Sequence Models Interview Question- – 2
Q.1 The main drawback of basic RNN is __________.
A. Exploding gradient problem
B. Vanishing gradient problem
C. Overfitting
D. Large training parameters
Ans : Vanishing gradient problem
Q.2 Feedforward network cannot be used on sequence data because ____________.
A. You cannot pass hidden states to next input
B. Dependencies on previous inputs will not be considered
C. Of the varying length of input data
D. All the options
Ans : All the options
Q.3 Which of the following are the applications of RNN?
A. Time series prediction
B. Machine translation
C. Intent extraction
D. All the options
Ans : Intent extraction
Q.4 Error function in RNN is computed as _________________.
A. Summation of errors at individual time step
B. Sum of errors at the first few time steps
C. Weighted sum of error values at each time step
D. Error value only at the final time step
Ans : Summation of errors at individual time step
Q.5 Back propagation through time is computed as ______________.
A. Summation of derivatives of parameters on error function over the first few time steps
B. Summation of derivatives of parameters on error function across each time step
C. Product of derivatives of parameters on error function across each time step
D. Derivatives of parameters on error function at the final time step
Ans : Summation of derivatives of parameters on error function across each time step
Q.6 Which of the following holds the memory in RNNs?
A. Error function
B. Hidden state
C. Output state
D. Input state
Ans : Error function
Q.7 Which of the following is not an example of sequential data?
A. Time series data
B. Text data
C. Speech data
D. Image data
Ans : Image data
Q.8 The basic RNN fails when _______________________.
A. Input sequence is too small
B. Input is of fixed length
C. Input sequence is very long
D. There is temporal dependencies in the data
Ans : Input sequence is very long
Q.9 In which of the following, the vanishing gradient problem is profound? —
A. GRU
B. Basic RNN
C. LSTM
Ans : Basic RNN
Q.10 In the case of LSTM, which of the following is used to compute the output of a cell?
A. Hidden state
B. Cell state
C. Input state
D. Reset state
Ans : Hidden state
Q.11 Which of the following gates (in LSTM) decides on keeping relevant features from the current input?
A. Input gate
B. Forget gate
C. Output gate
D. Update gate
Ans : Update gate
Q.12 Which of the following is an additional feature of LSTM when compared to basic RNN?
A. Hidden state
B. Cell state
C. Output state
D. Input State
Ans : Cell state
Q.13 Which of the following is not a feature of LSTM?
A. Hidden state
B. Reset gate
C. Update gate
D. Cell state
Ans : Hidden state
Q.14 GRU has more parameters to learn when compared to LSTM.
A. True
B. False
Ans : False
Q.15 LSTM has a large number of parameters to learn when compared to basic RNN.
A. True
B. False
Ans : True
Q.16 Which of the following is not the feature of GRU?
A. Hidden state
B. Reset gate
C. Update gate
D. Cell state
Ans : Cell state
Q.17 The reason feed forward network cannot be used on sequence data is _____________.
A. Dependencies on previous inputs will not be considered
B. Cannot pass hidden states to next input
C. Varying length of input data
D. All the options
Ans : Varying length of input data
Q.18 Which of the following is the component of GRU?
A. Forget gate
B. Reset gate
C. Input gate
D. Cell state
Ans : Reset gate
Q.19 In the case of LSTM, which of the following is used to compute the output of a cell?
A. Cell state
B. Hidden state
C. Input state
D. Reset state
Ans : Cell state
Q.20 Which of the following gates in LSTM decides on eliminating irrelevant features from previous information?
A. Input Gate
B. Update gate
C. Forget gate
D. Reset gate
Ans : Forget gate
Q.21 Which of the following gates in LSTM decides on keeping relevant features from the current input?
A. Output gate
B. Forget gate
C. Update gate
D. Input gate
Ans : Input gate
Q.22 Which of the following considers the information from previous timestep, current ones, as well as future timesteps?
A. GRU
B. LSTM
C. RNN
D. Bidirectional RNN
Ans : Bidirectional RNN
Q.23 What is meant by sequence data?
A. The data with a temporal relationship between them
B. The data with no relationship with other data
C. The data with a spacial relationship between them
D. Data arranged in a sequence
Ans : The data with a temporal relationship between them
Q.24 What is the entity that is carried from one timestep to next in RNN?
A. Concatenated inputs from previous timesteps
B. Final output
C. Previous timestep input
D. Hidden state
Ans : Hidden state
Q.25 De-noising and Contractive are examples of ______________
A. Shallow Neural Networks
B. Autoencoders
C. Convolution Neural Networks
D. Recurrent Neural Networks
Ans : Autoencoders
Q.26 How do RNTS interpret words?
A. One Hot Encoding
B. Lower Case Versions
C. Word Frequencies
D. Vector Representations
Ans : Vector Representations
Q.27 Autoencoders are trained using _______________.
A. Feed Forward
B. Reconstruction
C. Back Propagation
D. They do not require Training
Ans : Back Propagation
Q.28 The rate at which cost changes with respect to weight or bias is called __________________.
A. Derivative
B. Gradient
C. Loss
D. Rate of Change
Ans : Gradient
Q.29 A _________________ matches or surpasses the output of an individual neuron to a visual stimuli.
A. Max Pooling
B. Gradient
C. Cost
D. Convolution
Ans : Convolution
Q.30 Autoencoders cannot be used for Dimensionality Reduction.
A. True
B. False
Ans : False
Sequence Models Interview Question- – 2
