4. Neural Methods
Rnns — Quiz
Test your understanding of rnns with 5 practice questions.
Practice Questions
Question 1
In a Gated Recurrent Unit (GRU), which two gates are combined into a single gate?
Question 2
When training RNNs, what is the primary purpose of 'clipping gradients'?
Question 3
Consider an LSTM cell. If the input gate's activation is $i_t$ and the candidate cell state's activation is $\tilde{C}_t$, which of the following expressions correctly represents the contribution of the new input to the cell state at time $t$?
Question 4
Which of the following best describes the 'long-term dependencies' problem that RNNs struggle with and LSTMs/GRUs aim to solve?
Question 5
In the context of RNNs, what is the primary role of the 'softmax' activation function, typically applied to the output layer?