PiPSim: A Behavior-Level Modeling Tool for CNN Processing-in-Pixel Accelerators

Keyphrases

• Accurate Evaluation 33%
• Algorithmic Level 33%
• Behavioral Level 100%
• Circuit Level 66%
• Convolutional Neural Network 100%
• Design Decisions 33%
• Design Exploration 33%
• Design Optimization 33%
• Device Level 33%
• Early Design Phases 33%
• Error Rate 33%
• Flexible Interface 33%
• Hierarchical Structure 33%
• In-sensors 33%
• Inference Process 33%
• Learning Accuracy 33%
• Memory Technologies 33%
• Metric Accuracy 33%
• Model Performance 33%
• Modeling Tools 100%
• Neural Network Computation 100%
• Performance Learning 33%
• Performance Metrics 66%
• Pixel-based 33%
• Processing-in-pixel 100%
• Simulation Platform 33%
• Specialized Architectures 33%
• SPICE Model 33%
• Trade-off Relationship 33%
• Visual Internet of Things 33%

Computer Science

• Convolutional Neural Network 100%
• Deep Neural Network 50%
• Design Exploration 50%
• Inference Process 50%
• Internet of Things 50%
• Memory Technology 50%
• Performance Metric 100%
• Performance Model 50%
• Speed-up 100%