Tackling Small Eigen-Gaps: Fine-Grained Eigenvector Estimation and Inference under Heteroscedastic Noise

Mathematics

• Additive Noise 16%
• Asymmetric 16%
• Construct Confidence Interval 16%
• Eigenvalue 33%
• Eigenvector 100%
• Gaussian Distribution 16%
• Linear Functionals 33%
• Matrix (Mathematics) 83%
• Minimax 16%
• Perturbation Theory 16%
• Uncertainty Quantification 16%

Computer Science

• Data Matrix 16%
• Eigenvalue 33%
• Eigenvector 100%
• Gaussian White Noise 16%
• Noisy Observation 16%
• Quantification Procedure 16%

Keyphrases

• Generic Matrices 16%
• Heteroscedastic Noise 100%
• Statistical Reasoning 16%