Boosting Unsupervised Machine Translation with Pseudo-Parallel Data

Ivana Kvapilíková, Ondřej Bojar
10/22/2023
cs.CL

No confident code implementation found for this paper.

Abstract

Even with the latest developments in deep learning and large-scale language modeling, the task of machine translation (MT) of low-resource languages remains a challenge. Neural MT systems can be trained in an unsupervised way without any translation resources but the quality lags behind, especially in truly low-resource conditions. We propose a training strategy that relies on pseudo-parallel sentence pairs mined from monolingual corpora in addition to synthetic sentence pairs back-translated from monolingual corpora. We experiment with different training schedules and reach an improvement of up to 14.5 BLEU points (English to Ukrainian) over a baseline trained on back-translated data only.

View on arXivView PDF

Code Implementations

No confident code match yet

We couldn't find an author-owned or strongly-evidenced community implementation for this paper. Any repos shown below are weak matches — verify before relying on them.

No code implementations found yet.

Know of an implementation? Let us know in the comments below!

Similar Papers

Back-Translation Still Matters: Leveraging Monolingual Data for Low-Resource Machine Translation with LLMs

89% match2026

Mending the Holes: Mitigating Reward Hacking in Reinforcement Learning for Multilingual Translation

89% match2026

PEACH: Pre-Training Sequence-to-Sequence Multilingual Models for Translation with Semi-Supervised Pseudo-Parallel Document Generation

88% match1 repo32023

Neural Machine Translation For Low Resource Languages

88% match2023

Replicable Benchmarking of Neural Machine Translation (NMT) on Low-Resource Local Languages in Indonesia

88% match2023

Low-Resource Cross-Lingual Adaptive Training for Nigerian Pidgin

87% match2023

Cite this paper

@article{kvapilkov2023boosting,
  title  = {Boosting Unsupervised Machine Translation with Pseudo-Parallel Data},
  author = {Ivana Kvapilíková and Ondřej Bojar},
  year   = {2023},
  eprint = {2310.14262},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2310.14262v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2310.14262.svg)](https://codeofpaper.com/paper/2310.14262)

Discussion