ML Anthology

Two-View Feature Generation Model for Semi-Supervised Learning

Rie Kubota Ando, Tong Zhang
ICML 2007 pp. 25-32
doi:10.1145/1273496.1273500 /icml/2007/ando2007icml-two/

Abstract

We consider a setting for discriminative semisupervised learning where unlabeled data are used with a generative model to learn effective feature representations for discriminative training. Within this framework, we revisit the two-view feature generation model of cotraining and prove that the optimum predictor can be expressed as a linear combination of a few features constructed from unlabeled data. From this analysis, we derive methods that employ two views but are very different from co-training. Experiments show that our approach is more robust than co-training and EM, under various data generation conditions.

PDF ICML Semantic Scholar

Cite

Text

Ando and Zhang. "Two-View Feature Generation Model for Semi-Supervised Learning." International Conference on Machine Learning, 2007. doi:10.1145/1273496.1273500

Markdown

[Ando and Zhang. "Two-View Feature Generation Model for Semi-Supervised Learning." International Conference on Machine Learning, 2007.](https://mlanthology.org/icml/2007/ando2007icml-two/) doi:10.1145/1273496.1273500

BibTeX

@inproceedings{ando2007icml-two,
  title     = {{Two-View Feature Generation Model for Semi-Supervised Learning}},
  author    = {Ando, Rie Kubota and Zhang, Tong},
  booktitle = {International Conference on Machine Learning},
  year      = {2007},
  pages     = {25-32},
  doi       = {10.1145/1273496.1273500},
  url       = {https://mlanthology.org/icml/2007/ando2007icml-two/}
}