Loss function

An objective that assigns a scalar cost to a prediction, action, parameter setting, or comparison with a target.

An objective that assigns a scalar cost to a prediction, action, parameter setting, or comparison with a target.

is aObjective All entries Open in graph

A loss specifies what counts as error locally; an aggregate training objective specifies how those costs are combined and regularized.

Connections

has specialization

Adversarial training objective Contrastive objective Cross-entropy loss Orthogonality penalty

related to

Empirical risk