Not all Failure Modes are Created Equal: Training Deep Neural Networks for Explicable (Mis)Classification.
Alberto Olmo HernandezSailik SenguptaSubbarao KambhampatiPublished in: CoRR (2020)
Keyphrases
- neural network
- training process
- failure modes
- pattern recognition
- multi layer perceptron
- training set
- training phase
- supervised learning
- classification accuracy
- training algorithm
- image classification
- classification scheme
- training samples
- classification performances
- pattern classification
- decision trees
- classification method
- learning vector quantization
- training data
- classification models
- training patterns
- classification algorithm
- unsupervised learning
- text classification
- feature vectors
- artificial neural networks
- feature extraction
- support vector
- rule extraction
- feedforward neural networks
- feed forward neural networks
- neural nets
- test set
- feature space
- genetic algorithm
- machine learning
- neural network training
- deep architectures
- benchmark classification problems
- radial basis function network
- training dataset
- incremental learning
- svm classifier
- class labels
- back propagation
- support vector machine
- feature selection