A Structured Variational Autoencoder for Contextual Morphological Inflection

@inproceedings{Naradowsky2018ASV,
  title={A Structured Variational Autoencoder for Contextual Morphological Inflection},
  author={Jason Naradowsky and Ryan Cotterell and Sebastian J. Mielke and Lawrence Wolf-Sonkin},
  booktitle={ACL},
  year={2018}
}
Statistical morphological inflectors are typically trained on fully supervised, type-level data. One remaining open research question is the following: How can we effectively exploit raw, token-level data to improve their performance? To this end, we introduce a novel generative latent-variable model for the semi-supervised learning of inflection generation… CONTINUE READING