An Analytical Method to Determine Minimum Per-Layer Precision of Deep Neural Networks

Keyphrases

• Analytical Method 100%
• Order of Accuracy 100%
• Deep Neural Network 100%
• Numerical Results 50%
• Complexity Reduction 50%
• Fast Inference 50%
• Efficient Inference 50%
• Minimal Cost 50%
• 8-bit 50%
• Reduced Precision 50%
• Precision Analysis 50%
• Typical Model 50%
• MNIST 50%
• Resource-constrained Platforms 50%
• Accuracy Degradation 50%
• CIFAR10 Dataset 50%
• Binary Networks 50%
• Deep Learning System 50%
• Compression Mechanism 50%
• Fixed Point Network 50%

Computer Science

• Deep Neural Network 100%
• Learning Systems 50%
• Fixed Points 50%
• Deep Learning 50%
• Complexity Reduction 50%
• Binary Network 50%