Traditional Programming and Machine Learning

Traditional Programming

Machine Learning

Computer

The rules of the program (genre, etc.) are applied to the data (the list of songs).

Data

A list of songs.

Data

A list of 200 songs I like, or, a list of 400 songs labeled "like" or "don't like." **Machine learning requires large amounts of data!**

Output

A playlist of songs I like.

Computer

Looks for patterns such as genre, artist, tempo, etc. and uses those patterns to make predictions about what songs that aren't in the original data set that I may like.

Output

A playlist of songs I like.

Program

The program defines specific features of songs you like: genre, singer, tempo, etc. and has rules about how to choose those songs from the list (data).

Algorithm

An algorithm that can predict what songs I like based on the data (large list of songs I like or list of songs labeled "like" or "don't like."

1. A program is given to the computer.
2. Data is given to the computer.
3. The rules given in the program are applied to the data.
4. The result of the rules applied to that data is the output.

1. Data is given to the computer.
2. Output, or the correct answer, is given to the computer.
3. The computer creates algorithms.
4. The algorithm makes predictions for how to get the desired output from the input provided.

Make a Playlist of Songs I Like

Hover over the text in each shape to learn how that piece fits in to the two styles of computing