Automatic assessment on English-to-Chinese translation for personalized learning

Abstract

This project aims to investigate automatic assessment techniques for students’ English to Chinese translation writings (such as ESL English). In English to Chinese translations, students are required to submit the translated Chinese answers for marking. This project only aims to grade (assess) the students’ translations in Chinese (not automatic translation techniques). Manual marking of students translation writings is difficult as it is time consuming and difficult to grade. To assess the students' translation writings automatically, we need to look into different aspects such as rhetoric (readability, vocabulary, grammar, etc.), organization (conjunctions, etc.) and content (vocabulary related to the topic) of the translations. In this project, we will look into the different techniques in grading students’ translations automatically, especially to provide feedback on the translations. Currently automatic assessment techniques are mainly categorized as supervised such as Support Vector Machine and unsupervised such as Latent Semantic Analysis.