# Naver Labs Europe's Systems for the WMT19 Machine Translation Robustness Task

## TL;DR

This paper presents Naver Labs Europe's top-ranked ensemble systems for the WMT19 robustness task, focusing on improving machine translation's resilience to social media noise through specialized preprocessing and domain adaptation.

## Contribution

The paper introduces ensemble translation systems optimized for social media noise robustness, achieving top BLEU scores and discussing effective preprocessing and adaptation strategies.

## Key findings

- Ensemble models ranked first in all language pairs.
- Effective preprocessing improved noise robustness.
- Domain adaptation strategies enhanced translation quality.

## Abstract

This paper describes the systems that we submitted to the WMT19 Machine Translation robustness task. This task aims to improve MT's robustness to noise found on social media, like informal language, spelling mistakes and other orthographic variations. The organizers provide parallel data extracted from a social media website in two language pairs: French-English and Japanese-English (in both translation directions). The goal is to obtain the best scores on unseen test sets from the same source, according to automatic metrics (BLEU) and human evaluation. We proposed one single and one ensemble system for each translation direction. Our ensemble models ranked first in all language pairs, according to BLEU evaluation. We discuss the pre-processing choices that we made, and present our solutions for robustness to noise and domain adaptation.