Towards highly adaptive Edu-Chatbot
Tarek Ait Baha, Mohamed El Hajji, Youssef Es-Saady, Hammou Fadili
Abstract
Conversational Agents are widely used in different domains to automate tasks and help to improve user experience. In recent decades, AI systems, thanks to deep learning methods and Natural Language Processing (NLP) approaches, can interact with users, understand their needs, map their preferences and recommend an appropriate action with no human intervention. However, chatbots in the education field have received limited attention. In this work, we use Xatkit, a chatbot development framework, for the definition of our Chatbot and propose an Encoder-Decoder framework for intent recognition. For the encoder, we encode utterances as context representations using bidirectional transformer (CamemBERT). For the decoder, we use an intent classification decoder to detect the student’s intent. Our chatbot will be tested in the field of education to improve and simplify teaching for professors and learning for students as well as reducing faculty burnout and raising the speed of comprehension.