Highly accurate classification of chest radiographic reports using a deep learning natural language model pre-trained on 3.8 million text reports

Keyphrases

• Highly Accurate 100%
• Deep Learning 100%
• Accurate Classification 100%
• Natural Language Model 100%
• Bidirectional Encoder Representations from Transformers 100%
• Labeled Data 50%
• Natural Language Processing 50%
• Transformer Model 50%
• Neural Network 25%
• Source Code 25%
• Machine Learning 25%
• Unlabeled Data 25%
• Image Classification 25%
• Freely Available 25%
• Chest Radiograph 25%
• Processing Methods 25%
• Clinical Workflow 25%
• Congestion 25%
• Information Extraction 25%
• Outperformance 25%
• Receiver Operating Characteristic Curve 25%
• Radiology 25%
• Text Classification 25%
• Intensive Care 25%
• Free Text 25%
• Pneumothorax 25%
• Effusion 25%
• Medical Reports 25%
• Free-text Data 25%
• Deep Bidirectional Transformer 25%
• Annotation Effort 25%

Computer Science

• Language Modeling 100%
• Bidirectional Encoder Representations From Transformers 100%
• Deep Learning Method 100%
• Natural Language Processing 50%
• Transformer Model 50%
• Neural Network 25%
• Annotation 25%
• Unlabeled Data 25%
• Image Classification 25%
• Information Extraction 25%
• Text Classification 25%
• Processing Method 25%
• Characteristic Curve 25%
• Machine Learning 25%
• Learning System 25%

Engineering

• Natural Language Processing 100%
• Deep Learning Method 100%
• Image Classification 50%
• Processing Method 50%
• Good Result 50%
• Text Data 50%
• Characteristic Curve 50%
• Learning System 50%