Factorized CRF with Batch Normalization Based on the Entire Training Data.
Eran GoldmanJacob GoldbergerPublished in: ICASSP (2021)
Keyphrases
- training data
- conditional random fields
- data sets
- training set
- decision trees
- learning algorithm
- prior knowledge
- training examples
- training instances
- graphical models
- maximum entropy
- test data
- hidden markov models
- classification accuracy
- random fields
- information extraction
- naive bayes
- preprocessing
- generalization error
- classification models
- support vector machine
- semantic segmentation
- fully connected
- batch mode
- normalization method
- learned from training data
- reinforcement learning
- class labels
- missing data
- test set
- pairwise
- domain knowledge