WSNet: Compact and efficient networks through weight sampling

Mathematics

• Learning 97%
• Baseline 77%
• Sharing 75%
• Neural Networks 67%
• Filter 63%
• Learning Process 49%
• Model 46%
• Pruning 45%
• Performance 42%
• Architecture 39%
• Quantization 34%
• Compact Set 33%
• Factorization 29%
• Convolution 29%
• Verify 29%
• Experiment 26%
• Demonstrate 21%

Engineering & Materials Science

• Sampling 100%
• Factorization 35%
• Convolution 35%
• Deep neural networks 34%
• Convolutional neural networks 32%
• Set theory 22%
• Experiments 12%