Unit 3 Test Review

In supervised machine learning, humans provide an AI algorithm with data that is labeled (so the AI can associate the data with a specific name or class).

• True

• False

Machine learning is a type of AI algorithm used frequently today.

• True

• False

If we were training a supervised machine learning algorithm to identify different types of dogs, what might we use as labels for the different classes? (Select all correct answers.)

• Poodle

• Bulldog

• Garbage truck

• Cocker spaniel

• Elephant

If we were training a supervised machine learning algorithm to identify different types of fish what might we use as labels for the different classes? (Select all correct answers.)

• Goldfish

• Shark

• Boat

• Tetra fish

• Hamburger

How can humans supervise machine learning?

• By putting laptops in laptop bags

• By providing organized and labeled training data

• By explaining to an AI what steps to take, in human language

• It’s not possible, all machine learning is unsupervised

Supervised machine learning can use an algorithm to sort objects into different types – such as sorting baseball caps from top hats, or elephants from rhinoceroses. What do we call algorithms that do this?

• Recommendation algorithms

• Classification algorithms

• Colorful algorithms

• Search algorithms

If we wanted to train a machine learning algorithm to identify whether a person is sketching a mad or a joyful face, what should the dataset we use contain – to make sure it is diverse?

• Both mad and joyful face sketches drawn by many different types of people

• Only joyful face sketches drawn by many different types of people

• Only mad face sketches drawn by many different types of people

• Songs about people with mad or joyful faces

What should we do after training a machine learning algorithm using a training dataset?

• Nothing. We should just let it do its job without anything more — it’s now ready!

• Draw a picture of the dataset on paper.

• Test it on a different dataset (without labels) to pick out where it is making mistakes

• Stop using the dataset because it’s impossible to know whether it’s accurate or not.

A neural network is a type of machine learning algorithm similar to the human brain, that passes information from node to node.

• True

• False

One of the challenges of using neural networks is that it is a “black box” – we can’t see what features the algorithms are picking out.

• True

• False