![]() ![]() The consistent groups come from the original training data the inconsistent groups are obtained by sampling round-trip translations for each isolated sentence. It is trained as a monolingual sequence-to-sequence model that maps inconsistent groups of sentences into consistent ones. For training, the DocRepair model requires only monolingual document-level data in the target language. DocRepair performs automatic post-editing on a sequence of sentence-level translations, refining translations of sentences in context of each other. We propose a monolingual DocRepair model to correct inconsistencies between sentence-level translations. However, when put in context, these translations may end up being inconsistent with each other. Publisher = "Association for Computational Linguistics",Ībstract = "Modern sentence-level NMT systems often produce plausible translations of isolated sentences. Cite (Informal): Context-Aware Monolingual Repair for Neural Machine Translation (Voita et al., EMNLP 2019) Copy Citation: BibTeX Markdown MODS XML Endnote More options… PDF: Code = "Context-Aware Monolingual Repair for Neural Machine Translation",īooktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)", Association for Computational Linguistics. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 877–886, Hong Kong, China. Context-Aware Monolingual Repair for Neural Machine Translation. ![]() | IJCNLP SIG: SIGDAT Publisher: Association for Computational Linguistics Note: Pages: 877–886 Language: URL: DOI: 10.18653/v1/D19-1081 Bibkey: voita-etal-2019-context Cite (ACL): Elena Voita, Rico Sennrich, and Ivan Titov. Anthology ID: D19-1081 Volume: Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP) Month: November Year: 2019 Address: Hong Kong, China Venues: EMNLP Moreover, we analyze which discourse phenomena are hard to capture using monolingual data only. We also conduct a human evaluation and show a strong preference of the annotators to corrected translations over the baseline ones. We show that this approach successfully imitates inconsistencies we aim to fix: using contrastive evaluation, we show large improvements in the translation of several contextual phenomena in an English-Russian translation task, as well as improvements in the BLEU score. ![]() Let me know how I can help.Abstract Modern sentence-level NMT systems often produce plausible translations of isolated sentences. If you’re looking for a sensitivity reader but I’m not the best match for your text I would love to help you find another, whether you’re looking for a different language, different topic, or just a different perspective. Please email me at artemisqueerterpretercom or fill out this contact form if you’d like to work together. In the same way that you wouldn’t write a scientific text without help from scientists in the field, sensitivity readers are members of specific communities who can help make your text or product more inclusive and friendly for community members. Sensitivity readers are a little-known but incredibly valuable specialized resource. I am also a sensitivity reader for non-binary, trans, and queer topics both in English and in Spanish. ![]() I am available for bilingual editing, where I would review a translated text and use the original as a reference, and for monolingual editing, where I review a text either in English or in Spanish as an original. As a linguist and a translator, I have been editing texts both in English and in Spanish for years and would be happy to edit for you. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |