Question $1$

Which of the following are valid examples of using unsupervised learning to improve supervised learning? Select all that apply.

$1$ point

You apply dimensionality reduction using multi$-$dimensional scaling to a $1200-$dimensional dataset in order to train a k$-$nearest$-$neighbors classifier with a complex distance function more efficiently.

You perform hyperparameter tuning to optimize the recall of a regularized logistic regression text classifier.

Your site does product recommendation given a customer profile. Given a new customer to your site, you compute the probability of seeing that customer's profile in your training data. After seeing too many customers that are low$-$probability outliers, you consider re$-$training the classifier.

You apply k$-$means clustering to the colors in an image in order to learn an effective limited$-$size palette that can be used for compressing the image.