A hierarchical deep learning model for predicting pedestrian-level urban winds

Keyphrases

• Baseline Predictors 66%
• Global Flows 33%
• Hierarchical CNN 33%
• High-frequency Details 33%
• Kernel Size 33%
• L1 Loss 33%
• Multi-scale Architecture 33%
• Network Mapping 33%
• Peak Wind Speed 33%
• Pedestrian Level 100%
• Quantitative Enhancement 33%
• Sharp Peak 33%
• Simultaneous Learning 33%
• Turbulent Wind 33%
• Urban Geometry 33%
• Urban Wind 100%
• Wind Jet 33%
• Wind Statistics 33%

Computer Science

• Conditional Generative Adversarial Network 100%
• Convolutional Neural Network 50%
• Deep Learning Method 50%
• Deep Learning Model 100%
• Frequency Content 50%
• Hierarchical Model 50%
• Interactive 50%
• local feature 50%
• Prediction Accuracy 50%
• U-Net 50%
• Validation Set 50%
• Web Portal 50%

Engineering

• Kernel Size 50%