CETA: A Consensus Enhanced Training Approach for Denoising in Distantly Supervised Relation Extraction.
Ruri LiuShasha MoJianwei NiuShengda FanPublished in: COLING (2022)
Keyphrases
- relation extraction
- denoising
- weakly supervised
- automatic extraction
- information extraction
- named entities
- supervised learning
- semantic relations
- domain specific
- semantic role labeling
- named entity recognition
- question answering
- dependency trees
- semantic features
- tree kernels
- semi supervised
- broad coverage
- image processing
- unsupervised learning
- machine learning
- co occurrence
- high level
- structured data
- named entity recognizer
- parse tree
- semantic information
- training set
- object recognition
- training data
- learning algorithm