Neural Network Training with Stochastic Hardware Models and Software Abstractions

Keyphrases

• Neural Network Training 100%
• Hardware Complexity 100%
• Hardware Model 100%
• Stochastic Data 100%
• Model Abstraction 100%
• Hardware Resilience 100%
• Software Abstraction 100%
• Energy Throughput 50%
• Energy Restriction 50%
• Inference Performance 50%
• Benchmark Dataset 25%
• Statistical Model 25%
• Operating Point 25%
• Device Variations 25%
• Computationally Intensive 25%
• Energy Efficiency 25%
• Statistical Distribution 25%
• MNIST 25%
• Device Technology 25%
• Composable 25%
• Statistical Computing 25%
• Models of Computation 25%
• Constrained Application 25%
• Training Algorithm 25%
• Foundry 25%
• Variation Parameters 25%
• In-memory Computing 25%
• Parameter Learning 25%
• Deep Learning Model 25%
• CIFAR-10 25%
• MRAM Devices 25%
• Hardware Variation 25%
• Model Training 25%
• Machine Learning Inference 25%

Computer Science

• Energy Efficiency 100%
• Computer Architecture 100%
• training algorithm 100%
• Neural Network Training 100%
• Participatory Design 100%
• Model of Computation 100%
• Parameter Variation 100%
• Learning Parameter 100%
• Deep Learning Model 100%
• Training Instance 100%
• Machine Learning 100%
• Learning System 100%
• Statistical Distribution 100%