Skip to content
conftrace_
2021 EMNLP EMNLP 2021

NoahNMT at WMT 2021: Dual Transfer for Very Low Resource Supervised Machine Translation

Abstract

AbstractThis paper describes the NoahNMT system submitted to the WMT 2021 shared task of Very Low Resource Supervised Machine Translation. The system is a standard Transformer model equipped with our recent technique of dual transfer. It also employs widely used techniques that are known to be helpful for neural machine translation, including iterative back-translation, selected finetuning, and ensemble. The final submission achieves the top BLEU for three translation directions.

🌉 Interdisciplinary Bridge - Artificial Intelligence and Natural Language Processing

Authors

Meng Zhang, Minghao Wu, Pengfei Li, Liangyou Li, Qun Liu

Topics

Deep Learning > Architectures > Transformers Natural Language Processing > Applications > Machine Translation Machine Learning > Learning Types > Transfer Learning Deep Learning > Learning Types > Transfer Learning Deep Learning > Learning Types > Multi-Task Learning

Keywords

transfer learning ensemble learning neural machine translation model ensemble transformer model iterative back-translation dual transfer very low resource selected fine-tuning
Download PDF

More from EMNLP 2021

All EMNLP 2021 papers →
The LMU Munich Systems for the WMT21 Unsupervised and Very Low-Resource Translation Task
Language Model Pretraining and Transfer Learning for Very Low Resource Languages
NRC-CNRC Systems for Upper Sorbian-German and Lower Sorbian-German Machine Translation 2021
cushLEPOR: customising hLEPOR metric using Optuna for higher agreement with human judgments or pre-trained language model LaBSE
MTEQA at WMT21 Metrics Shared Task
Are References Really Needed? Unbabel-IST 2021 Submission for the Metrics Shared Task