedback connections?

A Recurrent Neural network
B Convolutional Neural Network
C Restricted Boltzmann Machine
D None of these
Answer
Marks 1.5
Unit II
Id 15
Question What is the sequence of the following tasks in a perceptron?
1. Initialize weights of perceptron randomly
2. Go to the next batch of datase
3. If the prediction does not match the output, change the weights
4. For a sample input, compute an output

A 1, 2, 3, 4
B 4, 3, 2, 1
C 3, 1, 2, 4
D 1, 4, 3, 2
Answer
Marks 1.5
Unit I
Id 16
Question Suppose that you have to minimize the cost function by changing the parameters. Which
of the following technique could be used for this?
A Exhaustive Search
B Random Search
C Bayesian Optimization
D Any of these
Answer
Marks 1.5
Unit I
Id 17
Question In which neural net architecture, does weight sharing occur?
A Convolutional neural Network
B Recurrent Neural Network
C Fully Connected Neural Network
D Both A and B
Answer
Marks 1.5
Unit II
Id 18
Question Batch Normalization is helpful because

A It normalizes (changes) all the input before sending it to the next layer
B It returns back the normalized mean and standard deviation of weights
C It is a very efficient backpropagation technique
D None of these
Answer
Marks 1.5
Unit III
Id 19
Question Instead of trying to achieve absolute zero error, we set a metric called bayes error which is
the error we hope to achieve. What could be the reason for using bayes error?

A Input variables may not contain complete information about the output variable
B System (that creates input-output mapping) may be stochastic
C Limited training data
D All the above
Answer
Marks 1.5
Unit III
Id 20
Question In a neural network, which of the following techniques is used to deal with overfitting?

A Dropout
B Regularization
C Batch Normalization
D All of these
Answer
Marks 1.5
Unit II
Id 21
Question What is a dead unit in a neural network?

A A unit which doesn't update during training by any of its neighbour

B A unit which does not respond completely to any of the training patterns
C The unit which produces the biggest sum-squared error
D None of these
Answer
Marks 1.5
Unit I
Id 22
Question Which of the following statement is the best description of early stopping?
A Train the network until a local minimum in the error function is reached
B Simulate the network on a test dataset after every epoch of training. Stop training when
the generalization error starts to increase
C Add a momentum term to the weight update in the Generalized Delta Rule, so that
training converges more quickly
D A faster version of backpropagation, such as the "Quickprop' algorithm
Answer
Marks 1.5
Unit III
Id 23
Question For a classification task, instead of random weight initializations in a neural network, we
set all the weights to zero. Which of the following statements is true?

A There will not be any problem and the neural network will train properly
B The neural network will train but all the neurons will end up recognizing the same thing
C The neural network will not train as there is no net gradient change
D None of these
Answer
Marks 1.5
Unit I
Id 24
Question There is a plateau at the start. This is happening because the neural network
gets stuck at local minima before going on to global minima.

To avoid this, which of the following strategy should work?

A Increase the number of parameters, as the network would not get stuck at local minima

B Decrease the learning rate by 10 times at the start and then use momentum
C Jitter the learning rate, i.e. change the learning rate for a few epochs
D None of these
Answer
Marks 1.5
Unit III
Id 25
Question For an image recognition problem (recognizing a cat in a photo), which
architecture of neural network would be better suited to solve the problem?
A Multi Layer Perceptron
B Convolutional Neural Network
C Recurrent Neural network
D Perceptron
Answer
Marks 1.5
Unit II
Id 26
Question

Which of the following is a decision boundary of Neural Network?

A B
B A
C D
D All of these
Answer
Marks 1.5
Unit I
Id 27
Question What are the factors to select the depth of neural network?
1.Type of neural network (eg. MLP, CNN etc)

2.Input data

3.Computation power, i.e. Hardware capabilities and software capabilities

4.Learning Rate

5.The output function to map

A 1, 2, 4, 5
B 2, 3, 4, 5
C 1, 3, 4, 5
D All of these
Answer
Marks 1.5
Unit II
Id 28
Question Consider the scenario. The problem you are trying to solve has a small amount of data.
Fortunately, you have a pre-trained neural network that was trained on a similar problem.
Which of the following methodologies would you choose to make use of this pre-trained
network?
A Re-train the model for the new dataset
B Assess on every layer how the model performs and only select a few of them
C Fine tune the last couple of layers only
D Freeze all the layers except the last, re-train the last layer
Answer
Marks 1.5
Unit III
Id 29
Question Artificial intelligence is--------------
A It uses machine-learning techniques. Here program can learn From past experience and
adapt themselves to new situations
B Computational procedure that takes some value as input and produces some value as
output.
C Science of making machines performs tasks that would require intelligence when
performed by humans
D None of these
Answer
Marks 1.5
Unit I
Id 30
Question Evolutionary computation is-----------
A Combining different types of method or information
B Approach to the design of learning algorithms that is structured along the lines of the
theory of evolution.
C Decision support systems that contain an information base filled with the knowledge of an
expert formulated in terms of if-then rules.
D None of these
Answer
Marks 1.5
Unit I
Id 31
Question Search space is-------
A The large set of candidate solutions possible for a problem
B The information stored in a database that can be, retrieved with a single query.
C Worth of the output of a machine learning program that makes it understandable for
humans
D None of these
Answer
Marks 1.5
Unit III
Id 32
Question Perceptron is____________
A General class of approaches to a problem.
B Performing several computations simultaneously
C Structures in a database those are statistically relevant
D Simple forerunner of modern neural networks, without hidden layers
Answer
Marks 1.5
Unit I
Id 33
Question Which of the following are universal approximators?
A Kernel SVM
B Neural Networks
C Boosted Decision Trees
D All of the above
Answer
Marks 1.5
Unit III
Id 34
Question In which of the following applications can we use deep learning to solve the problem?
A Protein structure prediction
B Prediction of chemical reactions
C Detection of exotic particles
D All of these
Answer
Marks 1.5
Unit III
Id 35
Question Which of the following statements is true when you use 1 X 1 convolutions in a CNN?
A It can help in dimensionality reduction
B It can be used for feature pooling
C It suffers less overfiting due to small kernel size
D All of the above
Answer
Marks 1.5
Unit III
Id 36
Question Statement 1: It is possible to train a network well by initializing all the weights as 0.
Statement 2: It is possible to train a network well by initializing biases as 0.
Which of the statements given above is true?
A Statement 1 is true while Statement 2 is false
B Statement 2 is true while statement 1 is false
C Both statements are true
D Both statements are false
Answer
Marks 1.5
Unit III
Id 37
Question The number of nodes in the input layer is 10 and the hidden layer is 5 .The maximum
number of connections from the input layer to the hidden layer are
A 50
B Less than 50
C More than 50
D It is an arbitrary value
Answer
Marks 1.5
Unit I
Id 38
Question The input image has been converted into a matrix of size 28 X 28 and a
kernel/filter of size 7 X 7 with a stride of 1. What will be the size of the convoluted
matrix?
A 22 X 22
B 21 X 21
C 28 X 28
D 7X7
Answer
Marks 1.5
Unit III
Id 39
Question In a simple MLP model with 8 neurons in the input layer, 5 neurons in the hidden
layer and 1 neuron in the output layer. What is the size of the weight matrices
between hidden output layer and input hidden layer?
A [1 X 5] , [5 X 8]
B [8 X 5] , [ 1 X 5]
C [8 X 5] , [5 X 1]
D [5 x 1] , [8 X 5]
Answer
Marks 1.5
Unit I
Id 40
Question Assume a simple MLP model with 3 neurons and inputs= 1,2,3. The weights to
the input neurons are 4,5 and 6 respectively. Assume the activation function is a
linear constant value of 3. What will be the output ?
A 32
B 643
C 96
D 48
Answer
Marks 1.5
Unit I
Id 41
Question Which of following activation function can't be used at output layer to classify an image
A Tanh
B If (x>5,1,0)
C ReLU
D None of the above
Answer
Marks 1.5
Unit III
Id 42
Question Which of the following would have a constant input in each epoch of training a
Deep Learning model?
A Weight between input and hidden layer
B Weight between hidden and output layer
C Biases of all hidden layer neurons
D Activation function of output layer
Answer
Marks 1.5
Unit II
Id 43
Question What value would be in place of question mark?

Here we see a convolutional function being applied to input.

A 3
B 4
C 5
D 6
Answer
Marks 1.5
Unit III
Id 44
Question For a binary classification problem, which of the following architecture would
you choose?

A 1
B 2
C Any one of these
D None of these
Answer
Marks 1.5
Unit I
Id 45
Question Which of the following statement is true regrading dropout?
1: Dropout gives a way to approximate by combining many different architectures
2: Dropout demands high learning rates
3: Dropout can help preventing overfitting
A Both 1 and 2
B Both 1 and 3
C Both 2 and 3
D All 1,2 and 3
Answer
Marks 1.5
Unit III
Id 46
Question What steps can we take to prevent overfitting in a Neural Network?
A Weight Sharing
B Early Stopping
C Dropout
D All of the above
Answer
Marks 1.5
Unit II
Id 47
Question You are building a binary classifier for recognizing cucumbers (y=1) vs. watermelons
(y=0). Which one of these activation functions would you recommend using for the
output layer?

A ReLU
B LeakyReLU
C sigmoid
D Parametric ReLU
Answer
Marks 1.5
Unit II
Id 48
Question Suppose you have built a neural network. You decide to initialize the weights and
biases to be zero. Which of the following statements is true?

A Each neuron in the first hidden layer will perform the same computation. So even after
multiple iterations of gradient descent each neuron in the layer will be computing the
same thing as other neurons.
B Each neuron in the first hidden layer will perform the same computation in the first
iteration. But after one iteration of gradient descent they will learn to compute different
things because we have “broken symmetry”.

C Each neuron in the first hidden layer will compute the same thing, but neurons in
different layers will compute different things, thus we have accomplished “symmetry
breaking” as described in lecture.

D The first hidden layer’s neurons will perform different computations from each other
even in the first iteration; their parameters will thus keep evolving in their own way.

Answer
Marks 1.5
Unit I
Id 49
Question What are the issues on which biological networks proves to be superior than Al
networks?
A robustness and fault tolerance
B flexibility
C collective computation
D all of the mentioned
Answer
Marks 1.5
Unit I
Id 50
Question GPU stands for
A Graphics Processing Unit
B Gradient Processing Unit
C General Processing Unit
D Good Processing Unit.
Answer
Marks 1.5
Unit II
Id 51
Question -------------is a Neural Nets way of classifying inputs.
A Learning
B Forward Propagation
C Activation
D Classification
Answer
Marks 1.5
Unit I
Id 52
Question Name the component of a Neural Network where the true value of the input is not
observed.
A Hidden Layer
B Gradient Descent
C Activation Function
D Output Layer
Answer
Marks 1.5
Unit I
Id 53
Question ________________ works best for Image Data.
A Random Forest
B Convolution Networks
C single Layer Perceptrons
D AutoEncoders
Answer
Marks 1.5
Unit II
Id 54
Question _______________ is a recommended Model for Pattern Recognition in Unlabeled Data.
A Random Forest
B Convolution Networks
C single Layer Perceptrons
D AutoEncoders
Answer
Marks 1.5
Unit II
Id 55
Question Process of improving the accuracy of a Neural Network is called.......................
A Training
B Random Walk
C Cross Validation
D Forward Propagation
Answer
Marks 1.5
Unit II
Id 56
Question Support Vector Machines, Naive Bayes and Logistic Regression are used for solving
problems.
A Regression Time Series
B Classification
C Clustering
D Image processing
Answer
Marks 1.5
Unit I
Id 57
Question What does LSTM stand for?
A Long Short Threshold Memory
B Least Square Time Mean
C Long Short Term Memory
D Least Squares Term Memory
Answer
Marks 1.5
Unit III
Id 58
Question What is the method to overcome the Decay of Information through time in RNN known
as?
A Gating
B Back Propagation
C Gradient Descent
D Activation
Answer
Marks 1.5
Unit III
Id 59
Question What is the best Neural Network Model for Temporal Data?
A Multi Layer Perceptrons
B Temporal Neural Networks
C Convolution Neural Networks
D Recurrent Neural Network
Answer
Marks 1.5
Unit III
Id 60
Question ReLU stands for …........
A Rectified Linear Unit
B Rectified Lagrangian Unit
C Regressive Linear Unit
D Regressive Lagrangian Unit
Answer
Marks 1.5
Unit II
Id 61
Question Why is the Pooling Layer used in a Convolution Neural Network? They are of no use in
CNN.
A Image Sensing
B Object Recognition
C Dimension Reduction
D Pattern Recognition
Answer
Marks 1.5
Unit II
Id 62
Question What are the two layers of a Restricted Boltzmann Machine called?
A Input and Output Layers
B Recurrent and Convolution Layers
C Activation and Threshold Layers
D Hidden and Visible Layers
Answer
Marks 1.5
Unit II
Id 63
Question The measure of Difference between two probability distributions is know as ….....
A Probability Difference
B Cost
C KL Divergence
D Error
Answer
Marks 1.5
Unit III
Id 64
Question A_______________matches or surpasses the output of an individual neuron to a visual
stimull.
A Convolution
B Gradient
C Cost
D Max Pooling
Answer
Marks 1.5
Unit I
Id 65
Question The rate at which cost changes with respect to weight or bias is called_____
A Loss
B Rate of Change
C Gradient
D Derivative
Answer
Marks 1.5
Unit I
Id 66
Question Autoencoders are trained using...........
A They do not require Training
B Back Propagation
C Reconstruction
D Feed Forward
Answer
Marks 1.5
Unit II
Id 67
Question De-noising and Contractive are examples of ____________
A Recurrent Neural Networks
B Convolution Neural Networks
C Autoencoders
D Shallow Neural Networks
Answer
Marks 1.5
Unit II
Id 68
Question What is the purpose of the Gradient Descent algorithm?
A To normalize the inputs
B To minimize the weights and bias
C To minimize the loss function
D To prevent model from overfitting
Answer
Marks 1.5
Unit III
Id 69
Question Which of the following pre-trained models in Keras can be fine-tuned for image
classification?
A VGG16
B MobileNet
C InceptionResNetV2
D All of the above
Answer
Marks 1.5
Unit III
Id 70
Question Least Squares Estimation minimizes:
A summation of squares of errors
B summation of errors
C summation of absolute values of errors
D All
Answer
Marks 1.5
Unit III
Id 71
Question Parameter Estimation problem is about:
A Identifying Input Parameters
B Identifying Output Parameters
C Identifying Model Parameters
D All
Answer
Marks 1.5
Unit III
Id 72
Question 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:
A 238
B 76
C 119
D 123
Answer
Marks 1.5
Unit I
Id 73
Question In the case of an algebraic model for a straight line, if a value for the x variable is
specified, then …..............
A the exact value of the response variable can be computed
B the computed response to the independent value will always give a minimal residual
C the computed value of y will always be the best estimate of the mean response
D none of these alternatives is correct.
Answer
Marks 1.5
Unit I
Id 74
Question How can states of units be updated in hopfield model?
A synchronously
B asynchronously
C synchronously and asynchronously
D none of the mentioned
Answer
Marks 1.5
Unit II
Id 75
Question What is synchronous update in hopfield model?
A all units are updated simultaneously
B a unit is selected at random and its new state is computed
C a predefined unit is selected and its new state is computed
D none of the mentioned
Answer
Marks 1.5
Unit II
Id 76
Question What is asynchronous update in hopfield model?
A all units are updated simultaneously
B a unit is selected at random and its new state is computed
C a predefined unit is selected and its new state is computed
D none of the mentioned
Answer
Marks 1.5
Unit II
Id 77
Question What is gradient descent?
A method to find the absolute minimum of a function
B method to find the absolute maximum of a function
C maximum or minimum, depends on the situation
D none of the mentioned
Answer
Marks 1.5
Unit 1.5
Id III
Id 78
Question method to find the absolute minimum of a function
A all units are updated simultaneously
B a unit is selected at random and its new state is computed
C a predefined unit is selected and its new state is computed
D none of the mentioned
Answer
Marks 1.5
Unit III
Id 79
Question If pattern is to be stored, then what does stable state should have updated value of?
A current sate
B next state
C both current and next state
D none of the mentioned
Answer
Marks 1.5
Unit I
Id 80
Question How can error in recall due to false minima be reduced?
A deterministic update for states
B stochastic update for states
C not possible
D none of the mentioned
Answer
Marks 1.5
Unit I
Id 81
Question Pattern storage problem which cannot be represented by a feedback network of given
size can be called as?
A easy problems
B hard problems
C no such problem exist
D none of the mentioned
Answer
Marks 1.5
Unit III
Id 82
Question For what purpose Feedback neural networks are primarily used?
A classification
B feature mapping
C pattern mapping
D none of the mentioned
Answer
Marks 1.5
Unit II
Id 83
Question Presence of false minima will have what effect on probability of error in recall?
A directly
B inversely
C no effect
D directly or inversely
Answer
Marks 1.5
Unit II
Id 84
Question How is effect false minima reduced
A deterministic update of weights
B stochastic update of weights
C deterministic or stochastic update of weights
D none of the mentioned
Answer
Marks 1.5
Unit II
Id 85
Question Match the following knowledge representation techniques with their applications:
List – I
(a) Frames
(b) Conceptual dependencies
(c) Associative networks
(d) Scripts

List – II
(i) Pictorial representation of objects, their attributes and relationships
(ii) To describe real world stereotype events
(iii) Record like structures for grouping closely related knowledge
(iv) Structures and primitives to represent sentences
code:
abcd
A (iii) (iv) (i) (ii)
B (iii) (iv) (ii) (i)
C (iv) (iii) (i) (ii)
D (iv) (iii) (ii) (i)
Answer
Marks 1.5
Unit II
Id 86
Question Slots and facets are used in
A Semantic Networks
B Frames
C Rules
D All of these
Answer
Marks 1.5
Unit III
Id 87
Question Consider the following statements:

(a) If primal (dual) problem has a finite optimal solution, then its dual (primal) problem has a
finite optimal solution.
(b) If primal (dual) problem has an unbounded optimum solution, then its dual (primal) has no
feasible solution at all.
(c) Both primal and dual problems may be infeasible.
Which of the following is correct?
A (a) and (b) only
B (a) and (c) only
C (b) and (c) only
D (a), (b) and (c)
Answer
Marks 1.5
Unit III
Id 88
Question Consider the following statements :

(a) Assignment problem can be used to minimize the cost.

(b) Assignment problem is a special case of transportation problem.
(c) Assignment problem requires that only one activity be assigned to each resource.

Which of the following options is correct?

A (a) and (b) only
B (a) and (c) only
C (b) and (c) only
D (a), (b) and (c)
Answer
Marks 1.5
Unit III
Id 89
Question What is purpose of Axon in Biological Neural network?
A receptors
B transmitter
C transmission
D none of the mentioned
Answer
Marks 1.5
Unit I
Id 90
Question A neural network model is said to be inspired from the human brain. The neural
network consists of many neurons, each neuron takes an input, processes it and
gives an output.
A A neuron has a single input and a single output only
B A neuron has multiple inputs but a single output only
C A neuron has a single input but multiple outputs
D All of the above statements are valid
Answer
Marks 1.5
Unit I
Id 91
Question Consider the following statements :
(a) If you increase the number of hidden layers in a Multi Layer Perceptron, the
classification error of test data always decreases.
(b) A neuron has multiple inputs and multiple outputs .
True or False
A a)False b) True
B a)False b) False
C a)True b) True
D a)True b) False
Answer
Marks 1.5
Unit II
Id 92
Question Consider the following statements :
(a) Convolutional Neural Networks can perform various types of transformation
(rotations or scaling) in an input”.
(b) Bagging perform similar operations as dropout in a neural network .
True or False
A a)False b) True
B a)False b) False
C a)True b) True
D a)True b) False
Answer
Marks 1.5
Unit II
Id 93
Question Consider the following statements :
(a) Convolutional Neural Networks can perform various types of transformation
(rotations or scaling) in an input”.
(b) a neural network model the function (y=1/x)
True or False
A a)False b) True
B a)False b) False
C a)True b) True
D a)True b) False
Answer
Marks 1.5
Unit II
Id 94
Question The network shown in Figure is trained to recognize the characters H and T as
shown below:

D Could be A or B depending on the weights of neural network

Answer
Marks 1.5
Unit II
Id 95
Question Consider the following statements :
(a) Suppose a convolutional neural network is trained on ImageNet dataset
(Object recognition dataset). This trained model is then given a completely white
image as an input.The output probabilities for this input would be equal for all
classes. True or False?
(b) When pooling layer is added in a convolutional neural network, translation in-
variance is preserved. True or False
A a)False b) True
B a)False b) False
C a)True b) True
D a)True b) False
Answer
Marks 1.5
Unit II
Id 96
Question Which tool is best suited for solving
Deep Learning problems
A R
B Sk-learn
C Excel
D TensorFlow
Answer
Marks 1.5
Unit III
Id 97
Question Kears is a deep learning framework on which tool
A R
B Sk-learn
C Excel
D TensorFlow
Answer
Marks 1.5
Unit II
Id 98
Question Which tool is a deep learning wrapper on TensorFlow
A Python
B Keras
C PyTourch
D Azure
Answer
Marks 1.5
Unit II
Id 99
Question How deep learning models are built on Keras
A by using sequential models
B by using feed_dict
C by creating place holders and computational
graphs
D by creating data frames
Answer
Marks 1.5
Unit II
Id 100
Question Can we use GPU for faster computations in TensorFlow
A No, not possible
B Possible only on cloud
C Possible only with small datasets
D Yes, possible
Answer
Marks 1.5
Unit II
Id 101
Question Boltzman learning is a?
A fast process
B steady process
C slow process
D none of the mentioned
Answer
Marks 1.5
Unit I
Id 102
Question In boltzman learning which algorithm can be used to arrive at equilibrium?
A hopfield
B mean field
C hebb
D none of the mentioned
Answer
Marks 1.5
Unit I
Id 103
Question What happens when we use mean field approximation with boltzman learning?
A it slows down
B it get speeded up
C nothing happens
D may speedup or speed down
Answer
Marks 1.5
Unit I
Id 104
Question For practical implementation what type of approximation is used on boltzman law?
A max field approximation
B min field approximation
C hopfield approximation
D none of the mentioned
Answer
Marks 1.5
Unit I
Id 105
Question What should be the aim of training procedure in boltzman machine of feedback networks?
A to capture inputs
B to feedback the captured outputs
C to capture the behaviour of system
D none of the mentioned
Answer
Marks 1.5
Unit I
Id 106
Question What consist of boltzman machine?
A fully connected network with both hidden and visible units
B asynchronous operation
C stochastic update
D all of the mentioned
Answer
Marks 1.5
Unit I
Id 107
Question For what purpose is pattern environment useful?
A determining structure
B determining desired outputs
C determining future inputs
D none of the mentioned
Answer
Marks 1.5
Unit I
Id 108
Question What is pattern environment?
A probability of desired patterns
B probability of given patterns
C behaviour of system
D none of the mentioned
Answer
Marks 1.5
Unit I
Id 109
Question Probability of error in recall of stored patterns can be reduced if?
A patterns are stored appropriately
B inputs are captured appropriately
C weights are chosen appropriately
D none of the mentioned
Answer
Marks 1.5
Unit I
Id 110
Question Why does change in temperature doesn’t effect stochastic update?
A shape landscape depends on the network and its weights which varies accordingly and
compensates the effect
B shape landscape depends on the network and its weights which is fixed
C shape landscape depends on the network, its weights and the output function which varies
accordingly and compensates the effect
D shape landscape depends on the network, its weights and the output function which is
fixed
Answer
Marks 1.5
Unit I
Id 111
Question As temperature increase, what happens to stochastic update?
A increase in update
B decrease in update
C no change
D none of the mentioned
Answer
Marks 1.5
Unit II
Id 112
Question In the case of an algebraic model for a straight line, if a value for the x variable is
specified, then
A the exact value of the response variable can be computed
B the computed response to the independent value will always give a minimal residual
C the computed value of y will always be the best estimate of the mean response
D none of these alternatives is correct.
Answer
Marks 1.5
Unit I
Id 113
Question Identify the activation function from the given plot

A Binary Step Function

B ReLU
C Signmoid
D Tanh
Answer
Marks 1.5
Unit II
Id 114
Question Identify the activation function from the given plot

A Binary Step Function

B ReLU
C Signmoid
D Tanh
Answer
Marks 1.5
Unit II
Id 115
Question Identify the activation function
f(x)=max(0,x)
A Binary Step Function
B ReLU
C Signmoid
D Tanh
Answer
Marks 1.5
Unit II
Id 116
Question Identify the activation function

A Binary Step Function

B ReLU
C Signmoid
D Tanh
Answer
Marks 1.5
Unit II
Id 117
Question Identify the activation function

A Binary Step Function

B ReLU
C Signmoid
D Tanh
Answer
Marks 1.5
Unit II
Id 118
Question Identify the activation function

A Binary Step Function

B ReLU
C Leaky ReLU
D Tanh
Answer
Marks 1.5
Unit II
Id 119
Question Identify the activation function given in the plot blow.

A Binary Step Function

B ReLU
C Swish
D Tanh
Answer
Marks 1.5
Unit II
Id 120
Question Identify the activation function given in the plot blow.

A Binary Step Function

B LeakyReLU
C Swish
D Tanh
Answer
Marks 1.5
Unit II
Id 121
Question Identify the activation function

A Binary Step Function

B ReLU
C Leaky ReLU
D Swish
Answer
Marks 1.5
Unit II
Id 122
Question Identify the activation function from the given plot

A Binary Step Function

B ReLU
C Signmoid
D Tanh
Answer
Marks 1.5
Unit II
Id 123
Question Statement 1: Recurrent Neural Networks are best suited for Text Processing.
Statement 2: Recurrent Networks work best for Speech Recognition.
Which of the statements given above is true?
A Statement 1 is true while Statement 2 is false
B Statement 2 is true while statement 1 is false
C Both statements are true
D Both statements are false
Answer
Marks 1.5
Unit I
Id 124
Question Statement 1: Neural Networks Algorithms are inspired from the structure and functioning of
the Human Biological Neuron.
Statement 2: In a Neural Network, all the edges and nodes have the same Weight and Bias
values.
Which of the statements given above is true?
A Statement 1 is true while Statement 2 is false
B Statement 2 is true while statement 1 is false
C Both statements are true
D Both statements are false
Answer
Marks 1.5
Unit I
Id 125
Question Data Collected from Survey results is an example of ___________________.
A Data
B Information
C Structured Data
D Unstructured Data
Answer
Marks 1.5
Unit I
Id 126
Question Statement 1: A Shallow Neural Network has only one hidden layer between Input and Output
layers.
Statement 2: All the Visible Layers in a Restricted Boltzmannn Machine are connected to
each other.
Which of the statements given above is true?
A Statement 1 is true while Statement 2 is false
B Statement 2 is true while statement 1 is false
C Both statements are true
D Both statements are false
Answer
Marks 1.5
Unit I
Id 127
Question Statement 1: All the neurons in a convolution layer have different Weights and Biases.
Statement 2: Recurrent Network can input Sequence of Data Points and Produce a
Sequence of Output.
Which of the statements given above is true?
A Statement 1 is true while Statement 2 is false
B Statement 2 is true while statement 1 is false
C Both statements are true
D Both statements are false
Answer
Marks 1.5
Unit II
Id 128
Question Statement 1: A Deep Belief Network is a stack of Restricted Boltzmann Machines.
Statement 2: Restricted Boltzmann Machine expects the data to be labeled for Training.
Which of the statements given above is true?
A Statement 1 is true while Statement 2 is false
B Statement 2 is true while statement 1 is false
C Both statements are true
D Both statements are false
Answer
Marks 1.5
Unit I
Id 129
Question Statement 1: Autoencoders cannot be used for Dimensionality Reduction.
Statement 2: Restricted Boltzmann Machine expects the data to be labeled for Training.
Which of the statements given above is true?
A Statement 1 is true while Statement 2 is false
B Statement 2 is true while statement 1 is false
C Both statements are true
D Both statements are false
Answer
Marks 1.5
Unit II
Id 130
Question What is hebb’s rule of learning ?
A the system learns from its past mistakes
B the system recalls previous reference inputs & respective ideal outputs
C the strength of neural connection get modified accordingly
D none of the mentioned
Answer
Marks 1.5
Unit I
Id 131
Question Which one of these plots represents a ReLU activation function?

Answer
Marks 1.5
Unit II
Id 132
Question What does a neuron compute?

A A neuron computes an activation function followed by a linear function (z = Wx + b)

B A neuron computes the mean of all features before applying the output to an activation
function

C A neuron computes a function g that scales the input x linearly (Wx + b)

D A neuron computes a linear function (z = Wx + b) followed by an activation function

Answer
Marks 1.5
Unit I
Id 133
Question Statement 1: Boltzman law is practical for implementation.
Statement 2: False minima can be reduced by deterministic updates
Which of the statements given above is true?
A Statement 1 is true while Statement 2 is false
B Statement 2 is true while statement 1 is false
C Both statements are true
D Both statements are false
Answer
Marks 1.5
Unit I
Id 134
Question Which of the following layers is NOT a part of CNN?

A Convolutional Layer
B Pooling Layer
C Code Layer
D Fully connected Layer
Answer
Marks 1.5
Unit II
Id 135
Question In which of the following situations, you should NOT prefer Keras over TensorFlow?

A When you want to quickly build prototype using neural networks

B When you want to implement simple neural networks in your initial learning phase
C When you are doing a critical and intensive research on any field
D When you want to create simple tutorials for your students and subscribers
Answer
Marks 1.5
Unit II
Id 136
Question Which of the following is an application of RNN?

A Natural Language Processing

B Audio and Video analysis
C Stock market prediction
D All of the above
Answer
Marks 1.5
Unit II
Id 137
Question Which of the following types of RNN is best suited for machine translation?

A One to One
B One to Many
C Many to One
D Many to Many
Answer
Marks 1.5
Unit II
Id 138
Question Which of the following is TRUE about TensorFlow?

A Tensor with rank zero is also called scaler

B Tensor with rank one is also called matrix
C Tensor with rank two is also called vector
D All of the above
Answer
Marks 1.5
Unit II
Id 139
Question Gradient Descent computes derivative of loss function w.r.t:

A input
B activation value
C weight
D bias
Answer
Marks 1.5
Unit I
Id 140
Question Which of the following is FALSE about step activation function?

A It is a differentiable activation function

B Derivative of a step function is always zero
C It is non-linear in nature
D None of the above
Answer
Marks 1.5
Unit I
Id 141
Question Which of the following is NOT an application of Autoencoders?

A Image coloring
B Image captioning
C Anomalies and outliers detection
D Dimensionality reduction
Answer
Marks 1.5
Unit II
Id 142
Question Which of the following is TRUE about Leaking ReLU?

A Leaky ReLU is the most common method to alleviate a dying ReLU

B Leaking ReLU adds a slight slope in the negative range to prevent dying ReLU issue
C Parametric ReLU (PReLU) is a type of leaky ReLU that uses parameterized slope value
D All of the above
Answer
Marks 1.5
Unit II
Id 143
Question ReLU activation function outputs zero when:

A Input is zero
B Input is less than or equal to zero
C Input is greater than or equal to zero
D Input is zero or one
Answer
Marks 1.5
Unit II
Id 144
Question Which of the following parameters is not required while compiling a model in Keras?

A activation function
B loss
C metrics
D optimizer
Answer
Marks 1.5
Unit II
Id 145
Question Which of the following is NOT an application of Restricted Boltzmann Machines RBM ?

A Dimensionality reduction
B Image captioning
C Feature learning and topic modelling
D Collaborative filtering for recommender systems
Answer
Marks 1.5
Unit I
Id 146
Question CNN is best suited for:

A Image Classification
B Natural Language Processing
C Image Captioning
D All of the above
Answer
Marks 1.5
Unit I
Id 147
Question Which of the following MUST be initialized in TensorFlow?

A Variables
B Placeholders
C Constants
D Sessions
Answer
Marks 1.5
Unit I
Id 148
Question Which of the following frameworks can be used as backend in Keras?

A Swift
B Gluon
C TensorFlow
D MATLAB
Answer
Marks 1.5
Unit I
Id 149
Question Which of the following neural networks uses supervised learning?
(A) Multilayer perceptron
(B) Self organizing feature map
(C) Hopfield network
A (A) only
B (B) only
C (A) and (B) only
D (A) and (C) only
Answer
Marks 1.5
Unit I
Id 150
Question Which of the following statement is true regrading dropout?
1: Dropout gives a way to approximate by combining many different architectures
2: Dropout demands high learning rates
3: Dropout can help preventing overfitting
A Both 1 and 2
B Both 1 and 3
C Both 2 and 3
D All 1,2 and 3
Answer
Marks 1.5
Unit II