From Partial to Strictly Incremental Constituent Parsing

Ana Ezquerro, Carlos Gómez-Rodríguez, David Vilares
2/5/2024
cs.CL

No official code found. Best implementation: anaezquerro/incpar (7 stars, MPL-2.0, last commit 2024-04-15).

Abstract

We study incremental constituent parsers to assess their capacity to output trees based on prefix representations alone. Guided by strictly left-to-right generative language models and tree-decoding modules, we build parsers that adhere to a strong definition of incrementality across languages. This builds upon work that asserted incrementality, but that mostly only enforced it on either the encoder or the decoder. Finally, we conduct an analysis against non-incremental and partially incremental models.

View on arXivView PDF

Code Implementations(1)

anaezquerro/incparCommunity100%
MPL-2.0

Fully Incremental Neural Dependency and Constituency Parsing

71PythonSep 13, 20232 years agoMPL-2.0
constituency-parsingdependency-parsinglarge-language-modelsnatural-language-understanding

Similar Papers

Tree Transformers are an Ineffective Model of Syntactic Constituency

78% match2024

Incremental Sentence Processing Mechanisms in Autoregressive Transformer Language Models

77% match1 repo82024

A Fundamental Algorithm for Dependency Parsing (With Corrections)

77% match2025

Towards Incremental Learning in Large Language Models: A Critical Review

76% match2024

A Systematic Study of Compositional Syntactic Transformer Language Models

76% match1 repo22025

Structured Code Representations Enable Data-Efficient Adaptation of Code Language Models

76% match2024

Cite this paper

@article{ezquerro2024from,
  title  = {From Partial to Strictly Incremental Constituent Parsing},
  author = {Ana Ezquerro and Carlos Gómez-Rodríguez and David Vilares},
  year   = {2024},
  eprint = {2402.02782},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2402.02782v1}
}

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/2402.02782.svg)](https://codeofpaper.com/paper/2402.02782)

Discussion