= Error tolerance (PAC learning)
{wiki=Error_tolerance_(PAC_learning)}
Error tolerance in the context of PAC (Probably Approximately Correct) learning relates to the ability of a learning algorithm to produce a hypothesis that is approximately correct with respect to a certain error rate. PAC learning is a framework introduced by Leslie Valiant in 1984 to formalize the concept of learning from examples in a statistical sense. In PAC learning, the goal is to learn a target function (or concept) from a set of training examples that are drawn from a probability distribution.
Back to article page