arXiv cs.AI· Jia Ling, Yangfan Wang, Chen Tang, Haoming Tan, Yang Yang, Yi Guan, Jingchi Jiang·· 6 天前AI 评分22
H2Table：用分层超图增强 LLM 进行复杂表格推理
H2Table: Hierarchical Hypergraph-Enhanced Large Language Models for Complex Table Reasoning
AI 导读
H2Table 将复杂表格表示为分层嵌套超图，并设计专用超图编码器在表头与单元格之间传递消息，再通过一组可学习查询向量把结构化嵌入轻量注入 LLM。在 HiTab 数据集上，该方法在嵌套深度为四的高复杂表格上较 SOTA 基线平均提升 22.88%，代码已开源，论文被 EMNLP 2026 主会接收。
正文 · 原文
Abstract:Tables are ubiquitous across diverse domains, yet reasoning over them remains a significant challenge for modern large language models (LLMs). Current approaches typically linearize tables into sequences, inherently overlooking their intrinsic two-dimensional and hierarchical structure. To address this, we propose H2Table (Hierarchical Hypergraph-Enhanced Table Reasoning), a novel framework that represents complex tables as hierarchical nested hypergraphs. To process this representation, we design a tailored hypergraph encoder to facilitate message passing between hyperedges (headers) and nodes (cells), thereby perceiving the semantic entailment relationships between them within complex tables. Furthermore, we introduce a set of learnable query vectors acting as a lightweight bridge to extract representative structural embeddings from the encoder into the LLM. Experimental results demonstrate that our approach effectively handles complex table question answering tasks with hierarchical nested headers. Notably, on the HiTab dataset, H2Table achieves an average improvement of 22.88% over state-of-the-art baselines on highly complex tables with a nesting depth of four. Our code is available at: this https URL.
| Comments: | Accepted to EMNLP 2026 (Main Conference). Camera-ready version |
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2609.01216 [cs.AI] |
| (or arXiv:2609.01216v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2609.01216 arXiv-issued DOI via DataCite |
Submission history
From: Jia Ling [view email]
[v1]
Tue, 1 Sep 2026 13:19:42 UTC (1,948 KB)
[v2]
Fri, 9 Oct 2026 14:10:37 UTC (1,948 KB)
来源：arXiv cs.AI · arxiv.org