Why ReLU Networks Yield High-Confidence Predictions Far Away From the Training Data and How to Mitigate the Problem.
Matthias HeinMaksym AndriushchenkoJulian BitterwolfPublished in: CVPR (2019)
Keyphrases
- high confidence
- training data
- class labels
- association rules
- prior knowledge
- decision trees
- training set
- learning algorithm
- network analysis
- test data
- data sets
- supervised learning
- network design
- community detection
- classification accuracy
- domain knowledge
- training instances
- training dataset
- feature selection
- network structure
- learning environment
- test cases
- test set
- training examples
- training samples
- labeled data
- text mining
- support vector machine