Logarithmic regret for online gradient descent beyond strong convexity

Keyphrases

• Online Gradient Descent 100%
• Logarithmic Regret 100%
• Strongly Convex 100%
• Strong Convexity 100%
• Polyhedral Set 66%
• Online Convex Optimization 33%
• Regret 33%
• Convex Optimization 33%
• Adversarial Setting 33%
• Gradient Method 33%
• Large-scale Problems 33%
• Polytope 33%
• Fast Rates 33%
• First-order Methods 33%
• Strongly Convex Function 33%
• Low-rank 33%
• Convex Loss Function 33%
• Non-convex Cost Function 33%
• Online Newton Step 33%
• Stochastic Sequences 33%
• Hoffman Bound 33%
• Step Algorithm 33%

Mathematics

• Polyhedral Set 100%
• Loss Function 50%
• Polytope 50%
• Step Newton 50%
• Stochastics 50%
• Scale Problem 50%
• Convex Function 50%