Definition
Passive Learning
Passive learning is the standard paradigm where the learner is provided with a fixed set of labelled data sampled i.i.d. from an underlying joint distribution . In this setting, the learner has no control over the data acquisition process and must optimise based on the provided sample.
Characteristics
Passive learning relies on large-scale, unbiased sampling to capture the full characteristics of the data distribution. Unlike active learning, it often includes redundant or low-information samples, resulting in lower sample efficiency. However, it remains the most common paradigm due to its simplicity and direct alignment with the i.i.d. assumption.