AI Debate Workshop @ Wu Ta-You Science Camp 2024
A multi-agent AI debate system using LangChain and RAG, introduced in a workshop and live competition with 100+ participants
I was honored to serve as a teaching assistant at Wu Ta-Yoh Science Camp 2024, where we hosted an AI debate workshop. For this event, we designed a multi-agent debating system powered by LangChain and Retrieval-Augmented Generation (RAG).
In this system, a router LLM first identifies the category of each question and then routes it to an expert LLM equipped with domain-specific knowledge, enhanced by RAG. This setup allowed participants to experience how specialized AI agents can collaborate to simulate dynamic debates.
Finally, we guided participants to build their own debate systems and concluded with an engaging AI debate competition, making the learning process both hands-on and fun.
