Design and Evaluation of a Multilingual Chatbot for Interactive English Language Learning

The study introduces an innovative Multilingual Chatbot designed for interactive English language learning, leveraging a diverse dataset from various conversational platforms to ensure adaptability across multiple languages while prioritizing English proficiency. Notably, the system integrates proficiency level labeling to personalize responses based on users' language skills, contributing to a more effective and adaptive learning experience. This personalized approach enhances the chatbot's ability to provide tailored language learning experiences, catering to users with different proficiency levels. The proposed chatbot architecture offers a structured and integrated framework, seamlessly connecting key components such as Natural Language Understanding, Dialog Management, and Natural Language Generation. Particularly noteworthy is the use of Support Vector Machines (SVM) for intent recognition, achieving an impressive accuracy of 98.3%, with SVM-based language classification demonstrating superior performance in identifying user languages. Overall, these advancements represent significant contributions to the field of language education technology, promising more dynamic and contextually relevant interactions for learners.