Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

DOCmT5: Document-Level Pretraining of Multilingual Language Models

About

In this paper, we introduce DOCmT5, a multilingual sequence-to-sequence language model pretrained with large scale parallel documents. While previous approaches have focused on leveraging sentence-level parallel data, we try to build a general-purpose pretrained model that can understand and generate long documents. We propose a simple and effective pretraining objective - Document reordering Machine Translation (DrMT), in which the input documents that are shuffled and masked need to be translated. DrMT brings consistent improvements over strong baselines on a variety of document-level generation tasks, including over 12 BLEU points for seen-language-pair document-level MT, over 7 BLEU points for unseen-language-pair document-level MT and over 3 ROUGE-1 points for seen-language-pair cross-lingual summarization. We achieve state-of-the-art (SOTA) on WMT20 De-En and IWSLT15 Zh-En document translation tasks. We also conduct extensive analysis on various factors for document pretraining, including (1) The effects of pretraining data quality and (2) The effects of combining mono-lingual and cross-lingual pretraining. We plan to make our model checkpoints publicly available.

Chia-Hsuan Lee, Aditya Siddhant, Viresh Ratnakar, Melvin Johnson• 2021

Related benchmarks

TaskDatasetResultRank
Document-Level Machine TranslationTED15 Zh-En 2010-2013 (test)
d-BLEU31.4
16
Document-Level Machine TranslationWMT20 De-En (test)
d-BLEU44.73
12
Cross-lingual SummarizationWikilingua Ru-En Seen Languages GEM
ROUGE-133.56
10
Cross-lingual SummarizationWikilingua Tr-En Seen Languages GEM
ROUGE-137.66
10
Cross-lingual SummarizationWikilingua Vi-En Seen Languages GEM
ROUGE-133.29
10
Cross-lingual SummarizationWikilingua Es-En Seen Languages GEM
ROUGE-136.79
10
Cross-lingual SummarizationWikilingua Fr-En
ROUGE-136.28
9
Cross-lingual SummarizationWikilingua Id-En
ROUGE-135.15
9
Cross-lingual SummarizationWikilingua Hi-En
ROUGE-134.16
9
Machine TranslationWMT20 JA-EN (test)
BLEU19.17
8
Showing 10 of 13 rows

Other info

Follow for update