Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
KGVoyager introduces a KG-agnostic architecture that autonomously generates SPARQL queries from natural language questions by leveraging a think-act-observe loop to explore and refine graph structures dynamically. This approach eliminates the need for formal ontologies or curated examples, making it applicable to a broader range of real-world knowledge graph endpoints. The system demonstrates significant improvements, achieving an ~8 point increase in F1 score while reducing cost and runtime by ~22% across four benchmarks.
KGVoyager can autonomously generate effective SPARQL queries without relying on pre-existing ontologies, revolutionizing KGQA in domain-specific contexts.
Knowledge Graph Question Answering (KGQA) over RDF graphs remains challenging in domain-specific settings, where formal ontologies and curated text-SPARQL pairs are often unavailable. We present KGVoyager, a KG-agnostic agentic architecture that generates SPARQL queries from natural language questions by dynamically discovering graph structure and semantics, requiring only a query endpoint of the underlying graph. Using a think-act-observe loop with search, exploration, and execution tools, KGVoyager maps terms to graph IRIs, uncovers structure, and refines queries through execution feedback - all without pre-existing ontologies or examples. Unlike the prior state of the art, KGVoyager requires only a lightweight class index which renders it applicable for far more real-world endpoints. Across four benchmarks, KGVoyager improves F1 by ~8 points while cutting cost and runtime by ~22% each.