liveBook · Manning

Generalization

As explained in Machines that Think

Generalization is the ability of a learned system to apply patterns or representations beyond the data on which it learned them.

Generalization is the capacity to carry learned patterns beyond the examples used for training. It distinguishes learning that captures relationships across examples from merely storing or reproducing them. section 3.3

This distinction matters because memorization can make a system brittle, while generalization supports responses to unfamiliar data. Yet performance beyond training examples does not by itself establish awareness or understanding; it may count as abstraction without resolving what understanding means. section 1.6.2 section 3.3

Memorizing or learning

A network that reproduces its training examples is confined to familiar cases; one that captures patterns across examples can respond to new data. section 3.3

How Machines that Think covers it

How deep it goes
Defines generalization through its contrast with memorization and connects it to abstraction, without detailing how to produce it.
Examples
Contrasts a network that reproduces training examples with one that responds to new data.
What you can do after
Evaluate it by distinguishing pattern extraction from memorization and treating generalization as evidence of abstraction, not awareness.
How it connects
Connects generalization to learning, memorization, abstraction, neural networks, and machine understanding.

Also explained in

Compare how 19 books explain generalization →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Machines that Think. The section links point to the parts of the book it draws on.