Nalazite se na CroRIS probnoj okolini. Ovdje evidentirani podaci neće biti pohranjeni u Informacijskom sustavu znanosti RH. Ako je ovo greška, CroRIS produkcijskoj okolini moguće je pristupi putem poveznice www.croris.hr
izvor podataka: crosbi

The impact of time pressure and source text access in post-editing of neural MT (CROSBI ID 668695)

Neobjavljeno sudjelovanje sa skupa | neobjavljeni prilog sa skupa | međunarodna recenzija

Antunović, Goranka ; Pavlović, Nataša The impact of time pressure and source text access in post-editing of neural MT // Translation research - translator training Budimpešta, Mađarska, 24.05.2018-26.05.2018

Podaci o odgovornosti

Antunović, Goranka ; Pavlović, Nataša

engleski

The impact of time pressure and source text access in post-editing of neural MT

Given the growing use of machine translation (MT) in the language industry, post-editing (PE) is increasingly seen as a standard task for translators. PE has been closely followed by researchers, who have studied its various aspects: productivity (O’Brien 2011 ; Guerberof 2014), overlaps and interactions between translation and PE competences (Almeida & O’Brien 2010 ; Rico & Torrejon 2012 ; Čulo et al. 2014 ; Daems et al. 2017), cognitive effort (O’Brien 2004, 2006, 2007 ; Vieira 2014 ; Lacruz et al. 2014 ; Moorkens et al. 2015 ; Alves et al. 2016), etc. Attempts have been made to develop error typologies and produce PE guidelines (e.g. TAUS 2016) to benefit practitioners and provide a sound basis for teaching PE. Modes of teaching and the place of PE in translator education have also been widely discussed (O’Brien 2002 ; Doherty & Kenny 2014 ; Flanagan & Christensen 2014 ; Koponen 2015 ; Witczak 2016). While the quality of MT output has greatly improved with the recent neural model, PE remains essential in the professional translation workflow, inviting further research. This study of post-editing aims to investigate the influence of time pressure and access to the source text on the type and number of errors that are identified and corrected in a neural MT product. In a three-stage experiment, two groups of subjects – one working with, the other without the ST – will post-edit the Croatian translation of an English ST produced by Google Translate. A comparison of errors detected at different stages and in different conditions is expected to indicate whether particular types of error require more temporal effort (Krings 2001) than others, which types of errors are most likely to remain unnoticed, and how access to the ST influences the completeness and speed of the PE process. The insights should be relevant for translators, translation educators and possibly MT system developers.

machine translation ; post-editing ; time pressure ; access to source text ; number of errors ; error types

Sudjelovanje na skupu financirano je iz sredstava za namjensko financiranje znanstvene i umjetničke aktivnosti na Sveučilištu u Zagrebu u 2018. godini - Filozofski fakultet, šifra projekta 43-921-1038

nije evidentirano

nije evidentirano

nije evidentirano

nije evidentirano

nije evidentirano

Podaci o prilogu

nije evidentirano

nije evidentirano

Podaci o skupu

Translation research - translator training

predavanje

24.05.2018-26.05.2018

Budimpešta, Mađarska

Povezanost rada

Filologija