Distant supervision tackles the data bottleneck in NER by automatically generating training instances via dictionary matching. Unfortunately, the learning of DS-NER is severely dictionary-biased, which suffers from spurious correlations and therefore undermines the effectiveness and the robustness of the learned models. In this paper, we fundamentally explain the dictionary bias via a Structural Causal Model (SCM), categorize the bias into intra-dictionary and inter-dictionary biases, and identify their causes. Based on the SCM, we learn de-biased DS-NER via causal interventions. For intra-dictionary bias, we conduct backdoor adjustment to remove the spurious correlations introduced by the dictionary confounder. For inter-dictionary bias, we propose a causal invariance regularizer which will make DS-NER models more robust to the perturbation of dictionaries. Experiments on four datasets and three DS-NER models show that our method can significantly improve the performance of DS-NER.
翻译:不幸的是,DS-NER的学习具有严重的字典偏见,这有虚假的关联性,因此破坏了所学模型的有效性和稳健性。在本文中,我们通过结构界别模型(SCM)从根本上解释字典偏见,将偏差分类为辖区内和辖区间偏差,并找出其原因。根据SCM,我们学习通过因果干预消除偏差的 DS-NER。对于词典内部偏差,我们进行后门调整,以消除词典汇合者引入的虚假关联性。关于词典间偏差,我们建议采用因果调节器,使DS-NER模型更能适应词典的渗透性。关于四个数据集和三个DS-NER模型的实验表明,我们的方法可以大大改善DS-NER的性能。