An accelerated non-Euclidean hybrid proximal extragradient-type algorithm for convex-concave saddle-point problems.
O. KolossoskiRenato D. C. MonteiroPublished in: Optim. Methods Softw. (2017)
Keyphrases
- saddle point
- convex concave
- primal dual
- discrete tomography
- variational inequalities
- solution path
- penalty function
- convex hull
- dynamic programming
- discrete space
- maximum margin
- global constraints
- learning algorithm
- probabilistic model
- optimal solution
- constrained optimization
- sensitivity analysis
- objective function
- genetic algorithm
- approximation algorithms
- convergence rate
- structured prediction
- expectation maximization
- linear programming
- structured output