Processing-In-Memory Acceleration of Convolutional Neural Networks for Energy-Effciency, and Power-Intermittency Resilience

Keyphrases

• Accumulation Method 33%
• Baseline Design 33%
• Battery-free 33%
• Bitwise 33%
• CMOS-only 33%
• Convolutional Layer 33%
• Convolutional Neural Network 100%
• Effciency 100%
• Energy Consumption Reduction 33%
• Forward Progress 33%
• High Energy Efficiency 66%
• In-memory 100%
• Inference Accuracy 33%
• Intermittency 33%
• IoT Nodes 33%
• Low Bitwidth 33%
• Magnetic Random Access Memory 33%
• Neural Network Inference 33%
• Partial State Information 33%
• Power Intermittency 100%
• Resilience 100%
• Resistive Random Access Memory (ReRAM) 33%
• Spin-orbit Torque 33%

Computer Science

• Accelerator Memory 33%
• Baseline Design 33%
• Convolutional Layer 33%
• Convolutional Neural Network 100%
• Energy Consumption 33%
• Energy Efficiency 66%
• In-Memory Processing 100%
• Internet of Things 33%
• Network Inference 33%
• Random Access Memory 33%
• State Information 33%

Engineering

• Battery (Electrochemical Energy Engineering) 33%
• Convolutional Layer 33%
• Convolutional Neural Network 100%
• Energy Engineering 100%
• High Energy Efficiency 66%
• Internet of Things 33%
• Nodes 33%
• Random Access Memory 33%
• Simulation Result 33%
• State Information 33%