Paper: Improving IBM Word Alignment Model 1

ACL ID P04-1066
Title Improving IBM Word Alignment Model 1
Venue Annual Meeting of the Association of Computational Linguistics
Session Main Conference
Year 2004

We investigate a number of simple methods for improving the word-alignment accuracy of IBM Model 1. We demonstrate reduction in alignment error rate of approximately 30% resulting from (1) giving extra weight to the probability of alignment to the null word, (2) smoothing probability esti- mates for rare words, and (3) using a simple heuris- tic estimation method to initialize, or replace, EM training of model parameters.