Noisy-Labeled NER with Confidence Estimation.
Kun LiuYao FuChuanqi TanMosha ChenNingyu ZhangSongfang HuangSheng GaoPublished in: NAACL-HLT (2021)
Keyphrases
- named entity recognition
- training data
- named entities
- maximum entropy
- information extraction
- estimation accuracy
- high confidence
- training set
- confidence level
- supervised learning
- parameter estimation
- conditional random fields
- natural language processing
- estimation error
- manually labeled
- machine learning
- proper names
- confidence measure
- manual labeling
- accurate estimation
- text summarization
- automatic extraction
- estimation algorithm
- active learning
- association rules
- expert systems
- learning algorithm