Experience AI Lessons - Summative Assessment

What is the difference between a traditional computer program and an ML model?

• ML models use rule-based IF/THEN statements

• ML models use a data-driven approach

• They are the same; the only difference is that ML models use more data

Which one of the following applications would most benefit from using an ML model?

• A. A login system that checks that the correct password has been entered for a valid user

• B. An online calculator that converts miles to kilometres

• C. A text prediction application that saves you time by predicting the next word you want to use in a sentence

AI applications can be used to generate works of art. What data does the model need to be able to generate the art?

• Vast quantities of example artwork

• Human-programmed IF/THEN rules

• One image to copy from each known style of art

A popular music streaming service has been designed using machine learning. Which one of the following services would benefit from using a data-driven approach?

• Recommending a playlist based on the songs you commonly listen to

• Adding a song to your favourites when you select the ‘thumbs up’ icon

• Calculating how many minutes of music you listen to during the day

Which type of machine learning involves the use of pre-labelled data?

• Supervised learning

• Unsupervised learning

• Reinforcement learning

Which one of the following describes the supervised learning approach to classification?

• Learning by trial and error

• Creating groups from similar items of data

• Categorising data by applying labels

The following diagram shows that a piece of data has been labelled as an orange. Which one of the following describes the term ‘label’?

• A category that data can be assigned to

• Applied to a single piece of data to indicate which class it belongs to

• The confidence score that is returned by the model

Which one of the following describes test data used during the creation of a classification model?

• Data used to check the accuracy of the model

• Data used to train the model

• Data entered into an application that uses the model

A machine learning model has been developed to identify birds from sound alone. Part A: Whilst testing and evaluating the model, developers have found that the model is not able to identify blackbirds. Which two of the following actions should the developers take to improve their model?

• A. Add more audio samples of all known varieties of birds

• B. Add audio samples of humans and other animals in a variety of environments

• C. Add audio samples of blackbirds in different conditions such as close up or far away

Part B: Whilst testing and evaluating the model, the developers found the following results. Which one of them is an indication of bias in the machine learning model?

• A. The app is 96% accurate for blackbirds and 48% accurate for pigeons

• B. The average accuracy across all birds is 68%

• C. The model is unable to identify human speech from sound