Work

Robust Deepfake Detection by Addressing Generalization and Trustworthiness Challenges: A Short Survey