ACL 2025short0 citations
Enhancing NER by Harnessing Multiple Datasets with Conditional Variational Autoencoders
Abstract
We propose a novel method to integrate a Conditional Variational Autoencoder (CVAE) into a span-based Named Entity Recognition (NER) model to model the shared and unshared information among labels in multiple datasets and ease the training on the datasets. Experimental results using multiple biomedical datasets show the effectiveness of the proposed method, achieving improved performance on the BioRED dataset.
BibTeX
@inproceedings{oi-miwa-2025-enhancing,
title = "Enhancing {NER} by Harnessing Multiple Datasets with Conditional Variational Autoencoders",
author = "Oi, Taku and
Miwa, Makoto",
editor = "Che, Wanxiang and
Nabende, Joyce and
Shutova, Ekaterina and
Pilehvar, Mohammad Taher",
booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)",
month = jul,
year = "2025",
address = "Vienna, Austria",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.acl-short.87/",
doi = "10.18653/v1/2025.acl-short.87",
pages = "1107--1117",
ISBN = "979-8-89176-252-7"
}