VIDEO DOI: https://doi.org/10.48448/0arw-kj38

findings / work in progress

ACL 2022

May 23, 2022

Dublin, Ireland

"That Is a Suspicious Reaction!": Interpreting Logits Variation to Detect NLP Adversarial Attacks

Please log in to leave a comment

Downloads

SlidesTranscript English (automatic)

Next from ACL 2022

The Moral Debater: A Study on the Computational Generation of Morally Framed Arguments
findings / work in progress

The Moral Debater: A Study on the Computational Generation of Morally Framed Arguments

ACL 2022

+1Henning WachsmuthRoxanne El BaffMilad Alshomary
Milad Alshomary and 3 other authors

23 May 2022

Similar lecture

Detecting Word-Level Adversarial Text Attacks via SHapley Additive exPlanations
workshop paper

Detecting Word-Level Adversarial Text Attacks via SHapley Additive exPlanations

ACL 2022

+1Lukas HuberGeorg GrohMarc Alexander Kühn
Marc Alexander Kühn and 3 other authors

26 May 2022