Neural Network Training as an Optimal Control Problem : - An Augmented Lagrangian Approach -.
Brecht EvensPuya LatafatAndreas ThemelisJohan A. K. SuykensPanagiotis PatrinosPublished in: CDC (2021)
Keyphrases
- optimal control
- neural network training
- augmented lagrangian
- constrained optimization
- neural network
- solving nonlinear
- total variation
- training algorithm
- constrained optimization problems
- convex optimization
- image denoising
- dynamic programming
- linear programming problems
- control strategy
- lagrange multipliers
- reinforcement learning
- denoising
- optimization method
- regularization term
- linear programming
- back propagation
- differential evolution
- particle swarm optimisation
- penalty function
- primal dual
- particle swarm optimization
- linear program