GenCNER: A Generative Framework for Continual Named Entity Recognition

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Yang, Yawen, Ma, Fukun, Meng, Shiao, Liu, Aiwei, Wen, Lijie
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914090292609024
author Yang, Yawen
Ma, Fukun
Meng, Shiao
Liu, Aiwei
Wen, Lijie
author_facet Yang, Yawen
Ma, Fukun
Meng, Shiao
Liu, Aiwei
Wen, Lijie
contents Traditional named entity recognition (NER) aims to identify text mentions into pre-defined entity types. Continual Named Entity Recognition (CNER) is introduced since entity categories are continuously increasing in various real-world scenarios. However, existing continual learning (CL) methods for NER face challenges of catastrophic forgetting and semantic shift of non-entity type. In this paper, we propose GenCNER, a simple but effective Generative framework for CNER to mitigate the above drawbacks. Specifically, we skillfully convert the CNER task into sustained entity triplet sequence generation problem and utilize a powerful pre-trained seq2seq model to solve it. Additionally, we design a type-specific confidence-based pseudo labeling strategy along with knowledge distillation (KD) to preserve learned knowledge and alleviate the impact of label noise at the triplet level. Experimental results on two benchmark datasets show that our framework outperforms previous state-of-the-art methods in multiple CNER settings, and achieves the smallest gap compared with non-CL results.
format Preprint
id arxiv_https___arxiv_org_abs_2510_11444
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle GenCNER: A Generative Framework for Continual Named Entity Recognition
Yang, Yawen
Ma, Fukun
Meng, Shiao
Liu, Aiwei
Wen, Lijie
Computation and Language
Traditional named entity recognition (NER) aims to identify text mentions into pre-defined entity types. Continual Named Entity Recognition (CNER) is introduced since entity categories are continuously increasing in various real-world scenarios. However, existing continual learning (CL) methods for NER face challenges of catastrophic forgetting and semantic shift of non-entity type. In this paper, we propose GenCNER, a simple but effective Generative framework for CNER to mitigate the above drawbacks. Specifically, we skillfully convert the CNER task into sustained entity triplet sequence generation problem and utilize a powerful pre-trained seq2seq model to solve it. Additionally, we design a type-specific confidence-based pseudo labeling strategy along with knowledge distillation (KD) to preserve learned knowledge and alleviate the impact of label noise at the triplet level. Experimental results on two benchmark datasets show that our framework outperforms previous state-of-the-art methods in multiple CNER settings, and achieves the smallest gap compared with non-CL results.
title GenCNER: A Generative Framework for Continual Named Entity Recognition
topic Computation and Language
url https://arxiv.org/abs/2510.11444