Faster decoding for subword level Phrase-based SMT between related languages

TL;DR

This paper explores optimizing decoder parameters and data formats in subword-based phrase SMT for related languages, significantly reducing decoding time with minimal impact on translation quality.

Contribution

It identifies optimal decoder settings and data formats that enhance decoding efficiency in subword phrase-based SMT systems for related languages.

Findings

Decoding time is significantly reduced with optimized settings.

Minimal loss in translation accuracy with the proposed optimizations.

Guidelines for choosing data formats and decoder parameters.

Abstract

A common and effective way to train translation systems between related languages is to consider sub-word level basic units. However, this increases the length of the sentences resulting in increased decoding time. The increase in length is also impacted by the specific choice of data format for representing the sentences as subwords. In a phrase-based SMT framework, we investigate different choices of decoder parameters as well as data format and their impact on decoding time and translation accuracy. We suggest best options for these settings that significantly improve decoding time with little impact on the translation accuracy.