PROVABLE FILTER PRUNING FOR EFFICIENT NEURAL NETWORKS

Engineering & Materials Science

• Importance sampling 54%
• Neural networks 50%
• Compressibility 47%
• Network performance 40%
• Network architecture 39%
• Convolutional neural networks 38%
• Sampling 29%
• Set theory 26%

Arts & Humanities

• Filter 87%
• Neural Networks 80%
• Sampling 22%
• Datum Point 20%
• Performance 15%
• Layer 11%
• Evaluation 7%

Social Sciences

• neural network 100%
• performance 15%
• guarantee 14%
• evaluation 7%