Generative AI for Researching Archaeology 

This project presents a Generative AI system for archaeological research using Retrieval-Augmented Generation (RAG). Leveraging advanced text-chunking strategies and metadata extraction, it indexes document segments into vector representations to enable precise retrieval and automated Q&A. Evaluated on a custom benchmark of 125 questions, the system achieves an F1-score of 93.44%, demonstrating significant improvements in the efficiency and accuracy of archaeological data analysis.

Click “Basic Info”, “Methodology” and “Results” in menubar for more information