BILGO: Bilateral greedy optimization for large scale semidefinite programming

Computer Science

• Computational Efficiency 50%
• Descent Direction 50%
• Desirable Property 50%
• Eigenvector 50%
• Experimental Result 50%
• Global Optimal Solution 50%
• Gradient Descent 50%
• Learning Problem 50%
• Linear Combination 50%
• Machine Learning 100%
• Search Strategies 50%
• Semidefinite Programming 100%
• Update Algorithm 50%

Keyphrases

• Descent Direction 20%
• Gradient Descent 20%
• KKT Conditions 20%
• Large-scale Data 20%
• Leading Eigenvector 20%
• Machine Learning Task 40%
• Mahalanobis Metric Learning 20%
• Manifold Learning 20%
• Matrix Solution 20%
• Maximum Variance Unfolding 20%
• Metric Learning 20%
• Optimization Iteration 20%
• Programming Algorithm 20%
• Rank-1 Matrices 20%
• Rank-1 Update 20%
• Scale Efficiency 20%
• Semidefinite Matrix 20%
• Updating Algorithm 20%