Overview
Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!
Learn about groundbreaking research from Google and Stanford University that reveals critical insights and practical solutions for enhancing Retrieval-Augmented Generation (RAG) systems in this 40-minute technical presentation. Explore how Graph Neural Networks (GNNs) can be utilized to evaluate and improve RAG systems' performance, particularly in complex question-answering scenarios involving scientific texts. Discover the current limitations of RAG frameworks, including their struggles with accurately mapping queries to relevant text passages, and understand the proposed solutions for developing more sophisticated, context-aware mechanisms. Delve into advanced computational strategies that bridge the gap between theoretical capabilities and practical performance, focusing on improving query-to-text mapping fidelity. Master the integration of cutting-edge machine learning techniques designed to parse, understand, and correlate nuanced relationships within data, ultimately leading to more accurate and relevant information retrieval in RAG systems.
Syllabus
Simple ideas to improve your RAG (Stanford, Google)
Taught by
Discover AI