On Convergence of Training Loss Without Reaching Stationary Points.
Jingzhao ZhangHaochuan LiSuvrit SraAli JadbabaiePublished in: CoRR (2021)
Keyphrases
- stationary points
- convergence analysis
- fixed point
- constrained optimization
- objective function
- nonlinear programming
- mathematical programming
- karush kuhn tucker
- training set
- training samples
- sufficient conditions
- linear programming
- cost function
- polynomial equations
- convergence speed
- variational inequalities
- neural network