Improving selective classification performance of deep neural networks through post-hoc logit normalization and temperature scaling.
Luís Felipe P. CattelanDanilo SilvaPublished in: CoRR (2023)
Keyphrases
- neural network
- pattern recognition
- post hoc
- classification accuracy
- rule extraction
- automatic classification
- machine learning
- pattern classification
- preprocessing
- feature extraction
- decision trees
- support vector machine
- classification systems
- text classification
- classification scheme
- decision rules
- learning vector quantization
- classification method
- classification models
- recurrent neural networks
- incremental learning
- classification process
- support vector machine svm
- supervised learning
- radial basis function network
- fuzzy decision trees
- training set
- genetic algorithm
- learning rules
- support vector
- classification algorithm
- fault diagnosis
- feature vectors