AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

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

Jinhee Park, Guisik Kim, Choongsang Cho, Junseok Kwon

ECCV 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Dense features and foundation priors

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add common-alignment calibration to residual routing (ablation)R2M · DF40 CDF (rank r0=0, rA=1)mean AUC: 0.398 -> 0.830Improved
add common-alignment calibration at best ranks (ablation)R2M · DF40 FF (r0=2, rA=3)mean AUC: 0.980 -> 0.992Improved
increase residual rank rA from 1 to 3 and r0 from 0 to 2 (ablation)R2M · DF40 Unseenmean AUC: 0.792 -> 0.818Improved
remove dominant Real-sensitive directions (r0>0) (ablation)R2M · DF40 CDFmean AUC (calibrated): 0.880 -> 0.901Improved
use R2M over best training-free baseline CARTR2M · DF40 Protocol 3mean AUC: 0.766 -> 0.818Improved
use R2M over CARTR2M · DF40 Protocol 2mean AUC: 0.877 -> 0.901Improved
use R2M over CARTR2M · DF40 Protocol 1mean AUC: 0.988 -> 0.992Improved
merge specialists instead of training all-in-one modelWeight Averaging · DF40 Protocol 3mean AUC: 0.665 -> 0.724Improved
R2M vs Task Arithmetic on CDF EFSR2M · DF40 CDF EFSAUC: 0.926 -> 0.898Hurt
add domain grouping before routing six heterogeneous specialists (ablation)R2M · six specialists, FF FSAUC: 0.778 -> 0.929Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy