In this work, we present the first results for neuralizing an Unsupervised Hidden Markov Model. We evaluate our approach on tag in- duction. Our approach outperforms existing generative models and is competitive with the state-of-the-art though with a simpler model easily extended to include additional context.
@article{arxiv.1609.09007,
title = {Unsupervised Neural Hidden Markov Models},
author = {Ke Tran and Yonatan Bisk and Ashish Vaswani and Daniel Marcu and Kevin Knight},
journal= {arXiv preprint arXiv:1609.09007},
year = {2016}
}
Comments
accepted at EMNLP 2016, Workshop on Structured Prediction for NLP. Oral presentation