摘要
由一名新闻记者-机器人与机器学习的工作人员新闻编辑每日新闻-调查人员发布了关于马翻译的新报道。根据NewsRx记者从沙特国王大学发回的新闻报道，研究表明：“利用预先训练的语言模型(PLMs)进行神经机器翻译的进展显示出改善不同语言之间翻译质量的前景。”我们的新闻编辑从沙特国王大学的研究中获得了一句话：“然而，从英语翻译成具有复杂形态的语言，如阿拉伯语，仍然具有挑战性。”本研究调查了当前最先进的PLM在不同文本领域从英语翻译到阿拉伯语时的错误模式。通过使用自动指标（chrF、BER TScore、COMET）的实证分析和多维质量指标(MQM)框架的手动评估，我们比较了Google Translate和五个PLM（赫尔辛基、马雷法、Fa Cebook、GPT-3.5-turbo和GPT-4）。根据新闻编辑的说法，这项研究的结论是：“关键的发现为当前PLM在处理阿拉伯语语法和词汇方面的局限性提供了宝贵的见解，同时也为未来提高英语-阿拉伯语机器翻译能力和可及性提供了信息。”
Abstract
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Investigators publish new report on ma chine translation. According to news reporting originating from King Saud Univer sity by NewsRx correspondents, research stated, "Advances in neural machine tran slation utilizing pretrained language models (PLMs) have shown promise in improv ing the translation quality between diverse languages." Our news editors obtained a quote from the research from King Saud University: " However, translation from English to languages with complex morphology, such as Arabic, remains challenging. This study investigated the prevailing error patter ns of state-of-the-art PLMs when translating from English to Arabic across diffe rent text domains. Through empirical analysis using automatic metrics (chrF, BER TScore, COMET) and manual evaluation with the Multidimensional Quality Metrics ( MQM) framework, we compared Google Translate and five PLMs (Helsinki, Marefa, Fa cebook, GPT-3.5-turbo, and GPT-4)." According to the news editors, the research concluded: "Key findings provide val uable insights into current PLM limitations in handling aspects of Arabic gramma r and vocabulary while also informing future improvements for advancing English- Arabic machine translation capabilities and accessibility."
NETL
NSTL