A variance minimization criterion to active learning on graphs

Keyphrases

• Active Learning 100%
• Variance Minimization 100%
• Learning on Graphs 100%
• Minimization Criterion 100%
• Performance Improvement 50%
• Selection Criteria 50%
• Gaussian Random Field 50%
• Probability Distribution 50%
• Total Variance 50%
• Common Graphs 50%
• Unlabeled Data 50%
• Feature Representation 50%
• Label Information 50%
• Multivariate Normal 50%
• Random Field Model 50%
• Label Selection 50%
• Graph Smoothness 50%
• Smoothness Assumption 50%
• Offline Mode 50%
• Harmonic Solution 50%
• Expected Prediction Error 50%

Mathematics

• Probability Distribution 100%
• Variance 100%
• Gaussian Random Field 100%
• Prediction Error 100%
• Total Variance 100%
• Real-World Data 100%
• Multivariate Normal 100%
• Unlabeled Vertex 100%

Computer Science

• Active Learning 100%
• Selection Criterion 50%
• Unlabeled Data 50%
• Prediction Error 50%
• Random Field Model 50%
• Total Variance 50%
• Existing Label 50%
• Smoothness Assumption 50%