EM Can Find Pretty Good HMM POS-Taggers (When Given a Good Start)

  title={EM Can Find Pretty Good HMM POS-Taggers (When Given a Good Start)},
  author={Yoav Goldberg and Meni Adler and Michael Elhadad},
We address the task of unsupervised POS tagging. We demonstrate that good results can be obtained using the robust EM-HMM learner when provided with good initial conditions, even with incomplete dictionaries. We present a family of algorithms to compute effective initial estimations p(tjw). We test the method on the task of full morphological disambiguation in Hebrew achieving an error reduction of 25% over a strong uniform distribution baseline. We also test the same method on the standard WSJ… CONTINUE READING

3 Figures & Tables



Citations per Year

72 Citations

Semantic Scholar estimates that this publication has 72 citations based on the available data.

See our FAQ for additional information.