Part 5

What is the building block of deep learning methods?

• Perceptron

• Regression

• Entropy

• Hyperparameters

Which error function is commonly used for regression problems?

• Squared error

• Cross-entropy

• Kullback-Leibler divergence

• SoftMax

What is the main limitation of machine learning algorithms?

• They can only interpret training data

• They are not accurate enough

• They have too many hyperparameters

• They cannot handle time series data

What is the phenomenon called when an algorithm underperforms on training data?

• Overfitting

• Underfitting

• Increasing complexity

• Training error

What is the phenomenon called when an algorithm underperforms on testing/validation data?

• Overfitting

• Underfitting

• Increasing complexity

• Training error

Machine learning algorithms can interpret aspects of the world that are not covered by their training data.

• True

• False

Increasing the complexity of a machine learning model always improves its performance.

• True

• False

Underfitting occurs when a model is too complex for a given task.

• True

• False

Overfitting results in the model memorizing the training data instead of generalizing.

• True

• False

A more complex algorithm is less likely to memorize a large dataset.

• True

• False