Multilingual Indian COVID-19 Chatbot

Albeit the COVID-19 era has ended, common people are ravaged by lingering misconceptions about the pre- and post-COVID effects. Chatbot is a computer program that simulates human conversation. Chatbots communicate with users via text or voice using natural language. This paper presents the development of a MILIC-19 chatbot to augur user-friendly interfaces for Indians, where there are many official languages. We have integrated the Google Translate API, allowing our chatbot to converse in 19 Indian languages. The MILIC-19 chatbot application responds to user queries by matching them with the keywords in the database to retrieve an appropriate response from the database. Handling Question-Answer tasks is configured in two layers, each serving a unique purpose in providing the best results. The first layer is an information retrieval model that leverages the power of Term Frequency-Inverse Document Frequency and Smooth Inverse Frequency to calculate inverted indices and Best Match to retrieve the top 15 articles related to the query. The second layer is a Bio-BERT model trained on the Stanford QA dataset. This combination of techniques enables the model to effectively provide relevant information by prioritizing the most important documents based on the frequency of the terms in the query.