BrambleXu / knowledge-graph-learning

A curated list of awesome knowledge graph tutorials, projects and communities.
MIT License
735 stars 120 forks source link

ACL-2018-Chinese NER Using Lattice LSTM #279

Open BrambleXu opened 4 years ago

BrambleXu commented 4 years ago

Summary:

把charecter和word信息通过LSTM学习,然后结合二者得到了Lattice LSTM。

Resource:

Paper information:

Notes:

主要使用LSTM-CRF构架,BIOES标注方法。

相比基于字符的方法(character-based),能够充分利用单词和词序信息;相比基于字的方法(word-based),不会因为分词错误影响识别结果。

Model Graph:

image

Result:

Thoughts:

Next Reading: