Abstract:
Library is a knowledge warehouse and various long past references can be found in it. Students, professors, kids, and
adults are regularly encouraged to visit the library as it provides a conducive environment for building the habit of reading books and
improving individual critical-thinking skills. As technology is getting more and more advanced nowadays, some common problems
faced by the librarians can be replaced by machines. For instance, the librarians may not be available all the time at the counter;
reduction of physical contact due to Covid19 infection et cetera, machines can take over the librarians’ roles to handle the tasks. In this
paper, an Artificial Intelligence (AI) chatbot is proposed and implemented on mobile application to answer library-related questions.
Bidirectional Encoder Representations from Transformers (BERT) algorithm is employed to classify the intent of the user’s messages.
Besides, many existing chatbot applications support only the text input. This paper proposes a speech-to-text recognition feature to
enable both text and voice input. If there are any queries that cannot be solved by the chatbot system, it will store the queries in
the database and the library admins can filter the queries and upload new training data for the AI model to cover a wider range of questions.
