Neuro-symbolic Syntactic Parsing: Shaping a Neural Network with the CYK Algorithm

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Zanzotto, Fabio Massimo, Ranaldi, Federico, Satta, Giorgio
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910272455704576
author Zanzotto, Fabio Massimo
Ranaldi, Federico
Satta, Giorgio
author_facet Zanzotto, Fabio Massimo
Ranaldi, Federico
Satta, Giorgio
contents In this paper, we show the possibility of a direct injection of algorithms into neural network architecture. We focus on a complex algorithm, that is, Cocke-Youger-Kasami (CYK) for parsing context-free grammars in Chomsky Normal Form and we propose CYKNN, a simple recurrent neural network architecture for encoding the CYK algorithm in trainable matrix-vector multiplications.We experimented with a very simple grammar with 4 variations showing that our approach outperforms existing LLMs with more than 20B parameters with an in-context learning setting and smaller LLMs of the Qwen family fine-tuned with LoRA. Our attempt paves the way to a different approach to neuro-symbolic methodologies.
format Preprint
id arxiv_https___arxiv_org_abs_2605_31421
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Neuro-symbolic Syntactic Parsing: Shaping a Neural Network with the CYK Algorithm
Zanzotto, Fabio Massimo
Ranaldi, Federico
Satta, Giorgio
Computation and Language
Artificial Intelligence
Data Structures and Algorithms
In this paper, we show the possibility of a direct injection of algorithms into neural network architecture. We focus on a complex algorithm, that is, Cocke-Youger-Kasami (CYK) for parsing context-free grammars in Chomsky Normal Form and we propose CYKNN, a simple recurrent neural network architecture for encoding the CYK algorithm in trainable matrix-vector multiplications.We experimented with a very simple grammar with 4 variations showing that our approach outperforms existing LLMs with more than 20B parameters with an in-context learning setting and smaller LLMs of the Qwen family fine-tuned with LoRA. Our attempt paves the way to a different approach to neuro-symbolic methodologies.
title Neuro-symbolic Syntactic Parsing: Shaping a Neural Network with the CYK Algorithm
topic Computation and Language
Artificial Intelligence
Data Structures and Algorithms
url https://arxiv.org/abs/2605.31421