Upper bounds on the Lipschitz constant of Neural Networks using Polynomial Optimization
Includes the original code used in the experimental section of the paper
Lipschitz constant estimation of Neural Networks via sparse polynomial optimization Fabian Latorre, Paul Rolland, Volkan Cevher. ICLR 2020
link: https://openreview.net/forum?id=rJe4_xSFDB
the original code can be found in the folder old_code, while at the top
level you can find a new (in progress) version with faster and more efficient
implementation.
IMPORTANT: a bug was found in the original version of the code, in the
_compute_grad_poly method in old_code/polynomials.py. The code
has been fixed and we expect changes in experimental results to be minor.