Efficient Communication-Computation Tradeoff for Split Computing: A Multi-Tier Deep Reinforcement Learning Approach

Keyphrases

• Multi-tier 100%
• Reinforcement Learning Approach 100%
• Efficient Communication 100%
• Deep Reinforcement Learning (deep RL) 100%
• Split Computing 100%
• User Equipment 60%
• Model Accuracy 40%
• Computation Capability 40%
• Splitting Point 40%
• Learning Latency 40%
• Energy Consumption 20%
• Neural Network 20%
• Recognition Task 20%
• Computational Load 20%
• Performance Optimization 20%
• Resource-constrained 20%
• Image Recognition 20%
• Promising Technology 20%
• Communication Link 20%
• Neural Network Training 20%
• Learning Scheme 20%
• Comprehensive Simulation 20%
• Computational Resources 20%
• Sub-optimization 20%
• Latency Model 20%
• Multiple Stations 20%
• Intelligent Services 20%
• Training Tasks 20%
• Edge Server 20%

Computer Science

• Learning Approach 100%
• Model Accuracy 100%
• Computation Capability 100%
• Deep Reinforcement Learning 100%
• Neural Network 50%
• Energy Consumption 50%
• Communication Link 50%
• Learning Scheme 50%
• Performance Optimization 50%
• Computation Load 50%
• Neural Network Training 50%
• Edge Server 50%