Nonconvex matrix factorization from rank-one measurements

Keyphrases

• Near-optimal 100%
• Non-convexity 100%
• Matrix Factorization 100%
• Rank-one Measurements 100%
• Spectral Initialization 100%
• Computational Complexity 50%
• Phase Retrieval 50%
• Gradient Descent 50%
• Problem Size 50%
• Low-rank Matrix 50%
• Measurement Vector 50%
• Quantum Tomography 50%
• Optimal Sample 50%
• State Learning 50%
• Implicit Regularization 50%
• Sample Splitting 50%
• Low-rank Factor 50%
• Shallow Neural Network 50%
• Least Squares Loss 50%
• Sample Complexity 50%
• Computational Guarantees 50%
• Near Optimality 50%
• Vanilla Gradient Descent 50%

Mathematics

• Regularization 100%
• Matrix (Mathematics) 100%
• Optimality 50%
• Loss Function 50%
• Neural Network 50%
• Quantum State 50%
• Step Size 50%
• Square Loss 50%
• Least Squares Method 50%