跳到正文
arXiv cs.CL· Yu Zou, Yan Chen, Jialei Zeng, Xiaorui Zhou, Aixi Zhong, Qi Zhou, Haoran You, Ke Jiang, Changyu Hu·· 10 小时前AI 评分29

TETHER:面向设计驱动长文档的免训练、弹性、以目录为锚的层级抽取与恢复框架

TETHER: Training-Free, Elastic, ToC-Anchored Hierarchy Extraction and Recovery in Design-Driven Long Documents

AI 导读

TETHER 是一个免训练框架,把目录(ToC)当作结构先验,通过 LMM 构建文档骨架、再用锚点对齐把内容块绑定到骨架上,实现设计驱动长文档的层级恢复。在多语言 ESG 报告、产品手册和年报基准上,它取得 SOTA 零样本层级恢复效果,优于深层级恢复模型、专用解析器和通用 LMM。相比端到端基线,TETHER 减少 90% token 并大幅提速。

正文 · 原文

View PDF HTML (experimental)

Abstract:Converting complex documents into structured knowledge constitutes the critical infrastructure urgently required for deploying Large Language Models (LLMs) in downstream research. However, blind to the Table of Contents (ToC) as a macro-structure, existing bottom-up paradigms rely on localized visual cues, inducing heuristic fragility in high-entropy layouts where visual prominence decouples from logical depth; meanwhile, direct LLM extraction lacks spatial grounding, triggering hallucinations in long contexts. To address this, we introduce TETHER, a training-free framework that elevates the ToC from discarded noise to a vital structural prior, enabling robust adaptation across design-driven long documents. TETHER operates via a two-stage neuro-symbolic mechanism: first, it constructs a document skeleton from the raw ToC using a Large Multimodal Model (LMM) empowered by spatio-semantic constraints; second, it employs an anchor-based alignment strategy to tether content blocks to this skeleton, effectively resolving boundary ambiguities. Evaluated across a multilingual benchmark of ESG reports, product manuals, and annual reports, TETHER delivers highly precise, state-of-the-art (SOTA) zero-shot hierarchy recovery, outperforming deep hierarchical recovery models, dedicated parsers, and general-purpose LMMs. Moreover, stress tests highlight TETHER's resilience to sparse omissions and textual noise, demonstrating that its precise semantic alignment is fundamentally anchored in topological and depth fidelity. Crucially, ablations confirm the framework's architectural irreducibility, while achieving 90% token reduction and drastic speedups over end-to-end baselines, establishing a compelling equilibrium between structural precision and deployment scalability.
Comments: EMNLP 2026 Main
Subjects: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:2611.24960 [cs.CL]
  (or arXiv:2611.24960v1 [cs.CL] for this version)
  https://doi.org/10.48550/arXiv.2611.24960

arXiv-issued DOI via DataCite (pending registration)

Journal reference: https://aclanthology.org/2026.emnlp-main.2354/
Related DOI: https://doi.org/10.18653/2026.emnlp-main.2354

DOI(s) linking to related resources

Submission history

From: Yu Zou [view email]
[v1] Wed, 11 Nov 2026 19:34:29 UTC (12,854 KB)

来源:arXiv cs.CL · arxiv.org