---
title: "Open-Source Large Language Models as Multilingual Crowdworkers: Synthesizing Open-Domain Dialogues in Several Languages With No Examples in Targets and No Machine Translation"
canonical_url: "https://www.modelscope.cn/papers/123482"
md_url: "https://www.modelscope.cn/papers/123482.md"
arxiv_id: 2503.03462
published: 2025-03-05
last_updated: 2025-03-05
authors:
  - "Ahmed Njifenjou"
  - "Virgile Sucal"
  - "Bassam Jabaian"
  - "Fabrice Lefèvre"
model_name: "MOUD (Multilingual Open-domain Unnatural Dialogue Dataset)"
model_developer: "阿维尼翁信息实验室 (LIA)，CERI-阿维尼翁大学"
domain:
  - "自然语言处理"
  - "深度学习"
  - "机器学习"
type:
  - "自然语言处理"
  - "深度学习"
  - "机器学习"
  - "Computation and Language (cs.CL)"
  - "Artificial Intelligence (cs.AI)"
  - "Human-Computer Interaction (cs.HC)"
  - "Machine Learning (cs.LG)"
arxiv_url: "https://arxiv.org/abs/2503.03462"
pdf_url: "https://arxiv.org/pdf/2503.03462.pdf"
---

# Open-Source Large Language Models as Multilingual Crowdworkers: Synthesizing Open-Domain Dialogues in Several Languages With No Examples in Targets and No Machine Translation

> The prevailing paradigm in the domain of Open-Domain Dialogue agents predominantly focuses on the English language, encompassing both models and datasets. Furthermore, the financial and temporal investments required for crowdsourcing such datasets for…

「Open-Source Large Language Models as Multilingual Crowdworkers: Synthesizing Open-Domain Dialogues in Several Languages With No Examples in Targets and No Machine Translation」是 ModelScope 魔搭社区收录的论文，arXiv 2503.03462，作者为 Ahmed Njifenjou, Virgile Sucal, Bassam Jabaian et al.，发表于 2025-03-05，属于 自然语言处理、深度学习、机器学习 领域。

- **ArXiv**: 2503.03462
- **Published**: 2025-03-05
- **Authors**: Ahmed Njifenjou, Virgile Sucal, Bassam Jabaian, Fabrice Lefèvre
- **Model**: MOUD (Multilingual Open-domain Unnatural Dialogue Dataset)
- **Developer**: 阿维尼翁信息实验室 (LIA)，CERI-阿维尼翁大学
- **Domain**: 自然语言处理, 深度学习, 机器学习
- **ArXiv URL**: https://arxiv.org/abs/2503.03462
- **PDF**: https://arxiv.org/pdf/2503.03462.pdf

Source: https://www.modelscope.cn/papers/123482

---

> 利用多语言大模型生成多样化开放域对话数据：超越传统机器翻译与目标语言示例

## 摘要

本文提出了一种利用多语言指令跟随的大型语言模型（LLMs）生成多种目标语言的开放域对话数据集的方法，旨在解决现有对话数据集主要集中在英语和其他少数语言的问题。通过这种方法，研究人员能够在不使用机器翻译和目标语言示例的情况下，生成符合特定语言特点的对话数据。该方法应用了PersonaChat数据集，并生成了一个包含28种目标语言的多语言开放域对话数据集MOUD。为了提高对话的真实性和多样性，作者引入了“共同背景”（common ground）和不同类型的“话语事件”（speech events）。实验结果表明，生成的数据在自动评估指标、语法分析和人类评价方面表现良好。此外，作者还进行了基线应用，评估了浅层微调模型在某些目标语言上的性能。

## Abstract

The prevailing paradigm in the domain of Open-Domain Dialogue agents predominantly focuses on the English language, encompassing both models and datasets. Furthermore, the financial and temporal investments required for crowdsourcing such datasets for finetuning are substantial, particularly when multiple languages are involved. Fortunately, advancements in Large Language Models (LLMs) have unveiled a plethora of possibilities across diverse tasks. Specifically, instruction-tuning has enabled LLMs to execute tasks based on natural language instructions, occasionally surpassing the performance of human crowdworkers. Additionally, these models possess the capability to function in various languages within a single thread. Consequently, to generate new samples in different languages, we propose leveraging these capabilities to replicate the data collection process. We introduce a pipeline for generating Open-Domain Dialogue data in multiple Target Languages using LLMs, with demonstrations provided in a unique Source Language. By eschewing explicit Machine Translation in this approach, we enhance the adherence to language-specific nuances. We apply this methodology to the PersonaChat dataset. To enhance the openness of generated dialogues and mimic real life scenarii, we added the notion of speech events corresponding to the type of conversation the speakers are involved in and also that of common ground which represents the premises of a conversation.