A Bengali Physics Chatbot Utilizing Sequence-to-Sequence Model with Beam Search
Javed Hossain · 2024
Chatbots, also known as talkbots or interactive agents, are software applications designed to facilitate communication between humans and machines. While most students in Bangladesh currently waste their valuable time engaging in casual conversations with each other, using a Bengali educational chatbot for such interactions would provide them with a more beneficial use of their time by providing hand note-wise answers for physics. The challenge faced by a Bengali chatbot lies in the necessity of creating a dataset in Bengali, given the scarcity of available Bengali datasets online. We have developed a physics-focused educational chatbot, akin to personalized hand notes, offering comprehensive answers to aid students in learning and revising complex physics concepts. The development of chatbots involves diverse techniques, including rule-based systems, machine learning-based approaches, and generative models. Our proposed system implements the seq2seq model with a beam search decoding algorithm, featuring BiLSTM layers for enhanced context understanding and efficient exploration of potential output sequences. This approach outperforms traditional seq2seq models, ensuring higher accuracy and delivering high-quality responses. Finally, we obtained the highest BLEU1, BLEU2, BLEU3, and BLEU4 scores, respectively, of 0.52, 0.41, 0.29, and 0.23.