KM news
On September 14, 2026, researchers at ETH Zurich published a breakthrough study in Nature Machine Intelligence introducing "Hyper--Dimensional Knowledge Mapping" (HDKM). This framework advances enterprise knowledge management by encoding organizational hierarchies and relational dependencies directly into high--dimensional vector spaces. Unlike traditional flat--vector embeddings used in Retrieval--Augmented Generation (RAG), HDKM preserves the structural context of corporate data, allowing AI systems to navigate complex reporting lines and project versioning with high precision. By utilizing "Algebraic Contextualization," the system enables real--time updates to organizational knowledge graphs without the need for full re--indexing. The research team is scheduled to present the full implementation framework at the Global AI & KM Symposium in Zurich on October 15, 2026.