A Theoretical Analysis on Independence-driven Importance Weighting for Covariate-shift Generalization

Keyphrases

• Importance Weighting 100%
• Covariate Shift 100%
• Weighting Algorithm 40%
• Source Code 20%
• Selection Strategy 20%
• Regression Algorithm 20%
• Loss Function 20%
• Asymptotic Properties 20%
• Feature Selection 20%
• Learning Model 20%
• Test Distribution 20%
• Deep Neural Network 20%
• Squared Loss 20%
• Synthetic Experiment 20%
• Training Distribution 20%
• Stable Learning 20%
• Binary Cross-entropy Loss 20%
• Out-of-distribution Generalization 20%

Computer Science

• Covariate Shift 100%
• Selection Process 20%
• Source Codes 20%
• Good Performance 20%
• Feature Selection 20%
• Deep Neural Network 20%