Reality Series: Machine translation quality estimation
In this episode of Reality Series, we focus on the myths and realities of machine translation quality estimation, hosted by Mei Zheng, data scientist, and Alex Yanishevsky, director of AI and machine translation.Over the past decade, the evolution of machine translation has significantly changed the industry. With that, there are a lot of factors to consider when we look at the quality estimation of machine translation – and this discussion is dives right in, addressing ways that quality is estimated and calculated, and questions like:- Does a high score mean high quality?- What role do semantics play in edit rate?- How does a fuzzy match affect translators?- How (and when) should we edit and translate profanities?Jump into the Episode[01:39] Busting myth #1: Is 90% quality good enough to publish?[09:58] Busting myth #2: "MT stinks!" [17:51] Busting myth #3: There is one MTQE. [24:13] Q&ASample translations discussed during Myth #2en: “Dr. Smith was found guilty of keeping a protected animal in the Atherton Magistrates Court after being charged with removing a scrub python from a resident’s property. She then went through the legal process to appeal the court ruling.”zh: 史密斯博士在阿瑟顿地方法院被裁定饲养受保护动物罪，因为她被指控从一个居民的财产中移走一条灌丛蟒蛇。随后，她通过法律程序对法院的裁决提出上诉。es: La Dra. Smith fue declarada culpable de tenencia de un animal protegido en el Tribunal de Magistrados de Atherton tras ser acusada de retirar una pitón de matorral de la propiedad de un residente. A continuación recurrió la sentencia judicial.ru: Доктор Смит была признана виновной в содержании охраняемого животного в магистратском суде Атертона после того, как ей было предъявлено обвинение в том, что она убрала камышового питона с участка жителя. Затем она прошла через судебный процесс, чтобы обжаловать решение суда.Resources and links:Smartling Website