Hurtful Words: Quantifying Biases in Clinical Contextual Word Embeddings.
Haoran ZhangAmy X. LuMohamed AbdallaMatthew B. A. McDermottMarzyeh GhassemiPublished in: CoRR (2020)
Keyphrases
- related words
- n gram
- english words
- word recognition
- word pairs
- word sense disambiguation
- word meaning
- unknown words
- word frequencies
- word segmentation
- multiword
- lexical information
- linguistic information
- text corpus
- word similarity
- chinese word segmentation
- word spotting
- stop words
- linguistic knowledge
- syntactic categories
- context sensitive
- noun phrases
- query words
- word co occurrence
- keywords
- speech recognition systems
- word meanings
- contextual information
- word level
- handwritten words
- spoken document retrieval
- natural language text
- distributional clustering
- compound words
- clinical data
- printed text
- co occurrence
- lexical features
- frequency counts
- latent topics
- chinese text
- automatic transcription
- language specific
- natural language processing
- language model
- training corpus
- vector space
- text documents
- out of vocabulary
- word frequency
- bilingual dictionaries
- parallel corpus
- historical documents
- translation model
- handwriting recognition
- language independent
- character recognition
- semantic relations
- text categorization
- dimensionality reduction