Training Conditional Random Fields for Maximum Labelwise Accuracy.
Samuel S. GrossOlga RussakovskyChuong B. DoSerafim BatzoglouPublished in: NIPS (2006)
Keyphrases
- conditional random fields
- structured prediction
- supervised training
- hidden markov models
- sequence labeling
- information extraction
- probabilistic model
- graphical models
- random fields
- error correcting output coding
- web page prediction
- crf model
- higher order
- maximum entropy
- named entity recognition
- approximate inference
- markov random field
- prediction accuracy
- generative model
- weighted sums
- pairwise
- structured learning
- image labeling
- parameter learning
- markov models
- markov networks
- training set
- restricted boltzmann machine
- data mining
- classification accuracy
- training process
- undirected graphical models
- training samples
- conditional likelihood
- clustering algorithm
- multi class
- superpixels