Towards Understanding the Generalization of Deepfake Detectors from a Game-Theoretical View

Kelu Yao, Jin Wang, Boyu Diao, Chao Li

Abstract

This paper aims to explain the generalization of deep-fake detectors from the novel perspective of multi-order interactions among visual concepts. Specifically, we propose three hypotheses: 1. Deepfake detectors encode multi-order interactions among visual concepts, in which the low-order interactions usually have substantially negative contributions to deepfake detection. 2. Deepfake detectors with better generalization abilities tend to encode low-order interactions with fewer negative contributions. 3. Generalized deepfake detectors usually weaken the negative contributions of low-order interactions by suppressing their strength. Accordingly, we design several mathematical metrics to evaluate the effect of low-order interaction for deepfake detectors. Extensive comparative experiments are conducted, which verify the soundness of our hypotheses. Based on the analyses, we further propose a generic method, which directly reduces the toxic effects of low-order interactions to improve the generalization of deepfake detectors to some extent.
BibTeX
@inproceedings{yao2023towards,
  author = {Kelu Yao and Jin Wang and Boyu Diao and Chao Li},
  title = {{Towards Understanding the Generalization of Deepfake Detectors from a Game-Theoretical View}},
  booktitle = {IEEE/CVF International Conference on Computer Vision},
  year = {2023},
  doi = {10.1109/iccv51070.2023.00194},
}

Loading related content...

Figure