面向低资源语言的文档驱动对话系统跨语言数据增强
Cross-lingual Data Augmentation for Document-grounded Dialog Systems in Low Resource Languages
AI总结:
本文提出跨语言增强模型 CLEM,通过对抗训练检索、Fid 生成器及翻译对齐架构,利用高资源语言增强低资源文档驱动对话系统,并在 DialDoc 2023 竞赛中取得第 4 名。
摘要:
本文提出了一个框架,用于解决文档驱动对话系统(DGDS)中数据稀缺的问题。我们的模型利用高资源语言来增强低资源语言的对话生成能力。具体来说,我们提出了一种新颖的流水线 CLEM(跨语言增强模型),包括对抗训练检索(Retriever 和 Re-ranker)以及 Fid(fusion-in-decoder)生成器。为了进一步利用高资源语言,我们还提出了一种创新架构,通过翻译训练在不同语言之间进行对齐。大量实验结果表明了我们模型的有效性,并在 DialDoc 2023 竞赛中获得了第 4 名。因此,CLEM 可以作为 DGDS 中资源稀缺问题的一种解决方案,并为多语言对齐任务提供有用的指导。
英文摘要:
This paper proposes a framework to address the issue of data scarcity in Document-Grounded Dialogue Systems(DGDS). Our model leverages high-resource languages to enhance the capability of dialogue generation in low-resource languages. Specifically, We present a novel pipeline CLEM (Cross-Lingual Enhanced Model) including adversarial training retrieval (Retriever and Re-ranker), and Fid (fusion-in-decoder) generator. To further leverage high-resource language, we also propose an innovative architecture to conduct alignment across different languages with translated training. Extensive experiment results demonstrate the effectiveness of our model and we achieved 4th place in the DialDoc 2023 Competition. Therefore, CLEM can serve as a solution to resource scarcity in DGDS and provide useful guidance for multi-lingual alignment tasks.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。