Certifiers make neural networks vulnerable to availability attacks
To achieve reliable, robust, and safe AI systems, it is vital to implement fallback strategies when AI predictions cannot be trusted. Certifiers for neural networks are a reliable way to check the robustness of these predictions. They guarantee for some predictions that a certain class of manipulati...
Κύριοι συγγραφείς: | Lorenz, T, Kwiatkowska, M, Fritz, M |
---|---|
Μορφή: | Conference item |
Γλώσσα: | English |
Έκδοση: |
Association for Computing Machinery
2023
|
Παρόμοια τεκμήρια
Παρόμοια τεκμήρια
-
FullCert: deterministic end-to-end certification for training and inference of neural networks
ανά: Lorenz, T, κ.ά.
Έκδοση: (2024) -
Certified Robustness to Text Adversarial Attacks by Randomized [MASK]
ανά: Jiehang Zeng, κ.ά.
Έκδοση: (2023-06-01) -
Attack Vulnerability of Network Controllability.
ανά: Zhe-Ming Lu, κ.ά.
Έκδοση: (2016-01-01) -
Bayesian inference with certifiable adversarial robustness
ανά: Wicker, M, κ.ά.
Έκδοση: (2021) -
Vulnerability analysis on noise-injection based hardware attack on deep neural networks
ανά: Liu, Wenye, κ.ά.
Έκδοση: (2020)