Transductive Support Vector Machines for Structured Variables

Abstract

We study the problem of learning kernel machines transductively for structured output variables. Transductive learning can be reduced to combinatorial optimization problems over all possible labelings of the unlabeled data. In order to scale transductive learning to structured variables, we transform the corresponding non-convex, combinatorial, constrained optimization problems into continuous, unconstrained optimization problems. The discrete optimization parameters are eliminated and the resulting differentiable problems can be optimized efficiently. We study the effectiveness of the generalized TSVM on multiclass classification and labelsequence learning problems empirically.

Cite

Text

Zien et al. "Transductive Support Vector Machines for Structured Variables." International Conference on Machine Learning, 2007. doi:10.1145/1273496.1273645

Markdown

[Zien et al. "Transductive Support Vector Machines for Structured Variables." International Conference on Machine Learning, 2007.](https://mlanthology.org/icml/2007/zien2007icml-transductive/) doi:10.1145/1273496.1273645

BibTeX

@inproceedings{zien2007icml-transductive,
  title     = {{Transductive Support Vector Machines for Structured Variables}},
  author    = {Zien, Alexander and Brefeld, Ulf and Scheffer, Tobias},
  booktitle = {International Conference on Machine Learning},
  year      = {2007},
  pages     = {1183-1190},
  doi       = {10.1145/1273496.1273645},
  url       = {https://mlanthology.org/icml/2007/zien2007icml-transductive/}
}