liveBook · Manning

Generalization

As explained in Machine Learning Algorithms in Depth

The ability of a learned model to perform accurately on previously unseen examples.

Generalization is a model’s ability to perform accurately on previously unseen examples. It requires low loss not only on labeled training data but also on new, unseen test data. This ability distinguishes useful machine-learning algorithms from systems that merely fit their training examples. section 1.1

Evaluate generalization by comparing performance across:

  • Training data, whose labels are already available and support fitting the model. section 1.1
  • Unseen test data, where low loss indicates that learned behavior extends beyond the training set. section 1.1

How Machine Learning Algorithms in Depth covers it

How deep it goes
Defines the training-versus-test objective but does not derive methods for improving generalization or analyze model internals.
Examples
Uses training examples and new test examples as the concrete comparison for judging accuracy and loss.
What it assumes
Assumes familiarity with labeled training data, predictions, ground-truth labels, and loss-based accuracy assessment.
What you can do after
Evaluate/understand whether a model performs accurately beyond its labeled training data by examining unseen test examples.
How it connects
Connects generalization to loss functions, training data, test data, supervised learning, and artificial general intelligence.
Cover of Machine Learning Algorithms in Depth
liveBookmanning.com
Read Machine Learning Algorithms in Depth on liveBook Get the book at manning.com

Also explained in

Compare how 13 books explain generalization →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Machine Learning Algorithms in Depth. The section links point to the parts of the book it draws on.