Skip to content Skip to main navigation Report an accessibility issue

ACE machine learning content

The online curriculum is composed of four modules:

  • What is machine learning?
  • Foundations of machine learning: Linear models
  • Foundations of machine learning: Model evaluation
  • Foundations of machine learning: Non-linear models.

Step 1: What is machine learning?

  • Download and watch the Introduction video. (Dropbox access is provided after registration).
  • Complete the Introduction to machine learning quiz.

Introduction to machine learning quiz

Answer the questions.
  • Your contact information

  • Enter your first name. For example, Sally or Joe.
  • Enter your last name (surname). For example, Smith.
  • Enter your school's name. Enter your company name if you are not currently a student.
  • Enter your registration email address. For example, sally.smith@utk.edu.
  • Questions

  • ML enables learning directly from data without being explicitly programmed, as opposed to physics-based approaches that require pre-defined rules and models.
  • Although physics-based approaches are generally reliable, using ML may be needed if the underlying relationships and patterns are unknown.
  • In supervised learning, the input data does not require the desired solution or labels.
  • Classification and regression are two types of unsupervised learning.
  • Supervised machine learning algorithms can be used to predict the label for a new data point.
  • This field is for validation purposes and should be left unchanged.

 

Step 2: Foundations of machine learning: Linear models

  • Download and watch the Linear models video. (Dropbox access is provided after registration).
  • Complete the Linear models quiz.

Linear models quiz

Answer the questions.
  • Your contact information

  • Enter your first name. For example, Sally or Joe.
  • Enter your last name (surname). For example, Smith.
  • Enter your school's name. Enter your company name if you are not currently a student.
  • Enter your registration email address. For example, sally.smith@utk.edu.
  • Questions

  • Linear regression is a supervised learning algorithm.
  • When using linear regression, the estimated/predicted values for the response variable may follow a quadratic function.
  • In logistic regression, the estimated/predicted values for the response variable are bounded.
  • In logistic regression, the model output is the estimated probability that the response variable falls in a certain class.
  • Logistic regression cannot be used for classification.
  • This field is for validation purposes and should be left unchanged.

 

Step 3: Foundations of machine learning: Model evaluation

  • Download and watch the Model evaluation video. (Dropbox access is provided after registration).
  • Complete the Model evaluation quiz.

Model evaluation quiz

Answer the questions.
  • Your contact information

  • Enter your first name. For example, Sally or Joe.
  • Enter your last name (surname). For example, Smith.
  • Enter your school's name. Enter your company name if you are not currently a student.
  • Enter your registration email address. For example, sally.smith@utk.edu.
  • Questions

  • A model may be tested objectively by re-applying it to the training data.
  • In k-fold cross validation, one fold is left out for testing at a time, while all other folds are used for training.
  • In binary classification, accuracy, sensitivity, and specificity metrics provide some measure of model performance.
  • In binary classification, mean absolute error and mean squared error are the best metrics to evaluate model performance.
  • An area under the receiver operating characteristic curve (AUC) that is less than 0.5 indicates a good model performance.
  • This field is for validation purposes and should be left unchanged.

 

Step 4: Foundations of machine learning: Non-linear models

  • Download and watch the Non-linear models video. (Dropbox access is provided after registration).
  • Complete the Nonlinear models quiz.

Non-linear models quiz

Answer the questions.
  • Your contact information

  • Enter your first name. For example, Sally or Joe.
  • Enter your last name (surname). For example, Smith.
  • Enter your school's name. Enter your company name if you are not currently a student.
  • Enter your registration email address. For example, sally.smith@utk.edu.
  • Questions

  • Decision trees are robust against overfitting by design.
  • Random forest is an ensemble classifier consisting of multiple decision trees.
  • Despite the term “random” in the name, random forests are completely deterministic.
  • Feature importance in random forests provides insight into the most important independent variables in the model.
  • Neural networks are an example of linear models.
  • This field is for validation purposes and should be left unchanged.