Calibration-Aided Edge Inference Offloading via Adaptive Model Partitioning of Deep Neural Networks

Keyphrases

• Adaptation Models 100%
• Calibration Algorithm 25%
• Communication Delay 25%
• Deep Neural Network 100%
• Early Exit 50%
• Edge Inference 100%
• Energy Conservation 25%
• Estimation Accuracy 25%
• Hardware Constraints 25%
• Inference Accuracy 25%
• Inference Time 25%
• Mobile Devices 50%
• Model Partition 100%
• Network Inference 25%
• Neural Network 25%
• Offloading Decision 25%
• Prediction Accuracy 25%
• Via Modeling 25%

Computer Science

• Communication Delay 16%
• Deep Neural Network 100%
• Mobile Device 33%
• Prediction Accuracy 16%