Sparse optimization for green edge ai inference

Keyphrases

• Sparse Optimization 100%
• Edge AI 100%
• Power Consumption 50%
• Task Selection 50%
• Group Sparsity 50%
• Sum Function 50%
• Logsum 50%
• Function-based 25%
• Energy Efficient 25%
• Energy Efficiency 25%
• Learning Task 25%
• Computational Power 25%
• Combinatorial Optimization Problem 25%
• Transmission Power 25%
• Mobile Users 25%
• Energy Efficiency Improvement 25%
• Computing Capabilities 25%
• Low Latency 25%
• Downlink Beamforming 25%
• Deep Learning 25%
• Intelligent Services 25%
• Overall Power 25%
• Edge Computing 25%
• Network Edge 25%
• Inference Tasks 25%
• Ergodic 25%
• Three-stage Approach 25%
• Transmit Beamforming 25%
• Inference System 25%
• Beamforming Vector 25%
• Effective Edge 25%
• Group Sparse Beamforming 25%
• Proximal Iteratively Reweighted Algorithm 25%
• Worst-case Convergence Rate 25%
• Global Convergence Analysis 25%
• Joint Inference 25%
• Challenging Problems 25%

Engineering

• Artificial Intelligence 100%
• Beamforming 100%
• Electric Power Utilization 66%
• Group Sparsity 66%
• Function Sum 66%
• Energy Efficiency 66%
• Network Edge 33%
• Optimisation Problem 33%
• Simulation Result 33%
• Convergence Rate 33%
• Power Transmission 33%
• Learning Task 33%
• Mobile User 33%
• Joints (Structural Components) 33%
• Edge Computing 33%
• Deep Learning Method 33%

Computer Science

• Artificial Intelligence 100%
• Power Consumption 66%
• Sparsity 66%
• Task Selection 66%
• Energy Efficiency 66%
• Energy Efficient 33%
• Combinatorial Optimization Problem 33%
• Deep Learning Method 33%
• Global Convergence 33%
• Transmission Power 33%
• Inference System 33%
• Mathematical Convergence 33%
• Inference Task 33%
• Edge Computing 33%
• Convergence Rate 33%