Perturbed Proximal Descent to Escape Saddle Points for Non-convex and Non-smooth Objective Functions.
Zhishen HuangStephen BeckerPublished in: CoRR (2019)
Keyphrases
- saddle points
- saddle point
- objective function
- convex functions
- penalty function
- scale space
- critical points
- primal dual
- convex optimization
- optimization problems
- variational inequalities
- endpoints
- convex sets
- optimal solution
- interior point
- original data
- constrained optimization
- numerical methods
- linear program
- feature space
- global constraints
- structured prediction
- convex relaxation
- convex hull
- discrete space
- linear programming
- genetic algorithm