A 2.25 TOPS/W fully-integrated deep CNN learning processor with on-chip training

Keyphrases

• Energy Efficiency 100%
• Fully Integrated 100%
• Processing Element 100%
• Deep Convolutional Neural Network (deep CNN) 100%
• Neural Network Learning 100%
• On-chip Training 100%
• Proposed Design 50%
• Processing Latency 50%
• High Energy Efficiency 50%
• Convolutional Neural Network 50%
• Memory Consumption 50%
• 40nm CMOS 50%
• Softmax 50%
• Security Privacy 50%
• High Security 50%
• Floating-point Arithmetic 50%
• Limited Area 50%
• Clock Gating 50%
• Data Gating 50%
• Low-precision 50%
• Max Pooling 50%
• Rectified Linear Unit (ReLU) 50%
• Art Learning 50%
• Unified Processing 50%
• Design Abstraction 50%
• Deep Learning Processor 50%

Computer Science

• Energy Efficiency 100%
• Convolutional Neural Network 100%
• Processing Element 66%
• Fixed Points 33%
• Floating Point 33%
• Security and Privacy 33%
• Design Abstraction 33%
• Softmax Function 33%
• Deep Learning Method 33%