R2M: Real-Aware Residual Model Merging for Robust and Generalizable Deepfake Detection

Citations

SCOPUS

0

초록

Deepfake generators evolve rapidly, making exhaustive data collection and repeated retraining impractical. Unlike generic multi-task settings, deepfake specialists share a common binary objective (Real vs. Fake) and mainly differ in generator-specific artifacts. However, naive parameter arithmetic can induce unintended decision-boundary shifts, causing unstable ranking behavior and degraded AUC under domain shift. We propose R2M, a training-free merging framework that decomposes task vectors into a shared component and generator-specific residuals, linking parameter-space updates to logit-space behavior. Offline spectral construction identifies shared and residual subspaces, and online routing selects residuals via first-order gradient-residual alignment, predicting per-sample logit updates under linearization. R2M is both efficient and interpretable: expensive computations are performed offline, while online inference requires only a single forward–backward pass with lightweight inner-product routing. The same formulation enables diagnostic analysis through margin and routing statistics. Experiments demonstrate consistently strong performance across in-domain, cross-domain, and unseen settings, highlighting R2M as an interpretable and scalable approach to training-free deepfake model merging.

제목
R2M: Real-Aware Residual Model Merging for Robust and Generalizable Deepfake Detection
저자
Park, Jinhee; Kim, Guisik; Cho, Choongsang; Kwon, Junseok
DOI
10.1007/978-3-032-37167-6_14
발행일
2026
유형
Conference Paper
저널명
Lecture Notes in Computer Science
권
17033
페이지
245 ~ 263