Raithubot: An RLHF-Fine-Tuned Telugu Chatbot for Farmers
摘要
In the current trend of technology, chatbots are a growing and evolving way to reach out to the consumers of the product. Chatbots are the computerized programs designed to solve the human problems by simulating a human-like experience, over the Internet. This paper proposes a chatbot built to simulate the conversation with a farmer who needs help in the day-to-day problems they face in the field of farming. It utilizes large language models and trained with reinforcement learning using a paradigm known as RLHF—Reinforcement Learning with Human Feedback and is further on agricultural datasets. This will later be extended with the application UI which will allow the user to chat or converse their queries and receive responses. The chatbot is further deployed on Google Cloud Platform, allowing it to efficiently handle and fulfill users’ requests with ease and respond to them in their own dialect.