Towards Deep Learning-Powered Chatbot for Translation Learning

As a consequence of the recent advances in artificial intelligence and educational technologies, the education sector is witnessing significant changes and transformations through massive use of intelligent systems with the goal to assist students in their learning experience and teachers in delivering academic knowledge in a better way while reducing burnout and stress. Communication, knowledge acquisition, and learning are now possible anytime and anywhere through modern technologies. However, learning translation is a process that requires continuous effort, enthusiasm, and motivation from both students and instructors. While some universities adopt traditional and outdated approaches to translation instruction, others use more innovative ways. Within this context and in order to foster a student-centered translation learning approach, our work focuses on taking advantage of recent advances in machine learning to develop a chatbot to help language learners, especially translators, develop their skills by having a conversation with a chatbot and getting the appropriate Arabic translations of English sentences under study. A general framework is proposed, and a first prototype is developed using a part of bilingual corpora. Another focus of this study is the preprocessing phase used to create a sentence-based paired dataset to train the machine learning model from bilingual corpora. The preliminary results are promising.