Rescue Implicit and Long-tail Cases: Nearest Neighbor Relation Extraction
Rescue Implicit and Long-tail Cases: Nearest Neighbor Relation Extraction
复制标题
DOI：
10.48550/arxiv.2210.11800
复制
发表时间：
2022-10
期刊：
影响因子：
--
通讯作者：
Zhen Wan;Qianying Liu;Zhuoyuan Mao;Fei Cheng;S. Kurohashi;Jiwei Li
中科院分区：
文献类型：
--
作者：
Zhen Wan;Qianying Liu;Zhuoyuan Mao;Fei Cheng;S. Kurohashi;Jiwei Li
关系抽取（RE）在预训练语言模型的帮助下取得了显着的进展。然而，现有的RE模型通常无法处理两种情况：隐式表达式和长尾关系类型，由语言复杂性和数据稀疏性造成的。在本文中，我们介绍了一个简单的增强RE使用k近邻（kNN-RE）。kNN-RE允许模型在测试时通过最近邻搜索来咨询训练关系，并提供了一种简单而有效的方法来解决上述两个问题。此外，我们观察到，kNN-RE作为一种有效的方式来利用远程监督（DS）数据的RE。实验结果表明，所提出的kNN-RE在各种监督RE数据集上实现了最先进的性能，即，ACE 05、SciERC和Wiki 80，沿着在允许使用DS的情况下，在i2 b2和Wiki 80数据集上的表现优于迄今为止最好的模型。我们的代码和模型可在https://github.com/YukinoWan/kNN-RE上获得。
Relation extraction (RE) has achieved remarkable progress with the help of pre-trained language models. However, existing RE models are usually incapable of handling two situations: implicit expressions and long-tail relation types, caused by language complexity and data sparsity. In this paper, we introduce a simple enhancement of RE using k nearest neighbors (kNN-RE). kNN-RE allows the model to consult training relations at test time through a nearest-neighbor search and provides a simple yet effective means to tackle the two issues above. Additionally, we observe that kNN-RE serves as an effective way to leverage distant supervision (DS) data for RE. Experimental results show that the proposed kNN-RE achieves state-of-the-art performances on a variety of supervised RE datasets, i.e., ACE05, SciERC, and Wiki80, along with outperforming the best model to date on the i2b2 and Wiki80 datasets in the setting of allowing using DS. Our code and models are available at: https://github.com/YukinoWan/kNN-RE.