Scientists are addressing the challenge of limited access to expert knowledge that hampers scalability in organizations. Researchers from Siemens AG and Eindhoven University of Technology have developed a groundbreaking software engineering framework that integrates human domain expertise into artificial intelligence systems. By augmenting Large Language Models (LLMs) with codified rules and a Retrieval-Augmented Generation (RAG) system, their framework significantly enhances simulation data visualizations, achieving a 206% improvement in output quality across engineering scenarios. This innovation allows non-experts to generate expert-level visualizations while reducing dependency on scarce expert time. The study also illustrates how capturing tacit expert knowledge can democratize complex insights, boosting productivity. They conducted evaluations with 12 testers across five engineering domains, showcasing superior performance and consistency in generating high-quality outputs. This research not only offers a solution to expert bottlenecks but also promotes more effective data analysis and informed decision-making across various industries, ultimately advancing AI’s capabilities in engineering contexts.
Source link
