An Interactive Learning Platform with Machine Translation for Practicing Text-Based Conversational English
Andre Rusli, Makoto Shishido · 2022 Joint 12th International Conference on Soft Computing and Intelligent Systems and 23rd International Symposium on Advanced Intelligent Systems (SCIS&ISIS) · 2022
Text-based conversation, such as chatting or texting, has become an essential means of written communication. This paper investigates the importance of practicing textual conversations for foreign language learners, focusing on Ll Japanese speakers learning English as a foreign language. We proposed a web-based interactive platform that utilizes neural machine translation (NMT) to support language learners practice text-based conversational English. Our proposed system enables the learners to practice texting with other users and conduct exercises for self-practice and assessments. The platform provides immediate feedback required by learners and also allows teachers to verify the assessment result, implementing the human-in-the-loop framework. The verified feedback from the teachers can then be used to improve the evaluation translation, especially for NMT models using metrics, such as BLEU, that depend on reference translations to calculate the scores.