Hierarchical Bracketing Encodings for Dependency Parsing as Tagging

Ana Ezquerro, David Vilares, Anssi Yli-Jyrä, Carlos Gómez-Rodríguez
5/16/2025
cs.CL

No official code found. Best implementation: anaezquerro/separ (7 stars, Apache-2.0, last commit 2025-10-01).

Abstract

We present a family of encodings for sequence labeling dependency parsing, based on the concept of hierarchical bracketing. We prove that the existing 4-bit projective encoding belongs to this family, but it is suboptimal in the number of labels used to encode a tree. We derive an optimal hierarchical bracketing, which minimizes the number of symbols used and encodes projective trees using only 12 distinct labels (vs. 16 for the 4-bit encoding). We also extend optimal hierarchical bracketing to support arbitrary non-projectivity in a more compact way than previous encodings. Our new encodings yield competitive accuracy on a diverse set of treebanks.

View on arXivView PDF

Code Implementations(1)

anaezquerro/separCommunity100%
Apache-2.0

Parsing as Sequence Labeling with Neural Models

70PythonOct 5, 20241 years agoApache-2.0

Similar Papers

Hierarchical Bracketing Encodings Work for Dependency Graphs

89% match2025

Dependency Graph Parsing as Sequence Labeling

81% match2024

Textless Dependency Parsing by Labeled Sequence Prediction

73% match1 repo22024

Modeling Text-Label Alignment for Hierarchical Text Classification

71% match1 repo302024

From Partial to Strictly Incremental Constituent Parsing

71% match1 repo72024

HILL: Hierarchy-aware Information Lossless Contrastive Learning for Hierarchical Text Classification

70% match1 repo212024

Cite this paper

@article{ezquerro2025hierarchical,
  title  = {Hierarchical Bracketing Encodings for Dependency Parsing as Tagging},
  author = {Ana Ezquerro and David Vilares and Anssi Yli-Jyrä and Carlos Gómez-Rodríguez},
  year   = {2025},
  eprint = {2505.11693},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2505.11693v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2505.11693.svg)](https://codeofpaper.com/paper/2505.11693)

Discussion