3.1 Machine learning
Machine learning is a technique for automatically finding patterns in large amounts of data. Watch the following video extract in which Wing discusses several applications of machine learning.
Activity 10 Machine learning
Wing gives the example of basketball coaches who use machine learning to find out which skills distinguish good players. Once they know this, they can teach those skills to their own players. Machine learning makes this possible by automatically finding patterns in the behaviour of professional players from a large collection of video recordings.
Describe this use of machine learning in terms of abstraction as modelling.
One can view this use of machine learning as an attempt to automatically build a model of a good basketball player. Such a model is an abstraction that needs to capture the skills that distinguish a good player and ignores anything else. At a different level, machine learning is itself based on a model of learning by humans or animals. Being an abstraction, it ignores many of the details of human and animal learning, while preserving some key properties (in particular, the idea that one can learn from examples).