Large Margin Training for Hidden Markov Models with Partially Observed States

Abstract

Large margin learning of Continuous Density Hidden Markov Models with a partially labeled dataset has been extensively studied in the speech and handwriting recognition fields. Yet due to the non convexity of the optimization problem, previous works usually rely on severe approximations so that it is still an open problem. We propose a new learning algorithm that relies on non convex optimization and bundle methods and allows tackling the original optimization problem as is. It is proved to converge to a solution with accuracy $\epsilon$ with a rate $O(1/\epsilon)$. We provide experimental results gained on speech recognition and on handwriting recognition that demonstrate the potential of the method.

Cite

Text

Do and Artières. "Large Margin Training for Hidden Markov Models with Partially Observed States." International Conference on Machine Learning, 2009. doi:10.1145/1553374.1553408

Markdown

[Do and Artières. "Large Margin Training for Hidden Markov Models with Partially Observed States." International Conference on Machine Learning, 2009.](https://mlanthology.org/icml/2009/do2009icml-large/) doi:10.1145/1553374.1553408

BibTeX

@inproceedings{do2009icml-large,
  title     = {{Large Margin Training for Hidden Markov Models with Partially Observed States}},
  author    = {Do, Trinh Minh Tri and Artières, Thierry},
  booktitle = {International Conference on Machine Learning},
  year      = {2009},
  pages     = {265-272},
  doi       = {10.1145/1553374.1553408},
  url       = {https://mlanthology.org/icml/2009/do2009icml-large/}
}