Recurrent Neural Networks - Introduction to RNN

1 Recurrent Neural Networks - Introduction to RNN

2 MCQ - RNN cannot work on sequences of arbitrary lengths of input

3 MCQ - RNN can analyze time series data such as stock prices

4 MCQ - RNN looks much like a feedforward neural network except it has connections pointing backward

5 MCQ - A part of a neural network that preserves some state across time steps is called a memory cell

6 Recurrent Neural Networks - Forecasting a Time Series

7 MCQ - In an encoder-decoder network, we have sequence-to-vector network, called an decoder followed by vector-to-sequence network, called a encoder

8 MCQ - To train an RNN, the trick is to unroll it through time and then simply use regular backpropagation. This strategy is called backpropagation through time (BPTT).

9 MCQ - When the data is a sequence of one or more values per time step it is called a time series

10 MCQ - Univariate time series involves two or more input variables

11 MCQ - The task to predict missing values from the past is called imputation

12 Recurrent Neural Networks - Forecasting several Time Steps ahead

13 MCQ - What is/are the option(s) to predict 10 time steps ahead using an RNN?

14 Recurrent Neural Networks - Handling Long Sequences

15 MCQ - The Long Short-Term Memory (LSTM) cell was proposed by Sepp Hochreiter and Jürgen Schmidhuber in the year

16 MCQ - The Gated Recurrent Unit (GRU) cell was proposed by Kyunghyun Cho et al. in the year

17 MCQ - GRU can be implemented in Keras using the keras.layers.GRU layer