Free Printable rnn_lstm_gru

Worksheets

rnn_lstm_gru_autoencoder

Total questions: 71

Worksheet time: 33mins

Name

Class

Date

What is the primary objective of an Autoencoder?

Classification

Clustering

Data Reconstruction

Dimensionality Increse

Which of the following is NOT a component of an Autoencoder?

Encoder

Decoder

Latent Space

Optimizer

What does the latent space in an Autoencoder represent?

An uncompressed version of the input data.

A random selection of the input data.

A compressed representation of the input data.

A detailed explanation of the input data.

Which of the following techniques is commonly used to train an Autoencoder?

Semi-supervised Learning

Supervised Learning

Reinforcement Learning

Unsupervised Learning

What is the role of the decoder in an Autoencoder?

To reconstruct the input data from the latent space.

To compress the input data.

To classify the input data.

To visualize the input data.

Which loss function is commonly used in Autoencoders for reconstruction?

Mean Squared Error (MSE)

Categorical Cross-Entropy

Hinge Loss

Binary Cross-Entropy

What is a common application of Autoencoders?

Data Augumentation

Image Denoising

Sentiment Analysis

Text Classification

What is a Denoising Autoencoder trained to do?

Reconstruct clean data from noisy input

Reconstruct the noise

Encode noisy data

Maximize reconstruction error

What distinguishes a Variational Autoencoder (VAE) from a standard Autoencoder?

It has no decoder

It uses probabilistic latent variables

It skips the encoding process

It does not use a loss function

10.

In a Variational Autoencoder, what distribution is typically assumed for the latent variables?

Exponential distribution

Bernoulli distribution

Uniform distribution

Gaussian distribution

11.

What is the purpose of the KL-divergence term in the VAE loss function?

To measure reconstruction error

To minimize noise

To regularize latent space distribution

To increase model complexity

12.

What is a Sparse Autoencoder trained to encourage?

Very few neurons Active at a time

More neurons Active at a time

No hidden Layers

Random Noise in Output

13.

What regularization technique is commonly used in Sparse Autoencoders?

L1 regularization

Gradient Clipping

Dropout regularization

L2 regularization

14.

What is the purpose of Contractive Autoencoders?

Maximize the size of latent space

Penalize sensitivity to input variations

Denoise input data

Increase reconstruction error

15.

Which of the following is NOT a regularized Autoencoder?

Sparse Autoencoder

Convolutional Autoencoder

Denoising Autoencoder

Contractive Autoencoder

16.

Feedback networks are used for?

auto association

pattern storage

both auto association & pattern storage

pattern recognition

17.

_________ is used to find local minima of the cost function

stochastic gradient descent

gradient descent

linear regression

logistic regression

18.

A = 1/(1 + e^-x) is an equation representing which activation function?

ReLU

Sigmoid

Leaky ReLu

Tanh

19.

Identify the activation function from the given diagram

Sigmoid, ReLU

ReLU, Leaky ReLU

ReLU, Tanh

TanH, ReLU

20.

What type of deep learning architecture is used in a multimodal multistream framework to tackle egocentric activity recognition using both video and sensor data?

A) Single Convolutional Neural Network (CNN)

B) Feedforward Neural Network

C) Dual CNNs and Long Short-Term Memory (LSTM)

D) Recurrent Neural Network (RNN)

21.

they perform the same task for every element of a sequence, with the output being depended on the previous computations.

RNN or Recurrent Neural Network

CNN or Convolution Neural Network

LSTMor Long-Short Term Memory

22.

What are general limitations of back propagation rule? Pick the right choice from the given options

i) local minima problem

ii) slow convergence

iii) scaling

only i

both ii and i

both ii and iii

i, ii and iii

23.

What does RNN stand for?

Recurring Neural Network

Removable Neural Network

Recurrent Neural Network

None

24.

Which type of Neural Network is suitable to solve spatial feature recognition?

CNN

Fully Connected Network

ANN

RNN

25.

These are all problems suitable to be solved by DL, except.....

Image processing

Syntax segmentation

NLP

Speech recognition

26.

Which statement best describes the training process of an LSTM model?

a) It uses gradient descent to optimize the model’s weights

b) It uses the forward-backward algorithm for sequence labeling

c) It leverages the attention mechanism for contextual information

d) It involves unfolding the LSTM through time and applying backpropagation through time (BPTT)

27.

Which variant of LSTM is designed to handle bidirectional sequence modeling?

a) Gated Recurrent Unit (GRU)

b) Clockwork RNN

c) Bidirectional LSTM

d) Stacked LSTM

28.

Which of the following techniques is commonly used to alleviate the vanishing gradient problem in LSTM networks?

a) Dropout regularization

b) Batch normalization

c) Gradient clipping

d) Learning rate decay

29.

Which component(s) is/are present in an LSTM cell?

a) Input gate

b) Output gate

c) Forget gate

d) All of the above

30.

Which activation function(s) are commonly used in LSTM networks?

a) Sigmoid

b) Tanh

c) ReLU

d) Both a) and b)

31.

Which of the following is not the promise of artificial neural network?

a) it can explain result

b) it can survive the failure of some nodes

c) it has inherent parallelism

d) it can handle noise

32.

Which of the following is the component of learning system?

a) goal

b) model

c) learning rules

d) all of the mentioned

33.

What is the purpose of the recurrent connection in an RNN?

a) To propagate the hidden state across different time steps

b) To adjust the weights and biases of the network

c) To reduce the dimensionality of the input data

d) None of the above

34.

Which layer type is commonly used in RNNs for speech recognition tasks?

a) Input layer

b) Hidden layer

c) Output layer

d) None of the above

35.

What is the primary purpose of a Recurrent Neural Network (RNN)?

a) Image classification

b) Text generation

c) Reinforcement learning

d) Object detection

36.

What is the basic concept of Recurrent Neural Network?

Use previous inputs to find the next output according to the training set.

Use a loop between inputs and outputs in order to achieve the better prediction.

Use recurrent features from dataset to find the best answers.

Use loops between the most important features to predict next output.

37.

According to the image, classify the type of connection we have in the example 1.

One to one

One to many

Many to one

Many to many

38.

According to the image, classify the type of connection we have in the example 3.

Many to one

Many to many

One to one

One to many

39.

What is 'gradient' when we are talking about RNN?

A gradient is a partial derivative with respect to its inputs

It is how RNN calls its features

The most important step of RNN algorithm

A parameter that can help you improve the algorithm's accuracy

40.

The other RNN's issue is called 'Vanishing Gradients'. What is that?

When the values of a gradient are too small and the model stops learning or takes way too long because of that.

When the values of a gradient are too big and the model stops learning or takes way too long because of that.

When the values of a gradient are too small and the model joins in a loop because of that.

When the values of a gradient are too big and the model joins in a loop because of that.

41.

LSTM? What is that?

LSTM networks are an extension for recurrent neural networks, which basically extends their memory. Therefore it is well suited to learn from important experiences that have very long time lags in between

LSTM networks are an extension for recurrent neural networks, which basically shorten their memory. Therefore it is well suited to learn from important experiences that have very low time lags in between

LSTM networks are an extension for recurrent neural networks, which basically extends their memory. Therefore it is not recommended to use it, unless you are using a small Dataset.

42.

Question 3

To which of these tasks would you apply a many-to-one RNN architecture?

Image classification (input an image and output a label)

Music genre recognition

Language recognition from speech (input an audio clip and output a label indicating the language being spoken)

Speech recognition (input an audio clip and output a transcript)

43.

Using this as the training model below, answer the following:
True/False: At the $t^{th}$ time step the RNN is estimating

$P\ \left(y^{<t>}\ \left|\ y^{<1>},\ y^{<2>},\ ...,\ y^{<t-1>}\right|\right)$
p/s: This is ' P( ....| ....)'

False

True

44.

Suppose your training examples are sentences (sequences of words). Which of the following refers to the $j^{th}$ word in the $i^{th}$ training example?

$x^{\left(i\right)<j>}$

$x^{<i>\left(j\right)}$

$x^{\left(j\right)<i>}$

$x^{<j>\left(i\right)}$

45.

Consider this RNN:
This specific type of architecture is appropriate when:

$T_x\ =\ T_y$

$T_{x\ }<T_y$

$T_{x\ }>T_y$

$T_{x\ }=\ 1$

46.

Question 3

To which of these tasks would you apply a many-to-one RNN architecture?

Image classification (input an image and output a label)

Music genre recognition

Language recognition from speech (input an audio clip and output a label indicating the language being spoken)

Speech recognition (input an audio clip and output a transcript)

47.

Using this as the training model below, answer the following:
True/False: At the $t^{th}$ time step the RNN is estimating

$P\ \left(y^{<t>}\ \left|\ y^{<1>},\ y^{<2>},\ ...,\ y^{<t-1>}\right|\right)$
p/s: This is ' P( ....| ....)'

False

True

48.

You have finished training a language model RNN and are using it to sample random sentences, as follows:

True/False: In this sample sentence, step t uses the probabilities output by the RNN to randomly sample a chosen word for that time-step. Then it passes this selected word to the next time-step.

False

True

49.

True/False: If you are training an RNN model, and find that your weights and activations are all taking on the value of NaN (“Not a Number”) then you have an exploding gradient problem.

False

True

50.

Suppose you are training an LSTM. You have an 80000 word vocabulary, and are using an LSTM with 800-dimensional activations $a^{<t>}$ . What is the dimension of Γu at each time step?

800

80000

100

51.

True/False: In order to simplify the GRU without vanishing gradient problems even when training on very long sequences you should remove the Γr i.e., setting Γr= 1 always.

False

True

52.

True/False: Using the equations for the GRU and LSTM below the Update Gate and Forget Gate in the LSTM play a different role to Γu and 1- Γu.

False

True

53.

Your mood is heavily dependent on the current and past few days’ weather. You’ve collected data for the past 365 days on the weather, which you represent as a sequence as $x^{<1>}$ ,…, $x^{<365>}$ . You’ve also collected data on your mood, which you represent as $y^{<1>}$ ,…, $y^{<365>}$ . You’d like to build a model to map from x→y. Should you use a Unidirectional RNN or Bidirectional RNN for this problem?

Bidirectional RNN, because this allows the prediction of mood on day t to take into account more information

Undirectional RNN, because the value of $y^{<t>}$ depends on $x^{<t>}$ , and not other days' weather

Bidirectional RNN, because this allows the prediction of mood on day t to take into account more information

Undirectional RNN, because the value of $y^{<t>}$ depends only on $x^{<1>}$ ,..., $x^{<t>}$ , but not on $x^{<1>}$ ,..., $x^{<365>}$

54.

Which type of Neural Network is suitable to solve spatial feature recognition?

CNN

Fully Connected Network

ANN

RNN

55.

Which statement best describes the training process of an LSTM model?

a) It uses gradient descent to optimize the model’s weights

b) It uses the forward-backward algorithm for sequence labeling

c) It leverages the attention mechanism for contextual information

d) It involves unfolding the LSTM through time and applying backpropagation through time (BPTT)

56.

Which variant of LSTM is designed to handle bidirectional sequence modeling?

a) Gated Recurrent Unit (GRU)

b) Clockwork RNN

c) Bidirectional LSTM

d) Stacked LSTM

57.

Which of the following techniques is commonly used to alleviate the vanishing gradient problem in LSTM networks?

a) Dropout regularization

b) Batch normalization

c) Gradient clipping

d) Learning rate decay

58.

Which component(s) is/are present in an LSTM cell?

a) Input gate

b) Output gate

c) Forget gate

d) All of the above

59.

Which activation function(s) are commonly used in LSTM networks?

a) Sigmoid

b) Tanh

c) ReLU

d) Both a) and b)

60.

What is the purpose of the recurrent connection in an RNN?

a) To propagate the hidden state across different time steps

b) To adjust the weights and biases of the network

c) To reduce the dimensionality of the input data

d) None of the above

61.

Which layer type is commonly used in RNNs for speech recognition tasks?

a) Input layer

b) Hidden layer

c) Output layer

d) None of the above

62.

What is the primary purpose of a Recurrent Neural Network (RNN)?

a) Image classification

b) Text generation

c) Reinforcement learning

d) Object detection

63.

How does the YOLO (You Only Look Once) algorithm improve object detection performance?

By dividing the image into a grid and predicting bounding boxes and class probabilities for each grid cell

By using a sliding window approach to scan the entire image for objects

By applying a separate classifier for each object class in the image

By resizing the input image to a fixed size before processing

64.

What is the purpose of non-maximum suppression in object detection algorithms?

To reduce the number of redundant bounding box predictions for the same object

To increase the resolution of the feature maps for better object localization

To introduce randomness in the training process for better generalization

To adjust the learning rate dynamically during optimization

65.

What is the primary advantage of using bidirectional LSTM networks over unidirectional ones?

Improved memory cell capacity

Ability to process sequence data in both forward and backward directions

Reduced computational complexity

Enhanced control over weight updates

66.

How does the concept of attention mechanism enhance the performance of LSTM networks?

By focusing on specific parts of the input sequence during processing

By increasing the number of memory cells in the network

By reducing the number of layers in the LSTM architecture

By using a different activation function for the forget gate

67.

What is the purpose of the forget gate in a Long Short-Term Memory (LSTM) network?

To store the output of the activation function

To control the flow of information through the network

To maintain long-term dependencies in the sequence data

To determine the learning rate for weight updates

68.

How does the LSTM architecture differ from traditional recurrent neural networks (RNNs)?

LSTM networks have a single hidden layer, while RNNs have multiple layers

LSTM networks use a memory cell to store information over time steps

RNNs have a forget gate to control the flow of information

LSTM networks do not support sequential data processing

69.

What is the purpose of the cell state in a Long Short-Term Memory (LSTM) network?

To store the output of the activation function

To control the flow of information through the network

To maintain long-term dependencies in the sequence data

To determine the learning rate for weight updates

70.

Why is regularization important in machine learning models?

To control the size of the weight updates during optimization

To optimize model parameters and reduce overfitting

To define the number of layers in the neural network

To select the type of activation function for the model

71.

What is the significance of the learning rate in gradient descent optimization?

To determine the number of epochs for training

To control the size of the weight updates during optimization

To define the number of layers in the neural network

To select the type of activation function for the model

Reset

rnn_lstm_gru_autoencoder

Total questions: 71

Worksheet time: 33mins

Which statement best describes the training process of an LSTM model?

Which variant of LSTM is designed to handle bidirectional sequence modeling?

Which of the following techniques is commonly used to alleviate the vanishing gradient problem in LSTM networks?

Which component(s) is/are present in an LSTM cell?

Which activation function(s) are commonly used in LSTM networks?

Which of the following is not the promise of artificial neural network?

Which of the following is the component of learning system?

Which statement best describes the training process of an LSTM model?

Which variant of LSTM is designed to handle bidirectional sequence modeling?

Which of the following techniques is commonly used to alleviate the vanishing gradient problem in LSTM networks?

Which component(s) is/are present in an LSTM cell?

Which activation function(s) are commonly used in LSTM networks?

How does the YOLO (You Only Look Once) algorithm improve object detection performance?

What is the purpose of non-maximum suppression in object detection algorithms?

What is the primary advantage of using bidirectional LSTM networks over unidirectional ones?

How does the concept of attention mechanism enhance the performance of LSTM networks?

What is the purpose of the forget gate in a Long Short-Term Memory (LSTM) network?

How does the LSTM architecture differ from traditional recurrent neural networks (RNNs)?

What is the purpose of the cell state in a Long Short-Term Memory (LSTM) network?

Why is regularization important in machine learning models?

What is the significance of the learning rate in gradient descent optimization?

Similar Resources on Wayground