Ensemble Methods as a Defense to Adversarial Perturbations Against Deep Neural Networks.
Thilo StraussMarkus HanselmannAndrej JungingerHolger UlmerPublished in: CoRR (2017)
Keyphrases
- ensemble methods
- neural network
- random forests
- prediction accuracy
- machine learning methods
- benchmark datasets
- decision trees
- ensemble learning
- random forest
- base classifiers
- pattern recognition
- generalization ability
- majority voting
- back propagation
- bootstrap sampling
- classifier ensemble
- drifting concepts
- base learners
- artificial neural networks
- genetic algorithm
- decision tree ensembles
- search space
- bp neural network
- multiple models
- ensemble classifier
- data sets
- inductive learning
- ensemble members
- ensemble pruning
- ensemble feature selection
- feature subset
- classification accuracy
- genetic algorithm ga