Foundations and practice of machine learning

Definition and learning from experience

“Definition and learning from experience” refers to a core, operational way of defining machine learning: a computer program learns from experience when its performance on a set of tasks improves as it gains that experience. The definition is tied to measurable task performance rather than to human-like cognition, making it practical and testable. In this framework, experience (E) can be data, feedback, or training interactions, and learning is evaluated by how well the program performs on tasks (T) according to a performance measure (P). This idea underpins most machine learning approaches, where models are trained by iteratively adjusting internal parameters to reduce prediction errors, so that they can generalize to new, unseen data.

0/2 modes complete

History and key milestones

0/2 modes complete