Modeling Bilingual Conversational Characteristics for Neural Chat Translation

  title={Modeling Bilingual Conversational Characteristics for Neural Chat Translation},
  author={Yunlong Liang and Fandong Meng and Yufeng Chen and Jinan Xu and Jie Zhou},
  • Yunlong Liang, Fandong Meng, +2 authors Jie Zhou
  • Published 2021
  • Computer Science
  • ArXiv
Neural chat translation aims to translate bilingual conversational text, which has a broad application in international exchanges and cooperation. Despite the impressive performance of sentence-level and context-aware Neural Machine Translation (NMT), there still remain challenges to translate bilingual conversational text due to its inherent characteristics such as role preference, dialogue coherence, and translation consistency. In this paper, we aim to promote the translation quality of… Expand
Towards Making the Most of Dialogue Characteristics for Neural Chat Translation
  • Yunlong Liang, Chulun Zhou, +4 authors Jie Zhou
  • Computer Science
  • ArXiv
  • 2021
This paper proposes to promote the chat translation by introducing the modeling of dialogue characteristics into the NCT model, and designs four auxiliary tasks including monolingual response generation, cross-lingual responsegeneration, next utterance discrimination, and speaker identification. Expand


Contextual Neural Model for Translating Bilingual Multi-Speaker Conversations
This work proposes the task of translating Bilingual Multi-Speaker Conversations, and explores neural architectures which exploit both source and target-side conversation histories for this task, including BLEU and manual evaluation. Expand
Context-Aware Monolingual Repair for Neural Machine Translation
This work proposes a monolingual DocRepair model to correct inconsistencies between sentence-level translations, and shows large improvements in the translation of several contextual phenomena in an English-Russian translation task, as well as Improvements in the BLEU score. Expand
Modeling Coherence for Discourse Neural Machine Translation
This paper proposes to use discourse context and reward to refine the translation quality from the discourse perspective, and generates the translation of individual sentences at first and trains the model to learn the policy that produces discourse coherent text by a reward teacher. Expand
Document-aligned Japanese-English Conversation Parallel Corpus
This work presents a document-aligned Japanese-English conversation corpus, including balanced, high-quality business conversation data for tuning and testing, and manually identifies the main areas where SL MT fails to produce adequate translations in lack of context. Expand
The University of Maryland’s Submissions to the WMT20 Chat Translation Task: Searching for More Data to Adapt Discourse-Aware Neural Machine Translation
The University of Maryland’s submissions to the WMT20 Shared Task on Chat Translation focus on translating agent-side utterances from English to German using an off-the-shelf BPE-based standard transformer model trained with WMT17 news and fine-tuned it with the provided in-domain training data. Expand
Findings of the WMT 2020 Shared Task on Chat Translation
The first edition of the WMT shared task on chat translation consisted of translating bilingual conversational text, in particular customer support chats for the English-German language pair (English agent, German customer). Expand
Automatically Annotate TV Series Subtitles for Dialogue Corpus Construction
  • Leilan Zhang, Qiang Zhou
  • Computer Science
  • 2019 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC)
  • 2019
An improved method to automatically annotate bilingual TV subtitles with speaker and scene tags using their corresponding scripts and can achieve an accuracy of 94.62% on speaker tag annotation, improving nearly 12% on the previous state-of-the-art result. Expand
Automatic Construction of Discourse Corpora for Dialogue Translation
A novel approach is proposed to automatically construct parallel discourse corpus for dialogue machine translation by mapping monolingual discourse to bilingual texts via an information retrieval approach and integrating speaker information into the translation. Expand
Evaluating Discourse Phenomena in Neural Machine Translation
This article presents hand-crafted, discourse test sets, designed to test the recently proposed multi-encoder NMT models’ ability to exploit previous source and target sentences, and explores a novel way of exploiting context from the previous sentence. Expand
Tencent AI Lab Machine Translation Systems for WMT20 Chat Translation Task
This paper proposed a hybrid data selection method to select high-quality and in-domain sentences from out-of-domain data and explored to transfer general knowledge from four different pre-training language models to the downstream translation task. Expand