AI Navigate
インサイト最新記事AI大全カオスマップ

DW-Bench: Benchmarking LLMs on Data Warehouse Graph Topology Reasoning

arXiv cs.AI / 4/22/2026

📰 NewsSignals & Early TrendsModels & Research
Read original →
共有:

Key Points

  • The paper introduces DW-Bench, a new benchmark designed to test LLM graph-topology reasoning over data warehouse schemas.
  • DW-Bench explicitly models both foreign-key (FK) relationships and data-lineage edges to better reflect real warehouse graph structure.
  • The benchmark includes 1,046 automatically generated questions that are verifiably correct, spanning five different schemas.
  • Experiments indicate that tool-augmented LLM approaches significantly outperform static (non-tool) methods, though performance plateaus on particularly hard compositional question subtypes.

Abstract

This paper introduces DW-Bench, a new benchmark that evaluates large language models (LLMs) on graph-topology reasoning over data warehouse schemas, explicitly integrating both foreign-key (FK) and data-lineage edges. The benchmark comprises 1,046 automatically generated, verifiably correct questions across five schemas. Experiments show that tool-augmented methods substantially outperform static approaches but plateau on hard compositional subtypes.

関連おすすめサービス

※当サイトはアフィリエイト広告を利用しています
🎧

Notta搭載AI議事録イヤホン ZENCHORD1

AI時代の仕事術。Notta搭載で会議の議事録を自動生成するスマートイヤホン。

🎙️

AI搭載ボイスレコーダー Plaud

世界100万人が愛用。AIで文字起こし・要約を自動化するボイスレコーダー。

🖼️

画像高画質化AIツール Aiarty Image Enhancer

AIで画像を高画質化。写真・イラストを簡単にアップスケール。

DW-Bench: Benchmarking LLMs on Data Warehouse Graph Topology Reasoning | AI Navigate