ML Anthology

The Asymptotics of Semi-Supervised Learning in Discriminative Probabilistic Models

Nataliya Sokolovska, Olivier Cappé, François Yvon
ICML 2008 pp. 984-991
doi:10.1145/1390156.1390280 /icml/2008/sokolovska2008icml-asymptotics/

Abstract

Semi-supervised learning aims at taking advantage of unlabeled data to improve the efficiency of supervised learning procedures. For discriminative models however, this is a challenging task. In this contribution, we introduce an original methodology for using unlabeled data through the design of a simple semi-supervised objective function. We prove that the corresponding semi-supervised estimator is asymptotically optimal. The practical consequences of this result are discussed for the case of the logistic regression model.

PDF ICML Semantic Scholar

Cite

Text

Sokolovska et al. "The Asymptotics of Semi-Supervised Learning in Discriminative Probabilistic Models." International Conference on Machine Learning, 2008. doi:10.1145/1390156.1390280

Markdown

[Sokolovska et al. "The Asymptotics of Semi-Supervised Learning in Discriminative Probabilistic Models." International Conference on Machine Learning, 2008.](https://mlanthology.org/icml/2008/sokolovska2008icml-asymptotics/) doi:10.1145/1390156.1390280

BibTeX

@inproceedings{sokolovska2008icml-asymptotics,
  title     = {{The Asymptotics of Semi-Supervised Learning in Discriminative Probabilistic Models}},
  author    = {Sokolovska, Nataliya and Cappé, Olivier and Yvon, François},
  booktitle = {International Conference on Machine Learning},
  year      = {2008},
  pages     = {984-991},
  doi       = {10.1145/1390156.1390280},
  url       = {https://mlanthology.org/icml/2008/sokolovska2008icml-asymptotics/}
}