Efficient and modularized training on FPGA for real-time applications

Keyphrases

• Time Application 100%
• Modularized 100%
• GPU 100%
• Energy Efficiency 33%
• Online Learning 33%
• Training Module 33%
• Intel 33%
• Training Procedure 33%
• Segmentation Method 33%
• Architecture Optimization 33%
• Complex Training 33%
• 16-bit 33%
• High Energy Efficiency 33%
• Training Tasks 33%
• Conventional Training 33%
• Deep Convolutional Neural Network (deep CNN) 33%
• Real-time Learning 33%
• Online Learning Algorithm 33%
• Convolution Neural Network 33%
• CIFAR-10 33%
• Stratix 10 33%
• FPGA Acceleration 33%
• Training Accelerator 33%
• Energy Restriction 33%
• Low Computational Cost 33%
• Model Segmentation 33%
• Segmented Learning 33%

Computer Science

• Real-Time Application 100%
• Field Programmable Gate Arrays 100%
• Graphics Processing Unit 60%
• Convolutional Neural Network 40%
• Energy Efficiency 40%
• Electronic Learning 40%
• Computation Cost 20%
• Training Module 20%
• Learning Algorithm 20%
• Fixed Points 20%
• Segmentation Technique 20%