Multi-step Transfer Learning in Natural Language Processing for the Health Domain

Keyphrases

• Natural Language Processing 100%
• Transfer Learning 100%
• Health Domain 100%
• Verbal Autopsy 100%
• Autopsy Report 60%
• Cause of Death 40%
• Word Embedding 40%
• Learning Task 40%
• Feature Extraction 40%
• Natural Language Processing Techniques 40%
• Bidirectional Encoder Representations from Transformers 40%
• Learning Words 40%
• Biomedical Domain 40%
• ELMo 40%
• Clinical Data 20%
• Performance Improvement 20%
• Classification Task 20%
• Healthcare Facilities 20%
• Multiple Domains 20%
• Initialize 20%
• Registration System 20%
• Restricted Access 20%
• Text Representation 20%
• Narrative Text 20%
• Death Registration 20%
• Healthcare Domain 20%
• Privacy Policy 20%
• Word-level 20%
• Character-level 20%
• Domain Adaptation 20%
• Data Annotation 20%
• Patient Confidentiality 20%
• Patient Privacy 20%
• Model Embedding 20%
• Two-component Domains 20%
• Confidentiality Policy 20%
• Non-clinicians 20%
• Access to Data 20%
• Technique Feature 20%
• General English 20%
• Transformer-Based Embeddings 20%
• Multi-step Learning 20%
• Field Agents 20%
• Cause of Death Classification 20%

Computer Science

• Natural Language Processing 100%
• Transfer Learning 100%
• Bidirectional Encoder Representations From Transformers 66%
• Word Embedding 66%
• Embedding From Language Model 66%
• Annotation 33%
• Multiple Domain 33%
• Restricted Access 33%
• Case Study 33%
• Classification Task 33%
• Domain Adaptation 33%
• Registration System 33%
• Feature Extraction 33%
• Clinical Data 33%
• Text Representation 33%