On obtaining sparse semantic solutions for inverse problems, control, and neural network training.
David A. B. HydeMichael BaoRonald FedkiwPublished in: J. Comput. Phys. (2021)
Keyphrases
- inverse problems
- neural network training
- image reconstruction
- global optimization
- convex optimization
- training algorithm
- optimization methods
- optimization method
- neural network
- optimal solution
- partial differential equations
- simulated annealing
- early vision
- computationally expensive
- sparse representation
- optimization problems
- benchmark problems
- high dimensional
- computer vision
- machine learning