Deep neural networks can stably solve high-dimensional, noisy, non-linear inverse problems.

Andrés Felipe Lerma Pineda Philipp Christian Petersen

Published in: CoRR (2022)

Keyphrases

inverse problems
neural network
high dimensional
image reconstruction
global optimization
convex optimization
high dimension
pattern recognition
parameter space
image segmentation
partial differential equations
optimization methods
computationally efficient
high dimensional data
early vision
emission tomography