Optimal Weak to Strong Learning

The field of boosting has been started from a classic question in learning theory asking whether classifiers that are just slightly better than random guessing can be used to create a classifier with arbitrarily high accuracy when given enough training data.