Efficient continual learning at the edge with progressive segmented training

Engineering

• Robotic System 100%
• Field Programmable Gate Arrays 100%
• Power Budget 100%
• Effective Approach 100%
• Model Parameter 100%
• Dynamic Structure 100%
• Drone 100%
• Regularization 100%
• Network Model 100%
• Learning Algorithm 100%
• Convolutional Neural Network 100%
• Data Stream 100%

Keyphrases

• Continual Learning 100%
• Segmented Learning 100%
• Single Network 40%
• CIFAR-10 40%
• System Dynamics 20%
• Network Model 20%
• Dynamic Structure 20%
• Robotic System 20%
• Learning Algorithm 20%
• Intel 20%
• Self-driving Cars 20%
• Multi-task 20%
• Limited Power 20%
• Effective Approach 20%
• Computational Efficiency 20%
• Power Budget 20%
• Single Model 20%
• New Task 20%
• Model-based Segmentation 20%
• Secondary Groups 20%
• Single Head 20%
• Future Learning 20%
• CIFAR-100 20%
• Network Segmentation 20%
• Catastrophic Forgetting 20%
• Redundant Capacity 20%
• Surveillance Drones 20%
• Stratix 10 20%
• Vehicle Surveillance 20%
• Drone System 20%

Computer Science

• Continual Learning 100%
• Single Network 40%
• Data Stream 20%
• Multiple Task 20%
• Convolutional Neural Network 20%
• Network Segmentation 20%
• Redundant Capacity 20%
• Drone Surveillance 20%
• System Dynamics 20%
• Learning Algorithm 20%
• Regularization 20%
• Effective Approach 20%
• Dynamic Structure 20%
• Field Programmable Gate Arrays 20%