Deep Active Learning by Leveraging Training Dynamics

Keyphrases

• Training Dynamics 100%
• Deep Active Learning 100%
• Generalization Performance 75%
• Active Learning 50%
• Real Application 50%
• Deep Learning Model 50%
• Learning Theory 25%
• Time Complexity 25%
• Subset Selection Problem 25%
• Learning Methods 25%
• Theory-driven 25%
• Learning Settings 25%
• Convergence Rate 25%
• Neural Data 25%
• Statistical Learning 25%
• Network Impact 25%
• Practical Effects 25%
• Deep Neural Network 25%
• Empirical Criteria 25%
• Active Learning Methods 25%
• Classical Statistical 25%
• Deep Network 25%
• Ultra-wide 25%

Computer Science

• Active Learning 100%
• Generalization Performance 50%
• Real Application 33%
• Deep Learning Model 33%
• Time Complexity 16%
• Feature Subset Selection 16%
• Speed Convergence 16%
• Deep Neural Network 16%

Chemical Engineering

• Deep Learning 100%
• Deep Neural Network 50%