For each of the following, state whether you would use binary classification, multi-class classification, or regression and give a one or two sentence justification.
What are the general steps of most learning algorithms?
What is the difference between online learning algorithms and bath learning algorithms?
How can binary classfiers be used to solve a multiclass classification problem?
How can linear models deal with data sets that are not linearly separable due to noise in the data?
What is the form of the hypothesis class of linear classifiers?
What are the functional forms and the loss functions for
What is a local minimum?
What is a global minimum?
What is the meaning of the learning rate hyperparameter to a learning algorithm?
What may happen if the learning rate is set too low?
What may happen if the learning rate is set too high?
Describe two desirable features of the sigmoid loss function for logistic regression.
What is a feature transform?
Can a linear model be used to separate into classes the feature vectors of instances that are not linearly separable? If so, how?
What happens to sample complexity (the number of training samples we need to maintain a bound our generalization error) with higher-order polynomial transforms?
Write up your homework in a format that can be converted to a PDF file (like LaTeX) and name the PDF file
linear-model.pdf. Submit your
linear-model.pdf file on Canvas as an attachment. When you’re ready, double-check that you have submitted and not just saved a draft.
Practice safe submission! Verify that your HW files were truly submitted correctly, the upload was successful, and that your program runs with no syntax or runtime errors. It is solely your responsibility to turn in your homework and practice this safe submission safeguard.
This procedure helps guard against a few things.