Non-Uniform Stochastic Average Gradient Method for Training Conditional Random Fields.
Mark SchmidtReza BabanezhadMohamed Osama AhmedAaron DefazioAnn CliftonAnoop SarkarPublished in: AISTATS (2015)
Keyphrases
- conditional random fields
- gradient method
- structured prediction
- error correcting output coding
- sequence labeling
- graphical models
- probabilistic model
- convergence rate
- hidden markov models
- web page prediction
- higher order
- markov random field
- negative matrix factorization
- crf model
- optimization methods
- generative model
- pairwise
- information extraction
- step size
- multi class
- training set
- named entities
- supervised learning
- bayesian networks