Subword-Delimited Downsampling for Better Character-Level Translation

TL;DR

This paper introduces a novel subword-informed downsampling method for character-level translation models, improving quality and efficiency, and achieving competitive results compared to subword models.

Contribution

It proposes a new downsampling technique based on subword information that enhances character-level translation models without sacrificing quality.

Findings

Outperforms existing downsampling methods in translation quality

Enables character-level models to be competitive with subword models

Reduces computational costs of character-level models

Abstract

Subword-level models have been the dominant paradigm in NLP. However, character-level models have the benefit of seeing each character individually, providing the model with more detailed information that ultimately could lead to better models. Recent works have shown character-level models to be competitive with subword models, but costly in terms of time and computation. Character-level models with a downsampling component alleviate this, but at the cost of quality, particularly for machine translation. This work analyzes the problems of previous downsampling methods and introduces a novel downsampling method which is informed by subwords. This new downsampling method not only outperforms existing downsampling methods, showing that downsampling characters can be done without sacrificing quality, but also leads to promising performance compared to subword models for translation.