Which of the following descriptions about the Recurrent Neural Network (RNN) is correct?

Question

Options

• ACan be used to process sequence data.
• BCannot process variable length sequence data.
• CUnlike convolutional neural networks. parameters of RNN cannot be shared
• DThe units above the hidden layer are not associated With each other.

• A
  76% (39)
• B
  14% (7)
• C
  4% (2)
• D
  6% (3)

Explanation

RNNs are specifically designed to handle sequential data by maintaining a hidden state that captures information from previous time steps, making them ideal for tasks like language modeling, speech recognition, and time-series prediction - so A is correct.

B is wrong: RNNs can process variable-length sequences; the recurrent loop simply runs for as many time steps as the input sequence requires, with no fixed-length constraint.

C is wrong: RNNs do share parameters - the same weight matrices (W, U, b) are applied at every time step, which is one of their defining characteristics and why they're efficient for sequential data.

D is wrong: Hidden units across time steps are directly connected through the recurrent link; that recurrent connection passing hidden state forward is precisely what makes RNNs "recurrent."

Memory tip: Think of RNN as "remembers the past" - the R stands for Recurrent (repeated connections), which means shared weights, variable-length input support, and hidden-to-hidden dependencies are all built-in features, not limitations.

No community discussion yet for this question.