Record Details

Predict translation directions using pupillometry associated with cognitive loading: A machine learning-based approach

Harvard Dataverse (Africa Rice Center, Bioversity International, CCAFS, CIAT, IFPRI, IRRI and WorldFish)

View Archive Info
 
 
Field Value
 
Title Predict translation directions using pupillometry associated with cognitive loading: A machine learning-based approach
 
Identifier https://doi.org/10.7910/DVN/Y1MKH9
 
Creator Chang, Vincent Chieh-Ying
 
Publisher Harvard Dataverse
 
Description Based on such physiological data as pupillometry collected in an eye-tracking experiment, the study has further confirmed the effect of directionality on cognitive loads during L1 and L2 textual translations by novice translators, a phenomenon called ‘translation asymmetry’ suggested by the Inhibitory Control Model, while revealing that machine learning-based approaches can be usefully applied to the field of Cognitive Translation and Interpreting Studies. Directionality was the only factor guiding the eye-tracking experiment where 14 novice translators with the language combination of Chinese and English were recruited to conduct L1 and L2 translations while their pupillometry were recorded. They also filled out a Language and Translation Questionnaire with which categorical data on their demographics were obtained. A nonparametric related-samples Wilcoxon signed rank test on pupillometry verified the effect of directionality, suggested by the model, during bilateral translations, verifying ‘translation asymmetry’ at a textual level. Further, using the pupillometric data, together with the categorical information, the XGBoost machine-learning algorithm yielded a model that could reliably and effectively predict translation directions. The study has shown that translation asymmetry suggested by the model was valid at a textual level, and that machine learning-based approaches can be gainfully applied to Cognitive Translation and Interpreting Studies.
 
Subject Arts and Humanities
pupil size data of the 14 participants during bilateral translations
 
Contributor Chang, Vincent Chieh-Ying