Deep Auto Encoder based Chatbot for Discrete Math Course
Sibghat Ullah, Momina Liaqat, Aqsa Asif, Asifullah Khan, Umair Aslam, Hifsa Asif · 2022
Chatbots are used for the facilitation of human agents by decreasing their involvement in manually done tasks. This Chatbot is developed for students of discrete mathematics. The dataset for this experiment is accumulated from 3 books of discrete mathematics. Our model is a variant of transformer known as Bert-base-NLI-mean-tokens. Bidirectional Encoder Representations from Transformers (BERT). SBERT library is used for the generation of queryresponse embeddings. This model is fine-tuned according to our dataset and provided this model to students with android based mobile App. This model results in increase in the knowledge of students. The same approach will be implemented for other subjects as well.