ML Anthology

Generalized Isotonic Conditional Random Fields

Yi Mao, Guy Lebanon
MLJ 2009 pp. 225-248
doi:10.1007/S10994-009-5139-1 /mlj/2009/mao2009mlj-generalized/

Abstract

Conditional random fields are one of the most popular structured prediction models. Nevertheless, the problem of incorporating domain knowledge into the model is poorly understood and remains an open issue. We explore a new approach for incorporating a particular form of domain knowledge through generalized isotonic constraints on the model parameters. The resulting approach has a clear probabilistic interpretation and efficient training procedures. We demonstrate the applicability of our framework with an experimental study on sentiment prediction and information extraction tasks.

PDF MLJ Semantic Scholar

Cite

Text

Mao and Lebanon. "Generalized Isotonic Conditional Random Fields." Machine Learning, 2009. doi:10.1007/S10994-009-5139-1

Markdown

[Mao and Lebanon. "Generalized Isotonic Conditional Random Fields." Machine Learning, 2009.](https://mlanthology.org/mlj/2009/mao2009mlj-generalized/) doi:10.1007/S10994-009-5139-1

BibTeX

@article{mao2009mlj-generalized,
  title     = {{Generalized Isotonic Conditional Random Fields}},
  author    = {Mao, Yi and Lebanon, Guy},
  journal   = {Machine Learning},
  year      = {2009},
  pages     = {225-248},
  doi       = {10.1007/S10994-009-5139-1},
  volume    = {77},
  url       = {https://mlanthology.org/mlj/2009/mao2009mlj-generalized/}
}