Adaptive boosting, also known as AdaBoost, is a machine learning algorithm used to improve the accuracy of a classification model. It works by iteratively training a series of weak classifiers on a dataset, and then combining their predictions to produce a more accurate final prediction.