NNSplit-SØREN: Supporting the model implementation of large neural networks in a programmable data plane

Keyphrases

• Neural Network 100%
• Implementation Model 100%
• Network Split 100%
• Wide Neural Network 100%
• Programmable Data Plane 100%
• Neural Network Model 60%
• Network Traffic 40%
• Processing Time 40%
• Cost Performance 40%
• Neural Network Architecture 40%
• Network Layer 40%
• Protocol Header 20%
• Prediction Accuracy 20%
• Multiple Data 20%
• Data Packet 20%
• Computational Cost 20%
• Model Design 20%
• Memory Size 20%
• Functional Performance 20%
• Computational Power 20%
• Computer Networks 20%
• Memory Consumption 20%
• Network Switch 20%
• Support Layer 20%
• Model Complexity 20%
• Multi-class Classification 20%
• Layer Splitting 20%
• Limited Memory 20%
• Average Processing 20%
• Round Trip Time 20%
• Training Costs 20%
• Performance Problems 20%
• P4 Language 20%
• Large-scale Neural Networks 20%
• Non-splitting 20%
• Traffic Throughput 20%

Engineering

• Network Model 100%
• Experimental Result 66%
• Processing Time 66%
• Layer Network 66%
• Neural Network Architecture 66%
• Cost Performance 66%
• Parameter Model 33%
• Input Parameter 33%
• Data Packet 33%
• Computational Cost 33%
• Functional Performance 33%
• Computational Power 33%
• Multiclass Classification 33%
• Complex Model 33%
• Round Trip Time 33%
• Support Layer 33%

Computer Science

• Neural Network 100%
• Implementation Model 100%
• Neural Network Model 60%
• Experimental Result 40%
• Network Traffic 40%
• Processing Time 40%
• Neural Network Architecture 40%
• Layer Neural Network 40%
• Use Case 20%
• Input Parameter 20%
• Computational Cost 20%
• Prediction Accuracy 20%
• Computer Network 20%
• Computational Power 20%
• Multi Class Classification 20%
• Multiclass Classification 20%
• Round-Trip Time 20%
• Performance Problem 20%