Dystri: A Dynamic Inference based Distributed DNN Service Framework on Edge

Keyphrases

• Network Services 100%
• Dynamic Inference 100%
• Distributed Deep Neural Networks 100%
• Service Framework 100%
• Deep Neural Network 50%
• Inference Methods 50%
• Deep Neural Network Inference 50%
• Edge Infrastructure 50%
• Quality of Service 25%
• Network Applications 25%
• Mobile Devices 25%
• Accuracy Improvement 25%
• Computational Cost 25%
• Multi-user 25%
• Multi-user Setting 25%
• User Heterogeneity 25%
• Existing State 25%
• Neural Network 25%
• Crucial Aspect 25%
• Number of Users 25%
• Computational Capability 25%
• Heterogeneous Devices 25%
• Distributed Controller 25%
• Edge Devices 25%
• Neural Network Applications 25%
• Device Type 25%
• User Request 25%
• Service Type 25%
• Heterogeneous Users 25%
• Content-based 25%
• Service Platform 25%
• Deadline Miss 25%
• Near-zero 25%
• Local Resources 25%
• Innovative Framework 25%
• Heterogeneous Services 25%
• Excel 25%
• Flexible Control 25%
• Instantaneous Control 25%
• Deep Inference 25%
• Discrete Control 25%
• Inference Service 25%
• Application Inference 25%

Computer Science

• Network Service 100%
• Deep Neural Network 100%
• Inference Technique 40%
• Network Inference 40%
• Quality of Service 20%
• Computational Cost 20%
• Service Platform 20%
• Neural Network Application 20%