Paper: Lattice-based Minimum Error Rate Training for Statistical Machine Translation

ACL ID D08-1076
Title Lattice-based Minimum Error Rate Training for Statistical Machine Translation
Venue Conference on Empirical Methods in Natural Language Processing
Session Main Conference
Year 2008
Authors

Minimum Error Rate Training (MERT) is an effective means to estimate the feature func- tion weights of a linear model such that an automated evaluation criterion for measuring system performance can directly be optimized in training. To accomplish this, the training procedure determines for each feature func- tion its exact error surface on a given set of candidate translations. The feature function weights are then adjusted by traversing the error surface combined over all sentences and picking those values for which the resulting error count reaches a minimum. Typically, candidates in MERT are represented as N- best lists which contain the N most probable translation hypotheses produced by a decoder. In this paper, we present a novel algorithm that allows for efficiently constructing and...