A deep contrastive learning approach to extremely-sparse disaster damage assessment in social sensing

Keyphrases

• Learning Approaches 100%
• Social Sensing 100%
• Extremely Sparse 100%
• Disaster Damage Assessment 100%
• Deep Contrastive Learning 100%
• Damage Severity 40%
• Damage Assessment 20%
• Online Social Media 20%
• Evaluation Results 20%
• Semi-supervised Learning 20%
• Fundamental Challenges 20%
• Training Data 20%
• Physical World 20%
• Affected Area 20%
• Timely Information 20%
• Training Samples 20%
• Human Sensors 20%
• Deep Learning 20%
• Augmentative 20%
• Neural Network Architecture 20%
• Disaster Events 20%
• Training Quality 20%
• Imagery Data 20%
• Sparse Training 20%
• Sparse Training Data 20%

Computer Science

• Learning Approach 100%
• Training Data 100%
• Contrastive Learning 100%
• Evaluation Result 50%
• Physical World 50%
• Neural Network Architecture 50%
• Semisupervised Learning 50%
• Application Scenario 50%
• Training Sample 50%
• Related Feature 50%
• Deep Learning Method 50%