In visually-rich documents, information is encoded not only in individual page objects such as tables, headers, and text blocks, but also in the structural relations among them, making document structure analysis fundamental to information retrieval and document understanding. However, accurately inferring such relations remains challenging in multi-page documents with long-range dependencies and heterogeneous layouts. To address this, we propose a tree-guided and self-regressive framework, termed DOcument Structure Analyzer (DOSA), for inferring relations among page objects and reconstructing document-level semantic trees. DOSA processes documents chunk-by-chunk, fusing visual, textual, and layout features for each page object and predicting hierarchical and ordering relations. The predicted relations are used to incrementally construct a semantic tree, which is then leveraged as structural context to guide inference on subsequent chunks. Experimental results on five benchmarks demonstrate the effectiveness of DOSA, with improvements of up to 4 F1 points and 19 TEDS points on DocHieNet, the most challenging multi-page hierarchy benchmark.
Long-document question-answering experiments show that human-verified TOC hierarchies and contextual relationships improve reasoning, with their combination providing complementary benefits.
Yuefeng Zou, Yichen Lu, Jingxiao Yang et al.· 0 citations
This work presents DocMaster, a hierarchical structure-aware document analysis system that parses documents into hierarchical document trees preserving original layouts and constructs a structure-aware semantic index that enables accurate document filtering and in-depth analysis.
Ziqi Chen, Yingli Zhou, Fangyuan Zhang et al.· 0 citations
Structured documents such as invoices, forms, reports, and scientific articles derive meaning from the interplay between spatial layout, textual content, and logical structure. Generative models operating at the pixel or token level often struggle to capture these dependencies effectively. We explore FRAGMENT, a generative framework that represents a document as a typed relational graph and factorizes its distribution as p(structure, content) = p(structure) * p(content | structure). The framework consists of two stages. The first stage, the Architect, is a causally masked Transformer conditioned on document category that autoregressively generates the graph topology and typed spatial relations. The second stage, the Builder, is a GATv2-based graph attention network that enriches the graph with normalized bounding boxes, text, and visual style attributes. Both stages define explicit likelihood models, yielding a tractable document-level likelihood that serves as an anomaly score for forgery detection. For controlled editing, a prompt-conditioned extension injects instruction embeddings into the Builder through cross-attention, enabling semantic and entity-aware modifications. We describe training on DocLayNet and fine-tuning on FUNSD and SROIE. Experiments on DocLayNet, FUNSD, and SROIE evaluate FRAGMENT alongside representative autoregressive, layout-only, and graph-based baselines, providing an empirical analysis of the characteristics and trade-offs of the proposed factorized graph generation framework.
Ayoub El Bouchtili, Guilhaume Leroy-Meline· 0 citations
Despite the remarkable generative capabilities of Large Language Models (LLMs), existing long-form document generation methods frequently suffer from structural collapse and logical drift as the document length increases. We identify the root cause as a persistent structural retrieval gap, where traditional Retrieval-Augmented Generation (RAG) paradigms focus almost exclusively on fine-grained factual snippets while neglecting the macro-level organizational logic required for coherent narratives. To bridge this gap, we propose a pioneering hierarchical structure-retrieved generation framework (HS-RAG) that reconceptualizes the generation task as a systematic retrieval-alignment-fusion process from template to text, marking a fundamental paradigm shift from spontaneous generation to grounded structural anchoring. The proposed framework introduces three core methodological innovations: 1) A novel dual-stream structural indexing mechanism that extracts retrievable prototypes from historical templates, effectively closing the structural retrieval gap; 2) An adaptive query-guided alignment algorithm that maps unstructured user intents into explicit hierarchical logical constraints; 3) A reinforcement learning (RL)-driven hierarchical relevance feedback optimization strategy designed to dynamically maximize the alignment between generated content, retrieved structures, and external knowledge. Extensive experiments across four diverse datasets demonstrate that our approach significantly outperforms state-of-the-art baselines, including GPT-5.2 and advanced RAG variants. Specifically, our framework achieves an exceptional Structure Match Rate (SMR) of 0.951 and a semantic fidelity score (BERTScore) of 0.921 on AcademicPapers dataset, outperforming the strongest SOTA baseline by 5.43% and 4.19%, respectively. These compelling quantitative indicators validate that explicit structural priors are essential for mitigating long-range incoherence in controllable document generation.
Yongpan Wang, Yu Tan, Mingli Song et al.· IEEE Access· 0 citations
Hyper-relational extraction aims to identify complex factual structures from unstructured text, where each instance consists of a core triplet and multiple qualified attributes. However, existing hyper-relational datasets often suffer from long-tail relation distributions and pseudo-negative samples, which limit the performance of hyper-relational extraction models. Although large language models (LLMs) provide new opportunities for data augmentation, they may introduce semantically inconsistent relations and structurally unreliable samples in complex scenarios, reducing the quality of generated data. To address these challenges, this paper proposes a structure-consistency-driven semi-automatic data augmentation framework, termed HRE-LSC. The framework consists of two key components: (1) a distribution-aware generation strategy that selectively generates samples for low-frequency relations according to relation frequency distributions, thereby alleviating data imbalance; and (2) a hierarchical logical consistency verification mechanism based on natural language inference (NLI), which progressively verifies the consistency of core triplets and qualified attributes to filter unreliable generated samples and reduce pseudo-negative samples. Experiments on the HyperRED dataset demonstrate that the proposed method improves the F1 score of Text2NKG from 83.81% to 84.7%, with more significant improvements observed on long-tail relation subsets. The results indicate that the proposed framework effectively enhances the quality and structural consistency of generated hyper-relational data while mitigating the effects of long-tail distributions and pseudo-negative samples without requiring additional manual annotations.
Xinping Du, Yan Chen, Siqi Xu et al.· Applied Sciences· 0 citations
Answering complex questions over large document collections requires assembling complementary evidence across sections and documents. GraphRAG offers structured retrieval but typically uses fixed traversal, while agentic RAG operates over weakly structured interfaces. Our key insight is that agents should navigate document structure within and across documents rather than repeatedly search from scratch. We introduce DocNavRAG, which organizes document hierarchies and cross-region relations into a navigable graph, exposes graph operations for locating, navigating, expanding, and fetching, and maintains an evolving evidence state to guide retrieval until sufficient evidence is collected. Across four long- and multi-document QA benchmarks, DocNavRAG improves answer quality and context sufficiency over the strongest baseline by 7.8\% and 17.7\% on average.