arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2210.03070cs.CL
收藏

大规模多语言机器翻译中的毒性问题

Toxicity in Multilingual Machine Translation at Scale

Meta AI;Universitat Politècnica de Catalunya(加泰罗尼亚理工大学)

Marta R. Costa-jussà, Eric Smith, Christophe Ropers, Daniel Licht, Jean Maillard, Javier Ferrando, Carlos Escolano

AI总结:

本文评估了将HOLISTICBIAS数据集从英语翻译至164种语言时出现的增加毒性问题,发现低资源语言及特定人口统计轴毒性较高,并通过源贡献分析提出减少毒性的建议。

摘要:

机器翻译系统会产生不同类型的错误,其中一些因其对用户造成的特定负面影响而被视为严重或灾难性错误。本文重点关注一种严重错误:增加的毒性。我们评估并分析了将大型评估数据集(HOLISTICBIAS,超过472k个句子,覆盖13个人口统计轴)从英语翻译成164种语言时增加的毒性。自动毒性评估显示,不同语言中增加的毒性在0%到5%之间变化。增加毒性最多的输出语言往往是低资源语言,而增加毒性最严重的人口统计轴包括性取向、性别与生理性别,以及能力。我们还对8个翻译方向的子集进行了人工评估,证实了真实增加毒性的普遍性。我们使用一种衡量源语言对翻译贡献程度的方法来解释导致毒性的原因,其中低源贡献意味着幻觉。利用输入归因使我们能够解释毒性,因为源贡献与所研究84%语言的毒性显著相关。基于我们的发现,我们建议通过整理训练数据以避免误译、缓解幻觉并检查不稳定翻译来减少增加的毒性。

英文摘要:

Machine Translation systems can produce different types of errors, some of which are characterized as critical or catastrophic due to the specific negative impact that they can have on users. In this paper we focus on one type of critical error: added toxicity. We evaluate and analyze added toxicity when translating a large evaluation dataset (HOLISTICBIAS, over 472k sentences, covering 13 demographic axes) from English into 164 languages. An automatic toxicity evaluation shows that added toxicity across languages varies from 0% to 5%. The output languages with the most added toxicity tend to be low-resource ones, and the demographic axes with the most added toxicity include sexual orientation, gender and sex, and ability. We also perform human evaluation on a subset of 8 translation directions, confirming the prevalence of true added toxicity. We use a measurement of the amount of source contribution to the translation, where a low source contribution implies hallucination, to interpret what causes toxicity. Making use of the input attributions allows us to explain toxicity, because the source contributions significantly correlate with toxicity for 84% of languages studied. Given our findings, our recommendations to reduce added toxicity are to curate training data to avoid mistranslations, mitigate hallucination and check unstable translations.

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。

↑