Maximum output discrepancy computation for convolutional neural network compression

Mathematics

• Compression 84%
• Discrepancy 83%
• Neural Networks 78%
• Output 62%
• Network Performance 19%
• Reachability Analysis 17%
• Pooling 15%
• Safety 13%
• Reachability 12%
• Training 12%
• Experimental Data 12%
• Correctness 11%
• Quantization 11%
• Convolution 9%
• Minimise 9%
• Experiment 8%
• Costs 8%
• Numerical Examples 7%
• Model 3%

Business & Economics

• Neural Networks 100%
• Compression 96%
• Discrepancy 79%
• Quantization 17%
• Convolution 16%
• Pooling 13%
• Incidents 11%
• Assurance 11%
• Safety 10%
• Experiment 7%
• Performance 4%
• Costs 4%

Engineering & Materials Science

• Convolutional neural networks 75%
• Neural networks 33%
• Convolution 27%
• Network performance 26%
• Concretes 18%
• Costs 10%
• Experiments 9%