DrFuse: Learning Disentangled Representation for Clinical Multi-Modal Fusion with Missing Modality and Modal Inconsistency

Keyphrases

• Inconsistency 100%
• Modal 100%
• Multimodal Fusion 100%
• Learning Disentangled Representations 100%
• Missing Modalities 100%
• Electronic Health Records 66%
• Medical Image 66%
• Data Modality 66%
• Prediction Task 33%
• Large-scale Dataset 33%
• Machine Learning Models 33%
• Model Performance 33%
• State-of-the-art Models 33%
• Clinical Prediction Model 33%
• Complementary Nature 33%
• Clinical Factors 33%
• Attention Layer 33%
• Administrative Factors 33%
• MIMIC-CXR 33%

Computer Science

• Experimental Result 100%
• Performance Model 100%
• Attention (Machine Learning) 100%
• Machine Learning Model 100%