Towards Speech Dialogue Translation Mediating Speakers of Different Languages

Shuichiro Shimizu, Chenhui Chu, Sheng Li, Sadao Kurohashi
5/16/2023
cs.CL

No official code found. Best implementation: ku-nlp/speechBSD (3 stars, last commit 2024-02-07).

Abstract

We present a new task, speech dialogue translation mediating speakers of different languages. We construct the SpeechBSD dataset for the task and conduct baseline experiments. Furthermore, we consider context to be an important aspect that needs to be addressed in this task and propose two ways of utilizing context, namely monolingual context and bilingual context. We conduct cascaded speech translation experiments using Whisper and mBART, and show that bilingual context performs better in our settings.

View on arXivView PDF

Code Implementations(1)

ku-nlp/speechBSDCommunity100%

An extension of the BSD corpus with audio and speaker attribute information

30May 25, 20222 years agoNOASSERTION

Similar Papers

A Multi-task Multi-stage Transitional Training Framework for Neural Chat Translation

86% match2023

IndoToD: A Multi-Domain Indonesian Benchmark For End-to-End Task-Oriented Dialogue Systems

84% match1 repo22023

Multi3WOZ: A Multilingual, Multi-Domain, Multi-Parallel Dataset for Training and Evaluating Culturally Adapted Task-Oriented Dialog Systems

83% match1 repo172023

Is Translation Helpful? An Empirical Analysis of Cross-Lingual Transfer in Low-Resource Dialog Generation

83% match2023

Cross-lingual QA: A Key to Unlocking In-context Cross-lingual Performance

83% match2023

Context-aware Neural Machine Translation for English-Japanese Business Scene Dialogues

82% match2023

Cite this paper

@article{shimizu2023towards,
  title  = {Towards Speech Dialogue Translation Mediating Speakers of Different Languages},
  author = {Shuichiro Shimizu and Chenhui Chu and Sheng Li and Sadao Kurohashi},
  year   = {2023},
  eprint = {2305.09210},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2305.09210v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2305.09210.svg)](https://codeofpaper.com/paper/2305.09210)

Discussion