Published on November 13, 2024, this entry describes a workflow using model-extracted entities, relevant documents, and knowledge-graph relations to generate context-aware query expansions.
On November 13, 2024, Radar recorded an approach called KAR for expanding queries used in information retrieval systems. The workflow extracts entities from a query with a language model, retrieves relevant documents, and traverses relations in a knowledge graph.
It then scores and filters those relations and generates query expansions informed by the documents and graph. The entry presents the proposal in general terms and provides no metrics or comparative results. To verify technical details, consult the original material associated with the entry and check its method and reported findings.