TraNNsformer: Clustered Pruning on Crossbar-Based Architectures for Energy-Efficient Neural Networks

Computer Science

• Neural Network 100%
• Energy Efficient 100%
• Deep Neural Network 100%
• Multilayer Perceptron 60%
• Convolutional Neural Network 60%
• Network Topology 20%
• Energy Consumption 20%
• Energy Saving 20%
• Computer Hardware 20%
• Reliable Operation 20%

Engineering

• Deep Neural Network 100%
• Convolutional Neural Network 60%
• Perceptron 60%
• Energy Conservation 40%
• Realization 40%
• Neuromorphic System 40%
• Reliable Operation 20%
• Network Topology 20%
• Array Size 20%
• Energy Benefit 20%
• Connectivity Matrix 20%
• Original Network 20%

Keyphrases

• Energy-efficient Neural Network 100%
• Network Pruning 50%
• Irregular Connection 25%
• Sparse Deep Neural Networks 25%
• Redundant Connection 25%
• Multi-layer Convolutional Neural Network 25%
• Memristive Technology 25%
• Crossbar Structure 25%