Computer scienceData scienceMachine learningEnsemble learning

Ensemble methods

Training ensembles in parallel

Report a typo

Ensembling involves training several machine learning models, or base learners.

In which ensembles is it possible to train those base learners in parallel to speed up the learning process? You can assume that the computational resources allow for parallel computation.

Select all correct answers.

Select one or more options from the list

Bagged ensembles, e.g. Random Forest.

Boosted ensembles, e.g. AdaBoost or gradient boosting.

It depends on the base learner we choose to use in our ensemble: some can be trained in parallel, and others can't.

No ensemble can be trained in parallel: each base model always depends on the previously trained one, making simultaneous training impossible.

___

Create a free account to access the full topic