Learning efficient detector with semi-supervised adaptive distillation

Keyphrases

• Distillation 100%
• Distillation Loss 75%
• Student Model 50%
• Popular 25%
• Neural Network 25%
• Convolutional Neural Network 25%
• ResNet50 25%
• Hard Samples 25%
• Object Detection Algorithm 25%
• Model Compression 25%
• ResNet101 25%
• Logit 25%
• Soft Targets 25%
• Existing Data 25%
• Teacher Model 25%
• Overall Accuracy 25%
• Semi-supervised Setting 25%
• One-stage Detector 25%
• Data Distillation 25%
• Hard Target 25%
• Unlabeled Data 25%
• Computation Complexity 25%
• Student Performance Prediction 25%
• Knowledge Distillation 25%
• Large Gap 25%

Computer Science

• Residual Neural Network 100%
• Convolutional Neural Network 50%
• Unlabeled Data 50%
• Speed-up 50%
• Object Detection 50%
• Computation Complexity 50%
• Model Compression 50%
• Knowledge Distillation 50%

Engineering

• Convolutional Neural Network 100%
• Computation Complexity 100%
• Detection Technique 100%