Introduction to Learning

Learning in AI is the process by which a system improves its performance on a task over time, using experience, examples, instruction, or its own reasoning — without being explicitly reprogrammed for every new case.

A Working Definition

"Learning is any change in a system that allows it to perform better the second time on repetition of the same task, or on another task drawn from the same population." — Herbert Simon

The General Learning Model

Why Learning Is Necessary

• Designers cannot anticipate every situation the system will face
• The environment can change over time — a static knowledge base becomes stale
• Some tasks (recognising handwriting, translating idioms) are too complex to hand-code exhaustively
• Learning lets a system get better rather than staying fixed at its initial competence

Forms of Learning Covered in This Unit

Classifying Learning Methods