Knowledge graph - Scientific news
Researchers at ETH Zurich have published a breakthrough study introducing "Hyper--Relational Graph Embeddings" (HRGE) for Large Language Models. Presented on May 29, 2026, this new mathematical framework allows LLMs to reason over "qualifiers" in knowledge graphs—such as the specific time period or conditions under which a relationship is valid. For the business sector, this development enables AI to perform sophisticated temporal analysis and strategic planning by navigating complex, multi--dimensional dependencies that traditional graphs cannot represent. By moving beyond simple binary relationships, HRGE provides the precision required for automated financial forecasting and risk assessment in volatile markets.