Implementation of an intelligent virtual assistant based on LLM models for irrigation optimization

This paper explores the application of new AI capabilities and Large Languages Models (LLMs) in agriculture by developing an intelligent virtual assistant. The assistant aims to provide essential information on crops, vulnerabilities, and optimization strategies. The study focuses on implementing a robust pipeline architecture that trains LLMs for practical scenarios. This architecture ensures data and demonstrates substantial domain knowledge. Central to this approach is the Retrieval-Augmented Generation (RAG) technology, which organizes information into a streamlined format for efficient retrieval. Unlike fine-tuning methods, RAG separates base knowledge from the model, enabling a faster and more adapted pipeline. This proposal transforms retrieved information into an accessible format, facilitating rapid access to comprehensive agricultural expertise.