Deep Learning and Reinforcement Learning | Coursera Quiz Answers

Deep Learning and Reinforcement Learning | Coursera (IBM)

This course provides an introduction to two highly sought-after fields in Machine Learning: Deep Learning and Reinforcement Learning. Deep Learning, a subset of Machine Learning, is utilized in both Supervised and Unsupervised Learning, powering many of the AI applications we use daily. The course begins by covering the theory behind Neural Networks, the foundation of Deep Learning, and explores various modern architectures. After building several Deep Learning models, the focus shifts to Reinforcement Learning, a Machine Learning approach that has gained significant attention recently. While Reinforcement Learning currently has limited practical applications, it is a promising research area in AI with potential future relevance.

Upon completing this course, and if you have followed the IBM Specialization sequence, you will have substantial practice and a solid understanding of the main types of Machine Learning: Supervised Learning, Unsupervised Learning, Deep Learning, and Reinforcement Learning. By the end of this course, you should be able to:

• Identify problems suitable for Unsupervised Learning
• Explain the curse of dimensionality and its impact on clustering with many features
• Describe and apply common clustering and dimensionality-reduction algorithms
• Perform clustering where appropriate and compare the performance of per-cluster models
• Understand metrics for characterizing clusters

This course is designed for aspiring data scientists looking to gain practical experience with Deep Learning and Reinforcement Learning. To benefit fully from this course, you should be familiar with programming in a Python development environment and have a fundamental understanding of Data Cleaning, Exploratory Data Analysis, Unsupervised Learning, Supervised Learning, Calculus, Linear Algebra, Probability, and Statistics.

Enroll Course

---

Notice!
Always refer to the module on your course for the most accurate and up-to-date information.

Attention!
If you have any questions that are not covered in this post, please feel free to leave them in the comments section below. Thank you for your engagement.

WEEK 1 QUIZ

1. What is another name for the “neuron” on which all neural networks are based?

• perceptron

2. What is an advantage of using a network of neurons?

• A network of neurons can represent a non-linear decision boundary.

3. A dataset with 8 features would have how many nodes in the input layer?

• 8

4. For a single data point, the weights between an input layer with 3 nodes and a hidden layer with 4 nodes can be represented by a:

• 3 x 4 matrix.

5. Use the following image for reference. How many hidden layers are in this Neural Network?

• Two

6. Use the following image for reference. How many hidden units are in this Neural Network?

• Eight

7. Which statement is TRUE about the relationship between Neural Networks and Logistic Regression?

• A single-layer Neural Network can be parameterized to generate results equivalent to Linear or Logistic Regression.

WEEK 2 QUIZ

1. The backpropagation algorithm updates which of the following?

• The parameters only.

2. What of the following about the activation functions is true?

• They add non-linearity into the model, allowing the model to learn complex pattern.

3. What is true regarding the backpropagation rule?

• The actual output is determined by computing the output of neurons in each hidden layer

4. Which option correctly lists the steps to build a linear regression model using Keras?
1. Use fit() and specify the number of epochs to train the model for.
2. Create a Sequential model with the relevant layers.
3. Normalize the features with  layers.Normalization() and apply adapt().
4. Compile using model.compile() with specified optimizer and loss.

• ANSWER: 3, 2, 4, 1

5. (True/False) Keras provides one approach to build a model: by defining a Sequential model.

• False

WEEK 3 QUIZ

1. What is the main function of backpropagation when training a Neural Network?

• Make adjustments to the weights

2. (True/False) The “vanishing gradient” problem can be solved using a different activation function.

• True

3. (True/False) Every node in a neural network has an activation function.

• True

4. These are all activation functions except:

• Leaky hyperbolic tangent

5. Deep Learning uses deep Neural Networks for all these uses, except

• Cases in which explainability is the main objective

6. These are all activation functions except:

• Pruning

7. (True/False) Optimizer approaches for Deep Learning Regularization use gradient descent:

• False

8. Stochastic gradient descent is this type of batching method:

• online learning

9. (True/False) The main purpose of data shuffling during the training of a Neural Network is to aid convergence and use the data in a different order each epoch.

• True

10. This is a high-level library that is commonly used to train deep learning models and runs on either TensorFlow or Theano:

• Keras

WEEK 4 QUIZ

1. What is the main function of backpropagation when training a Neural Network?

• Make adjustments to the weights

2.(True/False) The “vanishing gradient” problem can be solved using a different activation function.

• True

3. (True/False) Every node in a neural network has an activation function.

• True

4. What is the main function of backpropagation when training a Neural Network? These are all activation functions except:

•  Sigmoid
•  Hyperbolic tangent
•  Leaky hyperbolic tangent
•  ReLu

5. Deep Learning uses deep Neural Networks for all these uses, except:

• Cases in which explainability is the main objective

6. These are all activation functions for CNN, except:

• Pruning

7. (True/False) Optimizer approaches for Deep Learning Regularization use gradient descent:

• False

8. Stochastic gradient descent is this type of batching method:

• online learning

9. The main purpose of data shuffling during the training of a Neural Network is to aid convergence and use the data in a different order each epoch.

• True

10. Which of the following IS NOT a benefit of Transfer Learning?

• Improving the speed at which large models can be trained from scratch

11. Which of the following statements about using a Pooling Layer is TRUE?

• Pooling can reduce both computational complexity and overfitting.

WEEK 5 QUIZ

1. (True/False) In Keras, the Dropout layer has an argument called rate, which is a probability that represents how often we want to invoke the layer in the training.

• False

2. What is a benefit of applying transfer learning to neural networks?

• Save early layers for generalization before re-training later layers for specific applications.

3. By setting  layer.trainable = False for certain layers in a neural network, we____

• freeze the layers such that their weights don’t update during training.

4. Which option correctly orders the steps of implementing transfer learning?

1. Freeze the early layers of the pre-trained model.
2. Improve the model by fine-tuning.
3. Train the model with a new output layer in place.
4. Select a pre-trained model as the base of our training.

• ANSWER: 4, 1, 3, 2

5. Given a 100x100 pixels RGB image, there are _____ features.

• 30000 There is one feature per pixel of each channel, so 1001003=30000.

6. Before a CNN is ready for classifying images, what layer must we add as the last?

• Dense layer with the number of units corresponding to the number of classes

7. In a CNN, the depth of a layer corresponds to the number of:

• filters applied

WEEK 6 QUIZ

1. (True/False) RNN models are mostly used in the fields of natural language processing and speech recognition.

• True

2. (True/False) GRUs and LSTM are a way to deal with the vanishing gradient problem encountered by RNNs.

• True

3. (True/False) GRUs will generally perform about as well as LSTMs with shorter training time, especially for smaller datasets.

• True

4. (True/False) The main idea of Seq2Seq models is to improve accuracy by keeping necessary information in the hidden state from one sequence to the next.

• True

5. (True/False) The main parts of a Seq2Seq model are: an encoder, a hidden state, a sequence state, and a decoder.

• False

6. Select the correct option, in the context of Seq2Seq models:

• The Greedy Search algorithm selects one best candidate as an input sequence for each time step while the Beam Search produces multiple different hypothesis based on conditional probability.

7. Which is the gating mechanism for RNNs that include a reset gate and an update gate?

• GRUs

8. LSTM models are among the most common Deep Learning models used in forecasting. These are other common uses of LSTM models, except:

• Speech Recognition
•  Machine Translation
•  Image Captioning
•  Generating Images
•  Anomaly Detection
•  Robotic Control

WEEK 7 QUIZ

1. Select the correct option:
Statement 1: Autoencoders are a supervised learning technique.
Statement 2: Autoencoder’s output is exactly the same as the input.

• Both statements are false.

2. Select the correct option:
Statement 1: Autoencoders can be viewed as a generalization of PCA that discovers lower dimensional representation of complex data.
Statement 2: We can implement overcomplete autoencoder by constraining the number of units present in the hidden layers of the neural network.

• Statement 1 is true, statement 2 is false.

3. (True/False) Denoising autoencoders can be used as a tool for feature extraction.

• True

4. (True/False) An Autoencoder is a form of unsupervised deep learning.

• True

WEEK 8 QUIZ

1. Select the right assertion:

• Autoencoders learn from a compressed representation of the data, while variational autoencoders learn from a probability distribution representing the data.

2. (True/False) Variational autoencoders are generative models.

• True

3. When comparing the results of Autoencoders and Principal Component Analysis, which approach might best improve the results from Autoencoders?

• Add layers and epochs

4. (True/False) KL loss is used in Variatoinal Autoencoders to represent the measure of the difference between two distributions.

• True

5. A good way to compare the inputs and outputs of a Variational Autoencoder is to calculate the mean of a reconstruction function based on binary crossentropy.

• True

6. The main parts of GANs architecture are:

• generator and discriminator

7. (True/False) One of the main advantages of GANs over other adversarial networks is that it does not spend any time evaluating whether an input or image is fake or real. It only computes probability of being fake.

• True

WEEK 9 QUIZ

1. (True/False) Simulation is a common approach for Reinforcement Learning applications that are complex or computing intensive.

• True

2.(True/False) Discounting rewards refers to an agent reducing the value of the reward based on its uncertainty.

• False

3. (True/False) Successful Reinforcement Learning approaches are often limited by extreme sensitivity to hyperparameters.

• True

4. (True/False) Reinforcement Learning approaches are often limited by excessive computation resources and data requirements.

• True

5. Which type of Deep Learning approach is most commonly used for image recognition?

• Convolutional Neural Network

6. Which type of Deep Learning approach is most commonly used for forecasting problems?

• Recurrent Neural Network

7. Which type of Deep Learning approach is most commonly used for generating artificial images?

• Autoencoders

---

Related Articles