Transforming Student Advising in Smart Cities: A Deep Learning Conversational AI Chatbot

Citizens, including students, teachers and parents can all be significantly involved in the smart cities, as smart education is one of the main key drivers in developing smart citizens. Indeed, high school plays an essential role not only in shaping students’ future career, but also in contributing to the development of smart citizens. High school students who benefit from college-career advising will be more prepared and motivated to universities compared to students from other schools. Because the availability of college-career guidance position can vary in schools, students might not get the proper guidance equally. Therefore, in this paper a novel deep learning based chatbot is implemented by using the Seq2Seq model trained on 2944 students’ inquiries for taking the role of a college-career guidance in high schools. In this model a BiLSTM layer is configured with 400 units in each direction of LSTM layer and a single dense layer with Softmax activation is included in the decoder component. The model, when tested with the Adam optimizer, demonstrated a notable performance enhancement over the SGD optimizer. Evaluation with ROUGE-N metrics indicated a high precision score of 92% in the ROUGE-1 measure.