DiFiNet: Boundary-Aware Semantic Differentiation and Filtration Network for Nested Named Entity Recognition

Overview

This is the code for DiFiNet: Boundary-Aware Semantic Differentiation and Filtration Network for Nested Named Entity Recognition, accepted by ACL 2024.

Requirements

GPU=NVIDIA A100 Tensor Core
beautifulsoup4==4.9.3
FastNLP==1.0.1
fitlog==0.9.15
nltk==3.8.1
numpy==1.24.4
pandas==1.1.3
sparse==0.14.0
torch==1.13.1+cu117
torch_scatter==2.0.9
tqdm==4.65.0
transformers==4.20.1

Preprocess your datasets

Put ACE datasets in the preprocess/data directory, following a similar structure as demonstrated in CNN_Nested_NER

For GENIA dataset, we use the version from W2NER

Train

   bash train_arg_{dataset}.sh

The item {dataset} can be replaced with "04", "05" or "genia". The experiment results can be found in the directory logs after initiating the training process.

Citation

@inproceedings{cai-etal-2024-difinet,
    title = "{D}i{F}i{N}et: Boundary-Aware Semantic Differentiation and Filtration Network for Nested Named Entity Recognition",
    author = "Cai, Yuxiang  and
      Liu, Qiao  and
      Gan, Yanglei  and
      Lin, Run  and
      Li, Changlin  and
      Liu, Xueyi  and
      Luo, Da  and
      Jiaye, Yang",
    editor = "Ku, Lun-Wei  and
      Martins, Andre  and
      Srikumar, Vivek",
    booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    month = aug,
    year = "2024",
    address = "Bangkok, Thailand",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2024.acl-long.349",
    pages = "6455--6471"
}

Have any Questions？Please email cyx_yyy at foxmail dot com

Acknowledge

The code of CNN_Nested_NER