先使用Chinese Word Segmentation (CWS) first, and then to use a word-based NER model. However, the pipeline method suffers from error propagation, since the error of CWS may affect the performance of NER
The second one is to learn CWS and NER tasks jointly. However, the joint models must rely on CWS annotation datasets, which are costly and are annotated under many diverse segmentation criteria
The third one is to leverage an automatically constructed lexicon, which is pre-trained on large automatically segmented texts. Lexical knowledge includes boundaries and semantic information. Boundaries information is provided by the lexicon word itself, and semantic information is provided by pre-trained word embeddings
第三种使用lexicon的方法有两个问题:
The first challenge is to integrate self-matched lexical words. 比如“北京机场”这个词里的,“机场”
The second challenge is to integrate the nearest contextual lexical words direct. 通过距离lexical words最近的上下文单词来辅助预测。
Model Graph:
Result::
Thoughts:
EMNLP-2019/11-Leverage Lexical Knowledge for Chinese Named Entity Recognition via Collaborative Graph Network #278
Summary:
这篇论文讲如何将word level信息整合到NER模型里。
Resource:
Paper information:
Author: Dianbo Sui1,2 , Yubo Chen1 , Kang Liu1,2 , Jun Zhao1,2 , Shengping Liu3
Dataset:
keywords:
Notes:
中文不想英文,每个word之间有明确的间隔,所以大家针对中文做NER的时候,很自然地去用char level的feature。但是就这样舍弃word信息并不好,所以如何利用好这些信息就是这篇论文的创新点。
有三种方法将word信息导入到NER里:
第三种使用lexicon的方法有两个问题:
Model Graph:
Result::
Thoughts:
这三篇文章虽然都用到了lexicon,但是都还是基于sequence labeling的框架内的方案。而AutoNER则是新的方法。
Next Reading: