1.What is the primary purpose of a Recurrent Neural Network (RNN)?

Text generation
2.Which layer type is typically used to capture sequential dependencies in an RNN?
Hidden layer
3.What is the advantage of using recurrent layers in an RNN?
Answer: They can capture temporal dependencies in the input data
4.What is the purpose of the hidden state in an RNN?
Answer: To store the information from the previous time step
5.Which activation function is commonly used in the recurrent layers of an RNN?
Answer: Tanh
6.What is the purpose of the time step parameter in an RNN?
Answer: To specify the length of the input sequence
7.Which layer type is commonly used to initialize the hidden state in an RNN?
Answer: Hidden layer
8.What is the purpose of the bidirectional RNN architecture?
Answer: To handle sequential data in both forward and backward directions
9.Which layer type is responsible for making final predictions in an RNN?
Answer: Output layer
10.What is the purpose of the recurrent connection in an RNN?
Answer: To propagate the hidden state across different time steps
11.Which layer type is commonly used in RNNs for sequence-to-sequence tasks?
Answer: Attention layer
12.What is the purpose of the backpropagation through time (BPTT) algorithm in RNN training?
Answer: To compute the gradients and update the network's parameters
13.Which layer type is commonly used in RNNs to handle variable-length inputs?
Answer: Hidden layer
14.Which layer type is responsible for handling the output at each time step in an RNN?
Answer: Output layer
15.What is the purpose of the sequence-to-vector architecture in an RNN?
Answer: To process an input sequence and produce a fixed-length representation
16.What is the purpose of the peephole connections in a Long Short-Term Memory (LSTM) network?
Answer: To allow the cell state to influence the gating mechanisms
17.What is the purpose of the cell state in an LSTM network?
Answer: To store long-term dependencies in the input sequence
18.What is the purpose of the input gate in an LSTM network?
Answer: To control the flow of information from the current input
19.Which layer type is responsible for handling variable-length inputs and outputs in an RNN?
Answer: Hidden layer
20.What is the purpose of the output gate in an LSTM network?
Answer: To control the flow of information to the current output
21.What is the vanishing gradient problem in RNNs?
Answer: Gradients become very small during backpropagation
22.What is the exploding gradient problem in RNNs?
Answer: Gradients become very large during backpropagation
23.Which technique is used to address the exploding gradient problem?
Answer: Gradient clipping
24.What is the function of the forget gate in an LSTM network?
Answer: To decide what information to discard from the cell state
25.What is a major benefit of using LSTM over vanilla RNNs?
Answer: LSTM can capture long-term dependencies better
26.What distinguishes a GRU from an LSTM?
Answer: GRUs combine the forget and input gates into a single update gate
27.Which is more computationally efficient: GRU or LSTM?
Answer: GRU
28.What does the term “many-to-one” architecture in RNN refer to?
Answer: Sequence input, single output
29.What does the term “many-to-many” architecture in RNN refer to?
Answer: Sequence input, sequence output
30.What is the role of an embedding layer in sequence modeling?
Answer: To convert categorical tokens into dense vectors
31. Which type of neural network architecture is primarily used for image classification tasks?
Answer:Convolutional Neural Networks (CNNs)
32. Which of the following neural network architectures is considered one of the earliest
successful CNN models for handwritten digit recognition?
Answer:LeNet
33.What is the primary innovation introduced by AlexNet in the field of deep learning? a)The
use of max-pooling layer
Answer:The use of deep convolutional neural networks
34.Which of the following architectures is known for its simplicity and effectiveness in image
classification tasks, consisting mainly of 3×3 convolutional layers?
Answer:VGGNet
35. What distinguishes GoogLeNet from other CNN architectures?
a) It introduces Inception modules
36.Which neural network architecture is known for its extremely deep structure, consisting of
hundreds of layers?
b) ResNet
37. What characterizes recent trends in deep learning architectures?
Answer:Increasing model complexity b) Emphasis on interpretability
c) Focus on model efficiency d) All of the above
38. Which of the following is not a deep learning architecture specifically designed for image-related
tasks?
Answer:LSTM
39. Which activation function is often preferred in the output layer of a neural network for binary
classification tasks?
Answer:Sigmoid
40.Which of the following is not a characteristic of ReLU activation function?
Answer: Outputs range from 0 to 1
41. Which neural network architecture is commonly used for processing sequential data such as text
and speech?
a) Long Short-Term Memory Networks
42. Which of the following is a representation learning algorithm?
A) Neural network
43. Increase in size of a convolutional kernel would necessarily increase the performance of a
convolutional neural network.
Answer:FALSE
44. Deep learning can be applied to which of the following NLP tasks?
A) Machine translation B) Sentiment analysis
C) Question Answering system D) All of the above
45. What is the purpose of data augmentation in deep learning?
A:Expanding the training dataset by applying various transformations to the existing data
46. Which deep learning model is used for generating realistic images from random noise?
A:GAN (Generative Adversarial Network)
47. How many layers are there in Autoencoder?
b) 3
48. Select the correct option.
A. Supervised learning methods include autoencoders.
B. The output and input of the autoencoder are identical.
Answer:Both the statements are FALSE
49.Autoencoders are capable of learning nonlinear manifolds (a continuous, non-intersecting
surface.)
Answer:TRU
50.Autoencoders are trained using
c) Back Propagation