EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

GNLM: A Graph-Native Language Model With Road Name Address-Based Spatial Reasoning for Geographic Question Answering

Chae-Seok Lee, Dae-Seung Park, Jae-Min Choi, Ho-Jong Chang

IEEE Access · 2026 · 14 · 78339-78361

DOI: 10.1109/access.2026.3695549 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Chae-Seok Lee und 3 weiteren Autor:innen. Die Arbeit nennt 50 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Institute of Information and Communications Technology Planning and Evaluation, Korean Government (MSIT).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Geographic question answering (QA) requires factually accurate responses grounded in structured spatial knowledge. However, existing approaches based on large language models (LLMs) suffer from factual hallucination, while retrieval-augmented generation (RAG) lacks structural reasoning capabilities. We propose GNLM (Graph-Native Language Model), a novel architecture that restructures the traditional LLM pipeline by performing all reasoning through knowledge graph traversal and restricting the role of LLMs to natural language generation. The knowledge graph is constructed exclusively from government public data sources (building registry, road name addresses) and expert-curated domain knowledge, ensuring factual reliability without LLM-generated content. GNLM introduces four key contributions: 1) a Graph-First, LLM-Last architecture with fact validation that structurally prevents fabrication (all observed errors stem from knowledge gaps, not false generation); 2) an extended Resource Description Framework (RDF) triple attribute model that represents entity properties as independent graph nodes connected by semantic edges, enabling property-level querying and meta-reasoning; 3) coordinate-free spatial reasoning that leverages the Korean road numbering standard (20 m interval, $k = 10$ m per unit) for proximity ranking without Global Positioning System (GPS) coordinates; and 4) a 14-strategy reasoning router that classifies queries into specialized graph traversal algorithms. Evaluated on a real-world dataset of Daejeon, South Korea ( $58{,}882$ nodes, $80{,}088$ edges from $58{,}408$ buildings and 711 roads), GNLM achieves 88.4% accuracy on 250 diverse queries with a hallucination rate (HR) of only 5.7% on 35 fact-verification queries, compared to 60.0% for pure LLM, 25.7% for Dense RAG, and 17.1% for GraphRAG—demonstrating a consistent improvement as structural grounding increases. Cross-model validation with llama3.1:8b confirms LLM-agnostic generalizability. Multi-region validation on Sejong City ( $17{,}008$ buildings) with identical methodology confirms cross-region generalizability (Yuseong-gu avg $\rho = 0.976$ , Sejong avg $\rho = 0.973$ , both over 10 roads). Road-number-based proximity ranking achieves Spearman $\rho = 0.9878$ and Precision@ $10= 96.0\%$ against GPS ground truth, with short-range ( $\rho = 0.701$ —substantially lower than same-road accuracy, indicating a current practical limitation that motivates the multi-hop and automated intersection-detection extensions.

Förderung

  • Institute of Information and Communications Technology Planning and Evaluation
  • Korean Government (MSIT)

Zitieren

APA

Lee, C., Park, D., Choi, J., & Chang, H. (2026). GNLM: A Graph-Native Language Model With Road Name Address-Based Spatial Reasoning for Geographic Question Answering. IEEE Access, 14, 78339-78361. https://doi.org/10.1109/access.2026.3695549

Vancouver

Lee C, Park D, Choi J, Chang H. GNLM: A Graph-Native Language Model With Road Name Address-Based Spatial Reasoning for Geographic Question Answering. IEEE Access. 2026;14:78339-78361. doi:10.1109/access.2026.3695549

BibTeX
@article{Lee2026,
  author = {Lee, Chae-Seok and Park, Dae-Seung and Choi, Jae-Min and Chang, Ho-Jong},
  title = {GNLM: A Graph-Native Language Model With Road Name Address-Based Spatial Reasoning for Geographic Question Answering},
  journal = {IEEE Access},
  year = {2026},
  volume = {14},
  pages = {78339--78361},
  doi = {10.1109/access.2026.3695549}
}
Kostenlos starten →