CASE: CNN Acceleration for Speech-Classification in Edge-Computing

Mathematics

• Speech 95%
• Neural Networks 68%
• Accelerator 50%
• Computing 45%
• Latency 31%
• Chip 30%
• Server 25%
• Hardware Accelerator 24%
• Network Performance 20%
• Field Programmable Gate Array 18%
• Speech Recognition 18%
• High Performance 15%
• Engine 15%
• Learning Algorithm 14%
• Machine Learning 13%
• High Accuracy 12%
• Computational Cost 12%
• Model 11%
• Life 10%
• Requirements 9%
• Metric 7%
• Performance 7%
• Framework 6%
• Standards 6%

Business & Economics

• Accelerator 100%
• Neural Networks 86%
• Latency 46%
• Speech Recognition 29%
• Machine Learning 23%
• Learning Algorithm 21%
• Hardware 17%
• High Performance 17%
• Inference 15%
• Performance 13%
• Costs 6%
• Factors 6%

Engineering & Materials Science

• Edge computing 89%
• Convolutional neural networks 65%
• Particle accelerators 48%
• Internet of things 30%
• Servers 17%
• Network performance 11%
• Speech recognition 11%
• Learning algorithms 10%
• Field programmable gate arrays (FPGA) 9%
• Machine learning 8%
• Engines 7%
• Data storage equipment 7%
• Computer hardware 7%