AGRI-TECH

Purpose

The purpose of the Agri-Tech chatbot is to empower urban household farmers in Cambodia by providing them with easy access to reliable and localized farming advice in the Khmer language.

Features

• Khmer Language Support: The chatbot communicates fluently in Khmer, making it accessible to local farmers.
• Disease Detection: The chatbot can identify common plant diseases based on user descriptions and images.
• Reliable Information: All advice is sourced from official Ministry of Agriculture documents, ensuring accuracy and trustworthiness.
• Contextual Understanding: The chatbot can understand and respond to a wide range of agriculture-related queries specific to urban household farming.
• User-Friendly Interface: Designed for ease of use, even for those with limited technical skills.
• 24/7 Availability: Users can ask questions anytime and receive instant responses.

Technology Stack

Dataset Preparation

• We collect academic documents from the Ministry's eLibrary.
• The text is cleaned and pre-processed (removing noise, formatting, and irrelevant data) to ensure high-quality information.

Knowledge Base Creation

• Cleaning text from PDF using Khmer text OCR
• The cleaned text is stored in a vector database after being converted into embeddings.
• This allows the system to quickly search and retrieve the most relevant chunks of information when a query is asked.

User Query

• A user submits a question through the system interface.
• The query is also transformed into embeddings.

Retrieval Step

• The system searches the vector database for the most relevant documents or text chunks related to the query.
• These retrieved results provide factual grounding for the model.

Generation with Gemini API

• The retrieved documents are combined with the user's query and sent to the Gemini API.
• Gemini then generates a well-structured, context-aware, and accurate answer based on both the query and the retrieved data.

Final Response

• The system delivers a precise, grounded response back to the user.
• This ensures that answers are not only fluent but also reliable and tied to the eLibrary dataset.

System Architecture