A Layer-wise Training and Pruning Method for Memory Efficient On-chip Learning Hardware

Keyphrases

• Training Methods 100%
• Pruning Strategy 100%
• Memory Efficiency 100%
• Deep Neural Network 100%
• On-chip Learning 100%
• Layer-wise Training 100%
• Memory Space 66%
• Pruning 66%
• Limited Memory 66%
• Memory Resource 33%
• Physical Memory 33%
• Learning Methods 33%
• Classification Accuracy 33%
• Conventional Algorithm 33%
• Conventional Learning 33%
• Projection Algorithm 33%
• Initialize 33%
• MNIST 33%
• Edge Devices 33%
• Magnitude-based 33%
• Physical Weight 33%
• Target Projection 33%
• Random Target 33%

Computer Science

• Deep Neural Network 100%
• Computer Hardware 100%
• Memory Resource 33%
• Physical Memory 33%
• Classification Accuracy 33%
• Conventional Algorithm 33%
• Conventional Learning 33%