首页|Findings from Autonomous University Barcelona Yields New Data on Machine Transla tion (Evaluating Nmt Using the Non-inferiority Principle)
Findings from Autonomous University Barcelona Yields New Data on Machine Transla tion (Evaluating Nmt Using the Non-inferiority Principle)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning DailyNews Daily News-A new study on Machine Translation is now available. According to news originatingfrom Barcelona, Spain, by NewsRx co rrespondents, research stated, "The aim of this article is to proposea new neur al machine translation (NMT) evaluation method based on the non-inferiority prin ciple. Inorder to do that, we evaluate raw machine translation (MT) in terms of naturalness, which for this researchis defined as not just the lack of fluency errors but also meeting the linguistic expectations of Galician endusers when reading original texts in Galician."Our news journalists obtained a quote from the research from Autonomous Universi ty Barcelona, "Our main objective is, in the first place, to validate the new me thodology presented in our previous study byevaluating an NMT engine from Spani sh into Galician for the social media domain that was retrained witha new Twitt er corpus. This new methodology and NMT engine were applied after analyzing the conclusionsof a pilot survey conducted among Twitter users to evaluate their pe rception of tweets translated fromSpanish into Galician with our NMT engine cre ated with a corpus of tweets. As in our preliminarystudy, our aim is to propose a robust quality approximation method based on the reception parametersof end users' perceptions. This new survey was conducted in December of 2022 with the p articipationof 228 Galician-speaking Twitter users. Among the main changes prop osed are the inclusion of moreinformation about the participant profile, so the non-inferiority principle can be also evaluated accordingto these parameters; the inclusion of a new typology of tweets, the threads; the provision of context bymeans of presenting the tweets in their original display as shown in the Twi tter app; a change in thenumber of tweets evaluated and the number of different questionnaires; the change in the distribution ofthe questionnaires; and the i nclusion of an error classification human evaluation conducted by professionall inguists to correlate the findings. We will present the steps carried out follow ing the conclusions of thepilot study, describe the new study's design, analyze the new findings, and present the final conclusionsregarding the engine and th e evaluation method based on the non-inferiority principle."
BarcelonaSpainEuropeEmerging Techn ologiesMachine LearningMachine TranslationAutonomous University Barcelona
NETL
NSTL
