Knowledge quiz

It's time to put what you've learned to the test, get 8 right to pass this week.

Which step is performed third in supervised learning?

Choose the correct answer.

Label new data according to a pattern.

Find a pattern to match input and labels.

Receive rewards or punishments.

Collect labeled data.

Which of the following examples does not fit supervised learning?

Choose the correct answer.

Finding the optimal settings for an air conditioner

Automated detection of spam mails

Predicting real estate prices

Distinguishing between cat and dog images

How can color photos be encoded within a computer?

Choose the correct answer.

Using pixels: Each pixel has a red, green, and blue portion, which is specified as a number between 0 and 255.

Using pixels: Each pixel is either red, green, or blue.

By describing a wave similar to music files

Using a textual description of which coordinates the computer must and must not place pixels at

Which conditions mark the end of a decision tree branch?

There are two correct answers.

The subset of data in the branch is no longer distinguishable.

The subset of data in the branch contains only one category (label).

The subset of data in the branch can still be distinguished.

The subset of data in the branch contains more than one category (label).

Which of the following ideas would improve the algorithm implemented in Snap!?

There are three correct answers.

Decreasing the step size of the prototypes with an increasing number of iterations

Multiple iterations over the given input data

Optimizing the initial positioning of the prototypes

Using every second data point as an input

Fewer prototypes

Which of the following statements about overfitting is true?

Choose the correct answer.

Overfitting is a problem in supervised learning that makes it impossible to generalize from the training data.

Overfitted models also work for unsupervised learning.

The process of testing the model with retained data is called "overfitting".

Overfitting is the process of collecting labeled data.

Why is supervised learning "supervised"?

Choose the correct answer.

Because the labels of the inputs are known from the training set, the method receives feedback on whether and to what extent the rules established based on the features already label the input correctly.

Because the programmer is constantly checking that the learning process is correct

Because the learning process takes place only in secret labs that are extensively monitored

Because the inputs are known, programmers can check if the model classifies them correctly, and step in if necessary.

Which step is performed second in unsupervised learning?

Choose the correct answer.

Find similarities and patterns

Collect unlabeled data

Identify clusters and exceptions

Find rules that correctly label known input

What does a prototype describe?

Choose the correct answer.

A typical representative of a group in a dataset

A very old coin

A number of groups in a data set

An AI program

10.

What is unsupervised learning?

Choose the correct answer.

A type of Machine Learning where only unlabeled data is available as input. The process independently identifies similarities and patterns in the input data, such as to group the data or find outliers.

A process similar to the learning progress of children growing up without their parents.

A type of Machine Learning where the learning process receives feedback on whether and to what extent the established rules already provide the input with the correct label.

The fact that a programmer has no influence on the results of the learning process.