IPMix: Label-Preserving Data Augmentation Method for Training Robust Classifiers.
Zhenglin HuangXiaoan BaoNa ZhangQingqi ZhangXiao TuBiao WuXi YangPublished in: NeurIPS (2023)
Keyphrases
- synthetic data
- training samples
- input data
- test data
- training data
- data sets
- prior knowledge
- noisy data
- training examples
- labeling process
- machine learning methods
- data analysis
- support vector machine
- classification trees
- statistical methods
- classification method
- missing data
- high dimensional data
- classification algorithm
- missing values
- computationally efficient
- feature selection
- extracted features
- face detection
- probabilistic model
- classification accuracy
- similarity measure
- decision trees
- training process
- multi class
- learning algorithm