Abstract

The development of high-performance statistical machine translation (SMT) systems is contingent on the availability of substantial, in-domain parallel training corpora. The latter, however, are expensive to produce due to the labor-intensive nature of manual translation. We propose to alleviate this problem with a novel, semi-supervised, batch-mode active learning strategy that attempts to maximize in-domain coverage by selecting sentences, which represent a balance between domain match, translation difficulty, and batch diversity. Simulation experiments on an English-to-Pashto translation task show that the proposed strategy not only outperforms the random selection baseline, but also traditional active selection techniques based on dissimilarity to existing training data.

Date

2013

Authors

Sankaranarayanan Ananthakrishnan, Rohit Prasad, David Stallard, Prem Natarajan

Journal

Computer Speech & Language

Volume

27

Issue

2

Pages

397-406

Publisher

Academic Press