Summary
The Kentucky Agricultural Extension Program, part of the University of Kentucky College of Agriculture, Food, and Environment (CAFE), is designed to extend resources, services, and research-based knowledge to farmers, communities, and families across the commonwealth. Working in collaboration with the se statewide entities, we created an application that helps users across the commonwealth find relevant information, navigate the CAFE website and publications, and provide answers to questions in a conversational manner.
This is possible thanks to our recently developed tools leveraging RAG techniques. We can easily create vector datastores from papers, webpages, and data sets and retrieve the nearest semantically related and contextually relevant documents for a particular search query. The LLM can then use these records to enhance the prompt by giving it the relevant context needed to answer queries in a fashion that is consistent with the information stored in the vector datastore.
Demo
This project is still in development. A demo is available: https://rag.ai.uky.edu/agriguide
Dataset
The knowledge base of this model was informed by the University of Kentucky’s College of Agriculture, Food and Environment website and publication database.
Model
This application is supported by LLM Factory, the base model being used in Llama 3 8B
Send email to ai@uky.edu for more information.