Filters








3,488 Hits in 4.8 sec

Contrastive Learning for Many-to-many Multilingual Neural Machine Translation [article]

Xiao Pan, Mingxuan Wang, Liwei Wu, Lei Li
2021 arXiv   pre-print
To this end, we propose mRASP2, a training method to obtain a single unified multilingual translation model. mRASP2 is empowered by two techniques: a) a contrastive learning scheme to close the gap among  ...  In this work, we aim to build a many-to-many translation system with an emphasis on the quality of non-English language directions.  ...  The overall framework is illustrated in Figure 1 Multilingual Transformer A multilingual neural machine translation model learns a many-to-many mapping function f to translate from one language to  ... 
arXiv:2105.09501v3 fatcat:2yui6p4t3bhgzpbmxshb6xctfi

Contrastive Learning for Many-to-many Multilingual Neural Machine Translation

Xiao Pan, Mingxuan Wang, Liwei Wu, Lei Li
2021 Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)   unpublished
To this end, we propose mRASP2, a training method to obtain a single unified multilingual translation model. mRASP2 is empowered by two techniques: a) a contrastive learning scheme to close the gap among  ...  In this work, we aim to build a many-to-many translation system with an emphasis on the quality of non-English language directions.  ...  The overall framework is illustrated in Figure 1 Multilingual Transformer A multilingual neural machine translation model learns a many-to-many mapping function f to translate from one language to  ... 
doi:10.18653/v1/2021.acl-long.21 fatcat:qz3pskwwkfdaxjsyygdxf7qela

Improving Multilingual Neural Machine Translation For Low-Resource Languages: French,English - Vietnamese [article]

Thi-Vinh Ngo, Phuong-Thai Nguyen, Thanh-Le Ha, Khac-Quy Dinh, Le-Minh Nguyen
2021 arXiv   pre-print
Prior works have demonstrated that a low-resource language pair can benefit from multilingual machine translation (MT) systems, which rely on many language pairs' joint training.  ...  The first strategy is about dynamical learning word similarity of tokens in the shared space among source languages while another one attempts to augment the translation ability of rare words through updating  ...  Improved zero-shot neural machine translation via ignoring spurious c CoRR, abs/1906.01181. Effective Strategies in Zero-Shot Neural Machine Translation.  ... 
arXiv:2012.08743v2 fatcat:un6v4f72evb2nlodruvotawyeq

A Brief Survey of Multilingual Neural Machine Translation [article]

Raj Dabre, Chenhui Chu, Anoop Kunchukuttan
2020 arXiv   pre-print
Many approaches have been proposed in order to exploit multilingual parallel corpora for improving translation quality.  ...  We present a survey on multilingual neural machine translation (MNMT), which has gained a lot of traction in the recent years.  ...  A neural interlingua for multilingual machine translation. In Proceedings of the Third Conference on Machine Translation: Research Papers, pages 84-92. Association for Computational Linguistics.  ... 
arXiv:1905.05395v3 fatcat:cyufmt3y65bhjofvt5zeljahz4

Sicilian Translator: A Recipe for Low-Resource NMT [article]

Eryk Wdowiak
2021 arXiv   pre-print
With 17,000 pairs of Sicilian-English translated sentences, Arba Sicula developed the first neural machine translator for the Sicilian language.  ...  Then we supplemented our dataset with backtranslation and multilingual translation and pushed our scores into the mid 30s.  ...  I would like to thank them for their support and encouragement. Prof. Cipolla helped me learn Sicilian and he also helped me develop this recipe for low-resource neural machine translation.  ... 
arXiv:2110.01938v1 fatcat:2p5m5xmkcrcfzer4vzk5463zhy

An Evaluation of Language-Agnostic Inner-Attention-Based Representations in Machine Translation

Alessandro Raganato, Raúl Vázquez, Mathias Creutz, Jörg Tiedemann
2019 Proceedings of the 4th Workshop on Representation Learning for NLP (RepL4NLP-2019)  
In contrast to related previous work, we demonstrate that the performance in translation does correlate with trainable downstream tasks.  ...  In this paper, we explore a multilingual translation model with a cross-lingually shared layer that can be used as fixed-size sentence representation in different downstream tasks.  ...  The authors gratefully acknowledge the support of the Academy of Finland through project 314062 from the ICT 2023 call on Computation, Machine Learning and Artificial Intelligence and project 270354/273457  ... 
doi:10.18653/v1/w19-4304 dblp:conf/rep4nlp/RaganatoVCT19 fatcat:ibw6fd42dzhmtm2vritujjvh7y

Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation [article]

Biao Zhang, Philip Williams, Ivan Titov, Rico Sennrich
2020 arXiv   pre-print
Massively multilingual models for neural machine translation (NMT) are theoretically attractive, but often underperform bilingual models and deliver poor zero-shot translations.  ...  Experiments on OPUS-100 (a novel multilingual dataset with 100 languages) show that our approach substantially narrows the performance gap with bilingual models in both one-to-many and many-to-many settings  ...  Neural machine translation by jointly learning to align and translate.  ... 
arXiv:2004.11867v1 fatcat:lt3fiwqur5bmhbbbfvebduomym

Learning Language Representations for Typology Prediction [article]

Chaitanya Malaviya, Graham Neubig, Patrick Littell
2017 arXiv   pre-print
When a neural machine translation system learns to translate from one language to another, does it learn the syntax or semantics of the languages?  ...  Exploiting the existence of parallel texts in more than a thousand languages, we build a massive many-to-one neural machine translation (NMT) system from 1017 languages into English, and use this to predict  ...  Acknowledgments We thank Lori Levin and David Mortensen for their useful comments and also thank the reviewers for their feedback about this work.  ... 
arXiv:1707.09569v1 fatcat:vwtcfspokfedbfvrshxte5b2tm

Learning Language Representations for Typology Prediction

Chaitanya Malaviya, Graham Neubig, Patrick Littell
2017 Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing  
When a neural machine translation system learns to translate from one language to another, does it learn the syntax or semantics of the languages?  ...  Exploiting the existence of parallel texts in more than a thousand languages, we build a massive many-to-one neural machine translation (NMT) system from 1017 languages into English, and use this to predict  ...  Acknowledgments We thank Lori Levin and David Mortensen for their useful comments and also thank the reviewers for their feedback about this work.  ... 
doi:10.18653/v1/d17-1268 dblp:conf/emnlp/MalaviyaNL17 fatcat:ucmtp53ksze4jeoofa3rpcquui

A neural interlingua for multilingual machine translation [article]

Yichao Lu, Phillip Keung, Faisal Ladhak, Vikas Bhardwaj, Shaonan Zhang, Jason Sun
2018 arXiv   pre-print
We incorporate an explicit neural interlingua into a multilingual encoder-decoder neural machine translation (NMT) architecture.  ...  We demonstrate that our model learns a language-independent representation by performing direct zero-shot translation (without using pivot translation), and by using the source sentence embeddings to create  ...  Conclusion We incorporate a neural interlingua component into the standard encoder-decoder framework for multilingual neural machine translation, and demonstrate that the resulting model learns language-independent  ... 
arXiv:1804.08198v3 fatcat:mcxvz5zypvcydoikupzxzvetdy

A neural interlingua for multilingual machine translation

Yichao Lu, Phillip Keung, Faisal Ladhak, Vikas Bhardwaj, Shaonan Zhang, Jason Sun
2018 Proceedings of the Third Conference on Machine Translation: Research Papers  
We incorporate an explicit neural interlingua into a multilingual encoder-decoder neural machine translation (NMT) architecture.  ...  We demonstrate that our model learns a languageindependent representation by performing direct zero-shot translation (without using pivot translation), and by using the source sentence embeddings to create  ...  Introduction Multilingual Machine Translation Neural machine translation (NMT) relies on word and sentence embeddings to encode the semantic information needed for translation.  ... 
doi:10.18653/v1/w18-6309 dblp:conf/wmt/LuKLBZS18 fatcat:ceiv3kpejfh2rfva5a5orh7ncm

Meta-Learning for Low-Resource Neural Machine Translation

Jiatao Gu, Yong Wang, Yun Chen, Victor O. K. Li, Kyunghyun Cho
2018 Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing  
In this paper, we propose to extend the recently introduced model-agnostic meta-learning algorithm (MAML, Finn et al., 2017) for lowresource neural machine translation (NMT).  ...  We frame low-resource translation as a metalearning problem, and we learn to adapt to low-resource languages based on multilingual high-resource language tasks.  ...  Acknowledgement This research was supported in part by the Facebook Low Resource Neural Machine Translation Award.  ... 
doi:10.18653/v1/d18-1398 dblp:conf/emnlp/GuWCLC18 fatcat:mkiddurmxzgkhgwwwdmrwdcptu

Meta-Learning for Low-Resource Neural Machine Translation [article]

Jiatao Gu, Yong Wang, Yun Chen, Kyunghyun Cho, Victor O.K. Li
2018 arXiv   pre-print
In this paper, we propose to extend the recently introduced model-agnostic meta-learning algorithm (MAML) for low-resource neural machine translation (NMT).  ...  We frame low-resource translation as a meta-learning problem, and we learn to adapt to low-resource languages based on multilingual high-resource language tasks.  ...  Acknowledgement This research was supported in part by the Facebook Low Resource Neural Machine Translation Award.  ... 
arXiv:1808.08437v1 fatcat:4mxnnqu2dnbwdiegrp43r3p3c4

When do Contrastive Word Alignments Improve Many-to-many Neural Machine Translation? [article]

Zhuoyuan Mao, Chenhui Chu, Raj Dabre, Haiyue Song, Zhen Wan, Sadao Kurohashi
2022 arXiv   pre-print
Word alignment has proven to benefit many-to-many neural machine translation (NMT).  ...  Meanwhile, the contrastive objective can implicitly utilize automatically learned word alignment, which has not been explored in many-to-many NMT.  ...  Acknowledgements This work was supported by Grant-in-Aid for Young Scientists #19K20343, JSPS and Information/AI/Data Science Doctoral Fellowship of Kyoto University.  ... 
arXiv:2204.12165v1 fatcat:27fdplhoyjf7zpzkwhkz3nwrbe

XLM-T: Scaling up Multilingual Machine Translation with Pretrained Cross-lingual Transformer Encoders [article]

Shuming Ma, Jian Yang, Haoyang Huang, Zewen Chi, Li Dong, Dongdong Zhang, Hany Hassan Awadalla, Alexandre Muzio, Akiko Eriguchi, Saksham Singhal, Xia Song, Arul Menezes (+1 others)
2020 arXiv   pre-print
Multilingual machine translation enables a single model to translate between different languages.  ...  Moreover, extensive analysis of XLM-T on unsupervised syntactic parsing, word alignment, and multilingual classification explains its effectiveness for machine translation.  ...  Transfer learning for low-resource neural machine translation.  ... 
arXiv:2012.15547v1 fatcat:j52fkqqadzdlbohlkenwqndimu
« Previous Showing results 1 — 15 out of 3,488 results