Make sure the content matches the objectives for each episode

[svenvanderburg]

Necessary to meet the requirements for carpentries lab. See #264

[dsmits]

Episode 1:

1. [x] Recall the sort of problems for which Deep Learning is a useful tool
2. [x] List some of the available tools for Deep Learning
3. [x] Recall the steps of a Deep Learning workflow
4. [ ] Explain why it is important to test the accuracy of Deep Learning system.
5. [x] Identify the inputs and outputs of a Deep Learning system.
6. [x] Test that we’ve correctly installed the Keras, Seaborn and Sklearn libraries

For this one, I'd say we ditch number 4

[dsmits]

Episode 2:

1. [x] Use the deep learning workflow to structure the notebook
2. [x] Explore the dataset using pandas and seaborn
3. [x] Use one-hot encoding to prepare data for classification in Keras
4. [x] Describe a fully connected layer
5. [x] Implement a fully connected layer with Keras
6. [x] Use Keras to train a small fully connected network on prepared data
7. [x] Plot the loss curve of the training process
8. [x] Use a confusion matrix to measure the trained networks’ performance on a test set

[dsmits]

Episode 3:

1. [x] Explain the importance of splitting the training data
2. [x] Use the data splits to plot the training process
3. [ ] Set the training goal for your deep neural network
4. [x] Measure the performance of your deep neural network
5. [x] Interpret the training plots to recognize overfitting
6. [x] Implement basic strategies to prevent overfitting

Number 3 is there in principle but I think the lesson should explicitly explain what a training goal is and not just put it in the title of the section.

[dsmits]

Episode 4:

1. [x] Understand why convolutional and pooling layers are useful for image data
2. [ ] Use normalization as preparation step for Deep Learning
3. [x] Implement a convolutional neural network on an image dataset
4. [x] Plot and interpret the training process
5. [ ] Do a visual inspection of the results of your neural network
6. [ ] Understand strategies to improve your model based on the plots
7. [x] Use drop-out layer to prevent overfitting

• Number 2: should be moved to episode 3
• Number 5 and 6 are kind of there but are more an objective of the previous chapter