Exploiting Pre-trained Encoder-Decoder Transformers for Sequence-to-Sequence Constituent Parsing

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Fernández-González, Daniel, Cid, Cristina Outeiriño
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914563248619520
author Fernández-González, Daniel
Cid, Cristina Outeiriño
author_facet Fernández-González, Daniel
Cid, Cristina Outeiriño
contents To achieve deep natural language understanding, syntactic constituent parsing plays a crucial role and is widely required by many artificial intelligence systems for processing both text and speech. A recent approach involves using standard sequence-to-sequence models to handle constituent parsing as a machine translation problem, moving away from traditional task-specific parsers. These models are typically initialized with pre-trained encoder-only language models like BERT or RoBERTa. However, the use of pre-trained encoder-decoder language models for constituency parsing has not been thoroughly explored. To bridge this gap, we extend the sequence-to-sequence framework by investigating parsers built on pre-trained encoder-decoder architectures, including BART, mBART, and T5. We fine-tune them to generate linearized parse trees and extensively evaluate them on different linearization strategies across both continuous treebanks and more complex discontinuous benchmarks. Our results demonstrate that our approach outperforms all prior sequence-to-sequence models and performs competitively with leading task-specific constituent parsers on continuous constituent parsing.
format Preprint
id arxiv_https___arxiv_org_abs_2605_13373
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Exploiting Pre-trained Encoder-Decoder Transformers for Sequence-to-Sequence Constituent Parsing
Fernández-González, Daniel
Cid, Cristina Outeiriño
Computation and Language
68T50
I.2.7
To achieve deep natural language understanding, syntactic constituent parsing plays a crucial role and is widely required by many artificial intelligence systems for processing both text and speech. A recent approach involves using standard sequence-to-sequence models to handle constituent parsing as a machine translation problem, moving away from traditional task-specific parsers. These models are typically initialized with pre-trained encoder-only language models like BERT or RoBERTa. However, the use of pre-trained encoder-decoder language models for constituency parsing has not been thoroughly explored. To bridge this gap, we extend the sequence-to-sequence framework by investigating parsers built on pre-trained encoder-decoder architectures, including BART, mBART, and T5. We fine-tune them to generate linearized parse trees and extensively evaluate them on different linearization strategies across both continuous treebanks and more complex discontinuous benchmarks. Our results demonstrate that our approach outperforms all prior sequence-to-sequence models and performs competitively with leading task-specific constituent parsers on continuous constituent parsing.
title Exploiting Pre-trained Encoder-Decoder Transformers for Sequence-to-Sequence Constituent Parsing
topic Computation and Language
68T50
I.2.7
url https://arxiv.org/abs/2605.13373