Translate Smart, not Hard: Cascaded Translation Systems with Quality-Aware Deferral
Ant{\'o}nio Farinhas, Nuno M Guerreiro, Sweta Agrawal, Ricardo Rei, Andre Martins
Abstract
Larger models often outperform smaller ones but come with high computational costs. Cascading offers a potential solution. By default, it uses smaller models and defers only some instances to larger, more powerful models. However, designing effective deferral rules remains a challenge. In this paper, we propose a simple yet effective approach for machine translation, using existing quality estimation (QE) metrics as deferral rules. We show that QE-based deferral allows a cascaded system to match the performance of a larger model while invoking it for a small fraction (30% to 50%) of the examples, significantly reducing computational costs. We validate this approach through both automatic and human evaluation.
BibTeX
@inproceedings{emnlp2025_translatesmartno,
title = {Translate Smart, not Hard: Cascaded Translation Systems with Quality-Aware Deferral},
author = {Ant{\'o}nio Farinhas and Nuno M Guerreiro and Sweta Agrawal and Ricardo Rei and Andre Martins},
booktitle = {EMNLP 2025},
year = {2025}
}