Interaction AI with Retrieval-enhanced Generation for Knowledge Retrieval in EAST

The tokamak is the most representative device for achieving magnetic confinement fusion and is one of the most promising pathways to realizing fusion energy. Among the many tokamak devices, China's first fully superconducting tokamak experiment device, the Experimental Advanced Superconducting Tokamak (EAST), has garnered significant attention due to its advanced technology and outstanding experimental achievements. Over the years, EAST has accumulated a wealth of experimental data and literature records. These documents provide profound background knowledge and professional insights. To enhance the efficiency of researchers in retrieving documents and gaining inspiration from historical documents, we have utilized EAST-related experimental documents, combined with Retrieval-Augmented Generation (RAG) technology and state-of-the-art Large Language Models (LLMs), to create an "operational copilots" model. The model leverages Intelligent Augmented Intelligence (IAI) to not only provide accurate answers but also draw on historical documents to explain specialized issues. Through practical examples, it has been demonstrated that for highly specialized tokamak-related questions, the RAG retrieval model offers more credible responses than a basic model.