Chatting with Champions: An AI Conversational Football Historian

Want to grill a legend like Cristiano Ronaldo or Lionel Messi with questions? This project builds an AI that lets you chat with the 35 greatest footballers EVER! Ask anything about their careers, rivalries, or iconic moments, and this AI, powered by cutting-edge tech, will answer like a true football mastermind. Get ready to rewrite fan history!

VIEW PROJECT

Overview

This project delves into the world of football legends, offering a unique and interactive way to explore their careers and achievements. By leveraging cutting-edge AI techniques, it empowers users to ask insightful questions about any of the top 35 footballers of all time, as identified by a renowned source like The Guardian.

Tech Stack Used

Here’s a breakdown of the system’s core elements:

  • Data Foundation: The project starts by building a solid data foundation. It gathers information on the 35 greatest players through web scraping techniques, specifically targeting their Wikipedia pages. This rich textual data becomes the fuel for the AI engine.
  • RAG System Construction: A Retrieval Augmented Generation (RAG) system is then constructed. This innovative technology bridges the gap between traditional information retrieval and powerful language models. Simply put, the RAG system guides the AI towards relevant information within the collected data, enabling it to provide more focused and accurate responses.
  • OpenAI GPT-3.5 Turbo Power: As the engine for generating responses, the project utilizes OpenAI’s GPT-3.5 Turbo model. This state-of-the-art language model possesses remarkable capabilities in understanding and responding to natural language inquiries.
  • LangChain and Embeddings: To further enhance the system’s performance, langchain, a powerful framework, is employed. Additionally, free, open-source embeddings from Hugging Face are integrated. These embeddings essentially translate text into a numerical format, allowing the AI model to grasp the relationships between words and concepts with greater precision.
  • FAISS and Local Vectorstore: For efficient information retrieval, FAISS, a library for fast approximate nearest neighbors search, is utilized. This allows the system to quickly locate relevant information within the player data based on the user’s query.
  • Gradio Interface: Finally, a user-friendly Gradio interface is built. This interface acts as the bridge between the user and the AI system. Users can simply type their questions about any of the 35 legendary footballers, and the system will generate insightful and informative responses, drawing upon the vast knowledge base it has constructed.

In essence, this project offers a captivating example of how advanced AI techniques can be harnessed to create a truly engaging and informative experience. Football enthusiasts can now delve deeper into the stories of their favorite players through an interactive and intelligent interface, fostering a deeper appreciation for the sport’s rich history.