EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Focus on Evidence: Relational Structure Enhances LLM Effectiveness in TableQA

Zhen Yang, Ziwei Du, Minghan Zhang, Wei Du, Jie Chen, Fulan Qian, Shu Zhao

ACM Transactions on Information Systems · 30.11.2026 · 44 (8) · 1-31

DOI: 10.1145/3848125 ↗

Kurz gefasst

Fachartikel in der Zeitschrift ACM Transactions on Information Systems, online erschienen am 13.11.2026. Verfasst von Zhen Yang und 6 weiteren Autor:innen. Die Arbeit nennt 73 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Natural Science Foundation of China, Anhui Province Excellent Scientific Research and Innovation Team, Anhui Provincial Natural Science Foundation—Water Science Joint Fund.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Table Question Answering (TableQA) requires reasoning over natural language questions and structured tables and remains challenging due to noisy evidence and complex multi-step reasoning. Recent Large Language Model (LLM)-based approaches typically adopt decomposition–reasoning–validation pipelines that combine Chain-of-Thought (CoT) decomposition, Direct Prompting (DP), Python Agent (PyAgent) execution, and self-validation. However, these methods still largely rely on unconstrained LLM reasoning during decomposition, which may introduce unsupported sub-questions and irrelevant evidence expansion. We propose EV idence- A ware (EVA), a framework that explicitly grounds decomposition on question-relevant table evidence before reasoning begins. EVA first extracts evidence-aware relational structures that associate question semantics with supporting table attributes, thereby constraining sub-question generation within a grounded evidence space. EVA further introduces relation-guided validation and integrates both DP and PyAgent reasoning to improve robustness across different reasoning patterns. Experiments on datasets across multiple LLMs demonstrate that EVA consistently improves reasoning reliability and achieves strong performance across diverse TableQA settings.

Förderung

  • Natural Science Foundation of China
  • Anhui Province Excellent Scientific Research and Innovation Team
  • Anhui Provincial Natural Science Foundation—Water Science Joint Fund
  • University Synergy Innovation Program of Anhui Province
  • SMP-Zhipu.AI Large Model Cross-Disciplinary Fund

Zitieren

APA

Yang, Z., Du, Z., Zhang, M., Du, W., Chen, J., Qian, F., & Zhao, S. (2026). Focus on Evidence: Relational Structure Enhances LLM Effectiveness in TableQA. ACM Transactions on Information Systems, 44(8), 1-31. https://doi.org/10.1145/3848125

Vancouver

Yang Z, Du Z, Zhang M, Du W, Chen J, Qian F, et al. Focus on Evidence: Relational Structure Enhances LLM Effectiveness in TableQA. ACM Transactions on Information Systems. 2026;44(8):1-31. doi:10.1145/3848125

BibTeX
@article{Yang2026,
  author = {Yang, Zhen and Du, Ziwei and Zhang, Minghan and Du, Wei and Chen, Jie and Qian, Fulan and Zhao, Shu},
  title = {Focus on Evidence: Relational Structure Enhances LLM Effectiveness in TableQA},
  journal = {ACM Transactions on Information Systems},
  year = {2026},
  volume = {44},
  number = {8},
  pages = {1--31},
  doi = {10.1145/3848125}
}
Kostenlos starten →