Bi-Text Alignment of Movie Subtitles for Spoken English-Arabic Statistical Machine Translation

TL;DR

This paper presents a novel algorithm for aligning movie subtitles at the fragment level using time information, significantly improving the quality of English-Arabic machine translation resources and performance.

Contribution

It introduces a new alignment method leveraging timing data, creating the largest bi-text resource for English-Arabic translation and outperforming existing tools.

Findings

Alignment quality improved over existing tools

Adding the new bi-text increased BLEU scores by over two points

Developed the largest and most accurate subtitle bi-text resource

Abstract

We describe efforts towards getting better resources for English-Arabic machine translation of spoken text. In particular, we look at movie subtitles as a unique, rich resource, as subtitles in one language often get translated into other languages. Movie subtitles are not new as a resource and have been explored in previous research; however, here we create a much larger bi-text (the biggest to date), and we further generate better quality alignment for it. Given the subtitles for the same movie in different languages, a key problem is how to align them at the fragment level. Typically, this is done using length-based alignment, but for movie subtitles, there is also time information. Here we exploit this information to develop an original algorithm that outperforms the current best subtitle alignment tool, subalign. The evaluation results show that adding our bi-text to the IWSLT training bi-text yields an improvement of over two BLEU points absolute.