CycleNER: An Unsupervised Training Approach for Named Entity Recognition.
Andrea IovineAnjie FangBesnik FetahuOleg RokhlenkoShervin MalmasiPublished in: WWW (2022)
Keyphrases
- named entity recognition
- semi supervised
- information extraction
- natural language processing
- named entities
- supervised learning
- text summarization
- maximum entropy
- sequence labeling
- conditional random fields
- relation extraction
- annotated corpus
- training set
- unsupervised learning
- proper names
- pos tagging
- text mining
- maximum entropy classifier
- classifier ensemble
- domain specific
- active learning
- training samples
- co occurrence
- pairwise
- association rules
- information retrieval
- data sets