Deploying Secure RAG Systems: Generative AI for Higher Education

Session Abstract: This session delves into the University of Arizona’s deployment of a Retrieval-Augmented Generation (RAG) system, showcasing how generative AI can transform data access and decision-making in higher education. Key topics include organizing unstructured data, building vector databases, implementing semantic search, and optimizing prompt engineering. This session is crucial for those looking to deploy AI systems within their own institutions, offering insights into maintaining data privacy while enabling advanced applications. Attendees will explore real-world use cases, such as the ASK-KMAP research query tool, automated grant announcements, and an AI-powered news writing assistant. By the end, participants will understand how to harness RAG systems to improve research capabilities, streamline information access, and enhance decision-making processes at their institutions. The system is live at https://kmap.arizona.edu.

Keywords: RAG, unstructured data, semantic search, knowledge map, research metadata, AI applications

Contributing Authors: Thomas Harman – University of Arizona; Wesley Nguyen – University of Arizona