Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Orthogonal Hierarchical Decomposition for Structure-Aware Table Understanding with Large Language Models

Bin Cao, Huixian Lu, Chenwen Ma, Ting Wang, Ruizhe Li, Jing Fan

TL;DR

The paper tackles the problem of structuring and reasoning over complex, non-canonical tables that challenge traditional grid-based LLM inputs. It introduces Orthogonal Hierarchical Decomposition (OHD), which decouples table structure into independent row and column hierarchies via Orthogonal Tree Induction (OTI) and then reconstructs a cell’s semantics through a dual-pathway protocol, with a semantic arbitration step using an LLM. The approach yields structure-preserving representations that significantly improve performance on complex table QA benchmarks (AITQA and HiTab) over multiple baselines, with ablations validating the importance of semantic predicates, dual-axis pathways, and arbitration. This work advances robust, explainable table understanding and has practical implications for accurate data extraction and reasoning in domains with irregular table layouts, while highlighting the need for careful human oversight in high-stakes use cases.

Abstract

Complex tables with multi-level headers, merged cells and heterogeneous layouts pose persistent challenges for LLMs in both understanding and reasoning. Existing approaches typically rely on table linearization or normalized grid modeling. However, these representations struggle to explicitly capture hierarchical structures and cross-dimensional dependencies, which can lead to misalignment between structural semantics and textual representations for non-standard tables. To address this issue, we propose an Orthogonal Hierarchical Decomposition (OHD) framework that constructs structure-preserving input representations of complex tables for LLMs. OHD introduces an Orthogonal Tree Induction (OTI) method based on spatial--semantic co-constraints, which decomposes irregular tables into a column tree and a row tree to capture vertical and horizontal hierarchical dependencies, respectively. Building on this representation, we design a dual-pathway association protocol to symmetrically reconstruct semantic lineage of each cell, and incorporate an LLM as a semantic arbitrator to align multi-level semantic information. We evaluate OHD framework on two complex table question answering benchmarks, AITQA and HiTab. Experimental results show that OHD consistently outperforms existing representation paradigms across multiple evaluation metrics.

Orthogonal Hierarchical Decomposition for Structure-Aware Table Understanding with Large Language Models

TL;DR

The paper tackles the problem of structuring and reasoning over complex, non-canonical tables that challenge traditional grid-based LLM inputs. It introduces Orthogonal Hierarchical Decomposition (OHD), which decouples table structure into independent row and column hierarchies via Orthogonal Tree Induction (OTI) and then reconstructs a cell’s semantics through a dual-pathway protocol, with a semantic arbitration step using an LLM. The approach yields structure-preserving representations that significantly improve performance on complex table QA benchmarks (AITQA and HiTab) over multiple baselines, with ablations validating the importance of semantic predicates, dual-axis pathways, and arbitration. This work advances robust, explainable table understanding and has practical implications for accurate data extraction and reasoning in domains with irregular table layouts, while highlighting the need for careful human oversight in high-stakes use cases.

Abstract

Complex tables with multi-level headers, merged cells and heterogeneous layouts pose persistent challenges for LLMs in both understanding and reasoning. Existing approaches typically rely on table linearization or normalized grid modeling. However, these representations struggle to explicitly capture hierarchical structures and cross-dimensional dependencies, which can lead to misalignment between structural semantics and textual representations for non-standard tables. To address this issue, we propose an Orthogonal Hierarchical Decomposition (OHD) framework that constructs structure-preserving input representations of complex tables for LLMs. OHD introduces an Orthogonal Tree Induction (OTI) method based on spatial--semantic co-constraints, which decomposes irregular tables into a column tree and a row tree to capture vertical and horizontal hierarchical dependencies, respectively. Building on this representation, we design a dual-pathway association protocol to symmetrically reconstruct semantic lineage of each cell, and incorporate an LLM as a semantic arbitrator to align multi-level semantic information. We evaluate OHD framework on two complex table question answering benchmarks, AITQA and HiTab. Experimental results show that OHD consistently outperforms existing representation paradigms across multiple evaluation metrics.
Paper Structure (20 sections, 10 equations, 2 figures, 6 tables, 2 algorithms)

This paper contains 20 sections, 10 equations, 2 figures, 6 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. Orthogonal Hierarchical Decomposition (OHD)
  4. Construction Criteria: Semantic-Spatial Synergy
  5. Orthogonal Tree Induction (OTI)
  6. Structural Association Reconstruction
  7. Semantic Arbitration and Refinement
  8. Discussion and Case Analysis
  9. Experiments
  10. Datasets
  11. Baselines
  12. Evaluation Metrics
  13. Results
  14. Ablation Analysis.
  15. Conclusion
  16. ...and 5 more sections

Figures (2)

  • Figure 1: Illustration of table complexity and structural diversity. The examples encompass several challenging non-standard layouts. (a): Tables featuring multi-level nested column headers and merged data cells; (b): Tables characterized by deep hierarchical row header structures; (c): Complex instances with simultaneous multi-layer hierarchies in both rows and columns (dual-axis dependency); (d): Tables with flexible header positioning (column headers located in non-top sections) and highly irregular structural topologies.
  • Figure 2: Workflow of the OHD framework. The process begins with a Categorized Table Input where each cell is pre-identified as a Row Header, Column Header, or Data unit. The pipeline then proceeds in three stages: (1) Orthogonal Tree Induction (OTI) to decompose the table into independent row and column hierarchical trees; (2) Dual-Path Lineage Extraction to reconstruct the semantic lineage of each cell via synchronized tree traversal; and (3) Semantic Arbitration using an LLM to align hierarchical information into structure-aware prompts.