A Progressive Subnetwork Searching Framework for Dynamic Inference

Medicine & Life Sciences

• Datasets 100%
• Saturn 69%
• Work Performance 55%
• Computer Neural Networks 54%
• Entropy 50%
• Resource Allocation 50%
• Workload 42%
• Acceleration 41%
• Noise 38%

Engineering & Materials Science

• Deep neural networks 87%
• Computer hardware 31%
• Tuning 27%
• Sampling 25%
• Graphics processing unit 16%
• Program processors 15%
• Resource allocation 15%
• Entropy 15%
• Neural networks 10%
• Geometry 10%