CS-7643..Verified Multiple Choice and
Conceptual Actual Frequently Tested Exam
Questions With Reviewed 100% Correct
Detailed Answers
Guaranteed Pass!!Current Update!!
Q1. What is the primary characteristic of a Recurrent Neural Network (RNN)?
A. It can only process fixed-size vectors.
B. It has hidden states that allow sequential processing.
C. It ignores temporal dependencies in data.
D. It cannot model variable-length inputs.
Answer: B
Q2. In RNNs, the same parameters are reused at each time step. This property is
referred to as:
A. Parameter explosion
B. Backpropagation through time
C. Recurrence
D. Memory gating
Answer: C
Q3. What problem do LSTMs solve that vanilla RNNs struggle with?
A. Overfitting
B. Vanishing/exploding gradients
C. Model interpretability
D. Excessive parameterization
Answer: B
, Q4. How does the LSTM cell update rule differ from the RNN update rule?
A. LSTM uses multiplicative updates, RNN uses additive.
B. LSTM uses additive updates, RNN uses multiplicative.
C. Both use multiplicative updates.
D. Both use additive updates.
Answer: B
Q5. Which gate in an LSTM decides how much past memory to retain?
A. Input gate
B. Forget gate
C. Output gate
D. Candidate gate
Answer: B
Q6. Why do gradients in vanilla RNNs vanish or explode?
A. Because of linear activation functions.
B. Because of recurrent multiplication of weight matrices.
C. Because inputs are not normalized.
D. Because they use convolution layers.
Answer: B
Q7. If weight magnitude < 1 in an RNN, the gradient will:
A. Explode
B. Vanish
C. Remain constant
D. Oscillate
Answer: B
Conceptual Actual Frequently Tested Exam
Questions With Reviewed 100% Correct
Detailed Answers
Guaranteed Pass!!Current Update!!
Q1. What is the primary characteristic of a Recurrent Neural Network (RNN)?
A. It can only process fixed-size vectors.
B. It has hidden states that allow sequential processing.
C. It ignores temporal dependencies in data.
D. It cannot model variable-length inputs.
Answer: B
Q2. In RNNs, the same parameters are reused at each time step. This property is
referred to as:
A. Parameter explosion
B. Backpropagation through time
C. Recurrence
D. Memory gating
Answer: C
Q3. What problem do LSTMs solve that vanilla RNNs struggle with?
A. Overfitting
B. Vanishing/exploding gradients
C. Model interpretability
D. Excessive parameterization
Answer: B
, Q4. How does the LSTM cell update rule differ from the RNN update rule?
A. LSTM uses multiplicative updates, RNN uses additive.
B. LSTM uses additive updates, RNN uses multiplicative.
C. Both use multiplicative updates.
D. Both use additive updates.
Answer: B
Q5. Which gate in an LSTM decides how much past memory to retain?
A. Input gate
B. Forget gate
C. Output gate
D. Candidate gate
Answer: B
Q6. Why do gradients in vanilla RNNs vanish or explode?
A. Because of linear activation functions.
B. Because of recurrent multiplication of weight matrices.
C. Because inputs are not normalized.
D. Because they use convolution layers.
Answer: B
Q7. If weight magnitude < 1 in an RNN, the gradient will:
A. Explode
B. Vanish
C. Remain constant
D. Oscillate
Answer: B
