Format-Constraint Coupling in Knowledge Graph Construction from Statistical Tables

Jingxuan Qi, Zhiqiang Ye, Yuxiang Feng

arXiv:2605.21974 · 2026-05-23 공개 · arXiv · PDF

llm-evaluation knowledge-graph graphrag format-constraint-coupling csv-fidelity statistical-tables entity-inflation retrieval-modes

Abstract

An extraction schema should not reduce knowledge graph fidelity. On statistical CSV, however, it can. We study country-by-year time-series matrices, a common layout on open-data portals. In this setting, serialization format and schema constraints interact super-additively. Their joint effect exceeds the sum of independent effects by up to +1.180 (2x2 factorial, 6 datasets). Bootstrap 95% CIs are strictly positive on 4/6 datasets, with strongest evidence on wide Type-II matrices. More critically, a schema applied to a mismatched format can trigger catastrophic mismatch. Fact coverage falls below the unconstrained baseline on 4/6 datasets through entity inflation or extraction refusal. We call this observed pattern format-constraint coupling. Probing and token ablation support a surface-form anchoring explanation centred on column-name references. Controlled variants across format-schema pairings, GraphRAG hosts, and LLM families show the same direction within the measured scope; one LLM family shows only partial activation. The observation also has a diagnostic consequence. Three standard retrieval modes largely mask construction quality (delta <= 1pp), whereas direct graph access exposes gaps up to +47.6pp (p < 0.0001). To support fidelity-aware evaluation, we release CSVFidelity-Bench. It contains 15 datasets, 11 Type-II matrices, 4 Type-III tables, and 1,892 Gold Standard facts across 6 domains.

한국어 요약

한 줄 요약

**[Knowledge Graph / LLM 추출]** Format-constraint coupling이 통계 CSV에서 KG 구성 시 schema·serialization의 super-additive 상호작용으로 fidelity 저하 유발 — 최대 +1.180 결합 효과, 4/6 데이터셋에서 catastrophic mismatch, CSVFidelity-Bench 공개.

핵심 기여도

핵심 아이디어

LLM 기반 KG 추출의 fidelity 저하는 schema·format을 독립 변수로 취급해서는 안 되며 둘의 상호작용이 super-additive로 catastrophic 효과를 낼 수 있다 — column-name reference의 surface-form anchoring이 원인이며, 표준 retrieval은 이를 mask하므로 direct graph access 기반 평가가 필요하다.

기술적 접근법

주요 결과

의의 및 한계

**의의**: KG 추출 평가의 systematic 결함 노출, format-constraint coupling이라는 새 진단 개념, retrieval 평가가 construction 결함 masking을 정량 입증, 공개 벤치마크로 분야 발전 가속. **한계**: Country-by-year 통계 데이터에 한정, 한 LLM family에서 partial activation, surface-form anchoring 가설의 다른 도메인 일반화 필요.

실용적 활용