I would also like to ask three question. Do you know how well your model generalizes to video/audio deepfakes created by models that are not within your training sets? And also have you investigated whether your model can be used in a GAN setting to improve a deepfake generator towards creating better fakes? Or how robust your detectors are against adversarial attacks?
I would also like to ask three question. Do you know how well your model generalizes to video/audio deepfakes created by models that are not within your training sets? And also have you investigated whether your model can be used in a GAN setting to improve a deepfake generator towards creating better fakes? Or how robust your detectors are against adversarial attacks?