Airbnb breaks down Translation Engine progress

Airbnb分解翻译引擎进度

2022-09-27 10:00 multilingual

本文共441个字,阅读需5分钟

阅读模式 切换至中文

Earlier this month, hospitality and homestay company Airbnb rolled out the expansion of its Translation Engine to support the platform’s reviews feature. It’s the latest expansion in the ambitious translation project, announced last year, that aids communication between guests and hosts across more than 60 languages. First applied to the platform’s messaging feature, Translation Engine will eventually be used for all user-generated content, according to Airbnb officials. We caught up with the Airbnb team to learn just how work has progressed and what comes next. What lessons did you learn from implementing the translation feature into messages that applied to reviews? In what ways was this a similar challenge, and in what ways was it different? The main challenge we faced with implementing Translation Engine on messaging was to ensure best-in-class latency, in order to create a seamless user experience. Our second challenge was to maintain the highest level of translation quality. We were able to satisfactorily solve both challenges. The solutions and learnings from messaging were applied to reviews, making it an even better product experience. Reviews are an important tool for users to research their options and set appropriate expectations. But are there any other less obvious benefits this expansion represents? Reviews are, together with photos and pricing, one of the most critical factors that determine the completion of a booking. In this sense Translation Engine is also a business driver, as it helps convert potential guests into booked guests. This benefits Airbnb’s Hosts worldwide by removing language barriers. And as international travel continues to pick up, Translation Engine is particularly helpful in helping guests read and learn from reviews. Are there any ways that translation for reviews requires different considerations than the Messages app? The main challenge with applying Translation Engine to Reviews was the amount of content that needed to be translated. More than 550 million reviews have been written to date, and they are all now available for our users in their preferred language. With this achievement accomplished, what features are next for the Translation Engine? We will continue to roll out Translation Engine across all Airbnb product surfaces that contain user generated content. In parallel we continue focusing on improving translation quality. Anything you’d like to add? We are excited to have launched an innovative solution that is helping connect millions of hosts and guests globally. Guests can find their perfect stay or experience by being able to read all descriptions and reviews in their preferred language; hosts can reach more guests and increase their bookings; last but not least, hosts and guests can seamlessly communicate with each other.
本月早些时候,酒店和寄宿家庭公司Airbnb推出了其翻译引擎的扩展,以支持平台的评论功能。这是去年宣布的雄心勃勃的翻译项目的最新扩展,该项目有助于嘉宾和主持人之间跨越60多种语言的交流。Airbnb的官员表示,翻译引擎最初应用于平台的消息传递功能,最终将用于所有用户生成的内容。我们采访了Airbnb的团队,了解工作进展如何以及下一步会发生什么。 您从将翻译功能应用于评论的消息中学到了什么?这在哪些方面是一个类似的挑战,在哪些方面是不同的? 我们在消息传递上实施翻译引擎时面临的主要挑战是确保最佳的延迟,以便创建无缝的用户体验。我们的第二个挑战是保持最高水平的翻译质量。我们能够令人满意地解决这两个挑战。从消息传递中获得的解决方案和经验教训被应用到评论中,使其成为更好的产品体验。 评论是用户研究他们的选择和设置适当期望的重要工具。但是,这种扩张是否还有其他不那么明显的好处呢? 评论,连同照片和价格,是决定预订完成的最关键因素之一。从这个意义上说,翻译引擎也是一个业务驱动力,因为它有助于将潜在的客人转化为预订的客人。通过消除语言障碍,Airbnb的全球主机受益匪浅。随着国际旅行的不断增加,翻译引擎在帮助客人阅读和学习评论方面特别有用。 与消息应用程序相比,翻译评论是否需要考虑其他方面? 将翻译引擎应用于评论的主要挑战是需要翻译的内容量。到目前为止,我们已经撰写了超过5.5亿条点评,现在我们的用户可以使用他们的首选语言查看这些点评。 完成此成就后,翻译引擎得下一个功能是什么? 我们将继续在所有包含用户生成内容的Airbnb产品界面上推出翻译引擎。与此同时,我们将继续致力于提高翻译质量。 你还有什么要补充的吗? 我们很高兴推出了一个创新的解决方案,帮助连接全球数百万的主机和客人。客人可以找到他们的完美住宿或经验,能够阅读所有的描述和评论在他们的首选语言;主人可以接触到更多的客人并增加他们的预订;最后但并非最不重要的是,主机和客户机可以彼此无缝地通信。

以上中文文本为机器翻译,存在不同程度偏差和错误,请理解并参考英文原文阅读。

阅读原文