DeformRS: Certifying input deformations with randomized smoothing
Deep neural networks are vulnerable to input deformations in the form of vector fields of pixel displacements and to other parameterized geometric deformations e.g. translations, rotations, etc. Current input deformation certification methods either 1. do not scale to deep networks on large input da...
Главные авторы: | Alfarra, M, Bibi, A, Khan, N, Torr, P, Ghanem, B |
---|---|
Формат: | Conference item |
Язык: | English |
Опубликовано: |
Association for the Advancement of Artificial Intelligence
2022
|
Схожие документы
-
Data dependent randomized smoothing
по: Alfarra, M, и др.
Опубликовано: (2022) -
The bulk Higgs in the deformed RS model
по: F. Mahmoudi, и др.
Опубликовано: (2018-09-01) -
Deformations of smooth functions on 2-torus
по: Bohdan Feshchenko
Опубликовано: (2019-12-01) -
RANCER: non-axis aligned anisotropic certification with randomized smoothing
по: Rumezhak, T, и др.
Опубликовано: (2023) -
Combating adversaries with anti-adversaries
по: Alfarra, M, и др.
Опубликовано: (2022)