Authors:

Son Ha

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

This research project explores the development of a specialized chatbot for farmers, leveraging advancements in large language models (LLMs) like ChatGPT by OpenAI. The objective is to create a conversational artificial intelligence (AI)-based system capable of providing practical insights and engaging in near-human conversations to assist farmers in their decision-making processes. The chatbot utilizes natural language processing techniques to interpret farmers' queries. By integrating the text-embedding-ada-002 from OpenAI, the content of various agriculture articles from the Purdue online library is vectorized. Then, the vectorized inputs would be sent to Pinecone for storage. Additionally, by utilizing the search vectors from PineCone and the OpenWeatherAPI, the chatbot can provide personalized recommendations for crop selection, irrigation schedules, and pest control measures. Furthermore, the Word2Vec embedding and vectorization are also proposed and implemented manually as an alternative to OpenAI's model. The chatbot's user interface (UI) is developed using React Native, allowing it to be easily deployed on iOS and Android environments. By providing practical insights and engaging in near- human conversations, a specialized chatbot is an advanced agricultural extension and outreach tool for farmers, aiding their decision-making processes for farm operations.

Source:

Purdue University / 2023

Topics:

No topics listed

Co-authors:

Son Ha