Machine Learning MT23, Paradigms
Flashcards
@Define the active learning ML paradigm.
Supervised learning, but where the data is initially unlabelled and the algorithm tries to identify the most useful data to label in order to learn the most effectively.
@Define the semi-supervised learning ML paradigm.
Like supervised learning, but where only a portion of the data is labelled.
@Define the collaborative filtering ML paradigm.
Used in recommender systems, trying to calculate estimates for whether to recommend something to a user based on more complete data from other similar users.
@Define the reinforcement learning ML paradigm.
Learning the optimal behaviour in an environment in order to obtain the maximum reward (e.g. given by some reward function).