Neural Networks are Convex Regularizers: Exact Polynomial-time Convex Optimization Formulations for Two-Layer Networks.
Mert PilanciTolga ErgenPublished in: CoRR (2020)
Keyphrases
- convex optimization
- convex formulation
- neural network
- total variation
- convex optimization problems
- convex relaxation
- interior point methods
- convex programming
- primal dual
- special case
- convex constraints
- alternating direction method of multipliers
- image restoration
- low rank
- augmented lagrangian
- convex sets
- norm minimization
- computational complexity
- semi definite programming
- semidefinite program
- operator splitting
- denoising
- object recognition
- gradient method
- image denoising
- linear combination
- wavelet transform
- basis pursuit
- multiscale
- image processing