Developement of a conversational agent for the university using LLM models

Abstract

A chatbot is a computer program designed to facilitate conversations between hu- mans and machines. It can be used across various platforms, such as messaging apps and virtual assistants. Over the years, chatbots have evolved significantly, transitioning from being mere entertainment to performing important tasks. When creating a chatbot, several design considerations should be taken into account, including its purpose, target audience, communication channels, conversational flow, and the need for testing and it- erative improvements to ensure accuracy and user-friendliness. Based on their domain, model, and conversation style, chatbots can be categorized into various types, including customer service, sales, informational, personal assistant, entertainment, health, and edu- cational chatbots. Each type serves a specific function and caters to the needs of different user groups. Despite technological advancements, chatbot technology still faces several challenges. These challenges include contextual understanding, seamless integration with backend systems, personalization for individual users, ensuring security, and gaining user acceptance and trust. This dissertation aims to develop an information chatbot that can answer different questions related to university studies. The latter can be installed on the university home page.