@inproceedings{yang-etal-2024-pga,
    title = "{PGA}-{S}ci{RE}:基于大语言模型的数据增强框架进行科学领域的关系({PGA}-{S}ci{RE}:Harnessing {LLM} on Data Augmentation for Enhancing Scientific Relation Extraction)",
    author = "Yang, Zhou  and
      Shimin, Dan  and
      Hongkui, Wei  and
      Zhehuan, Zhao  and
      Wenshuo, Feng",
    editor = "Sun, Maosong  and
      Liang, Jiye  and
      Han, Xianpei  and
      Liu, Zhiyuan  and
      He, Yulan",
    booktitle = "Proceedings of the 23rd Chinese National Conference on Computational Linguistics (Volume 1: Main Conference)",
    month = jul,
    year = "2024",
    address = "Taiyuan, China",
    publisher = "Chinese Information Processing Society of China",
    url = "https://preview.aclanthology.org/jlcl-multiple-ingestion/2024.ccl-1.27/",
    pages = "352--369",
    language = "zho",
    abstract = "{\textquotedblleft}关系提取旨在识别文本中提到的实体对之间的关系。大语言模型的进步对自然语言处理任务产生了巨大的影响。在这项工作中,我们针对科学领域的关系抽取任务,提出一个名为PGA的数据增强框架,用于提升模型在科学领域的关系抽取的性能。框架引入了两种数据增强的方式,利用大语言模型通过转述原训练集样本,得到句意相同但具备不同表述和形式的伪样本。以及指导大语言模型根据原训练集样本的关系和实体标签,生成暗含对应标签信息的句子。这两种伪样本分别与原数据集共同参与关系抽取模型的训练。实验中PGA框架提高了三个主流模型的科学领域内关系抽取的F1分数。同时,使用大语言模型获得样本也能有效减少人工标注数据的成本。{\textquotedblright}"
}