Phase-change memory models for deep learning training and inference

Keyphrases

• Accurate Model 25%
• Analog Memory 25%
• Array-based 25%
• Conductance 25%
• Conductance Drift 25%
• Conductance States 25%
• Connection Strength 25%
• Connectivity Matrix 25%
• Deep Learning Framework 25%
• Deep Learning Hardware 25%
• Deep Learning Inference 100%
• Deep Learning Training 100%
• Deep Neural Network 25%
• Device Modeling 25%
• Efficient Inference 25%
• Hardware Efficiency 25%
• Hardware Implementation 50%
• In-memory Computing 50%
• Inference Engine 25%
• Inference Performance 50%
• Memory Array 25%
• Memory Device 50%
• Memory Models 100%
• Memory-based 25%
• Non-ideality 25%
• Non-volatile 25%
• On chip 25%
• Phase Change Random Access Memory (PCRAM) 100%
• Read Noise 25%
• Simple Device 25%
• State Dependence 25%
• Temporal Evolution 25%
• TensorFlow 25%
• Training Performance 50%

Computer Science

• Connection Strength 33%
• Connectivity Matrix 33%
• Deep Learning 100%
• Deep Neural Network 100%
• Hardware Implementation 66%
• Inference Engines 33%
• Learning Framework 33%
• Memory Array 33%
• Memory Model 100%
• Nonvolatile 33%
• Temporal Evolution 33%