 |
Lipschitz stability of Deep Neural Networks in view of applications
1 : Laboratoire Jacques-Louis Lions
(LJLL (UMR_7598))
Laboratoire Jacques-Louis Lions (LJLL)
Sorbonne-Université, Boîte courrier 187 - 75252 Paris Cedex 05 -
France
The use of functions constructed by deep neural networks is attractive for data-driven applications, and is an active area of research at present.
However, it is almost universally observed that the stability of these functions is difficult to guarantee.
In fact, we can expect this stability problem to grow in tandem with applications in sciences and technology.
With Moreno Pintore, we focused on deriving computable upper bounds on the Lipschitz constant of deep neural networks and obtained new estimators.
The optimality of the estimators will be discussed and compared with the literature.
