3-Dimensional convolutional neural networks for predicting StarCraft II results and extracting key game situations

Keyphrases

• Game Situations 100%
• StarCraft II 100%
• 3-dimensional Convolutional Neural Network 100%
• 3D Residual Network 75%
• Gradient-weighted Class Activation Mapping (Grad-CAM) 50%
• Real-time Strategy Games 50%
• Data Providing 25%
• Popular 25%
• Outcome Prediction 25%
• Prediction Model 25%
• Video Data 25%
• Performance Prediction 25%
• Extracting Information 25%
• Result Prediction 25%
• Multiple Time Points 25%
• Spatial-temporal Data 25%
• Game Player 25%
• Game Outcome 25%
• Residual Network 25%
• Resource Control 25%
• Winning Strategy 25%

Computer Science

• Residual Neural Network 100%
• 3-Dimensional Convolutional Neural Networks 100%
• Information Point 50%
• Prediction Model 25%
• Extract Information 25%
• Prediction Performance 25%
• Gradient-Weighted Class Activation Mapping 25%
• Winning Strategy 25%
• Feature Map 25%