fesarikaya / CEIAView on GitHub
This novel adversarial attack method that I have developed is called CEIA (Contextual Embedding Inversion Attack). This method is a sophisticated and novel approach to creating adversarial examples for BERT-based models, particularly targeting question-answering systems.
16Nov 19, 2024Updated last year

Alternatives and similar repositories for CEIA

Users that are interested in CEIA are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.

Sorting:

Are these results useful?