AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

Disrupting Model Merging: A Parameter-Level Defense Without Sacrificing Accuracy

Wei Junhao, Yu Zhe, Jun Sakuma

ICCV 2025 · Image generation › Text-to-image diffusion · approach: Fairness and bias in generative models · 5 citations (OpenAlex)

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 MLP rearrangement and random head scaling (PaRaMS) to the defender modelTA-merged ViT-B-32 · GTSRB/DTD pairfree-rider task accuracy of merged model: 67.82 -> 2.66Improved
add scaling module to rearrangement-only protection (ablation)PaRaMS on ViT-L-14 · Cars defender, avg MMP+accuracy (lower is better defense): 5.90 -> 4.99Improved
add rearrangement module to scaling-only protection (ablation)PaRaMS on ViT-L-14 · Cars defender, avg MMP+accuracy (lower is better defense): 10.92 -> 4.99Improved
add 10% dropout pruning to PaRaMS against an adaptive free-rider (ablation)PaRaMS ViT-B-32 · SVHN defender merged accuracyaccuracy (lower is better defense): 80.32% -> 64.35%Improved
add 10% dropout pruning to the unmerged defender model (ablation)PaRaMS ViT-B-32 · SVHNunmerged accuracy: 97.46% to 96.33%Hurt
free-rider uses adaptive recovery against PaRaMSPaRaMS ViT-B-32 · SVHN defendermerged accuracy: 7.12% -> 80.32%Hurt
free-rider uses knowledge distillation against PaRaMS+dropoutPaRaMS+dropout · SVHN defendermerged accuracy: 7.01% -> 82.27%Hurt
increase TA merging coefficient from 0.3 to 0.8 (ablation)PaRaMS ViT-B-32 · EuroSAT defenderMMP+ accuracy: MMP+ at most less than 25%No clear effect
increase number of merged models from 2 to 7 (ablation)PaRaMS ViT-B-32/ViT-L-14 · MNIST defenderMMP+ accuracy: around 6-7%No clear effect
apply PaRaMS to Stable Diffusion 1.5 merged by TAPaRaMS · Prompt1 alignmentCLIP alignment: 0.3416 -> 0.1277Improved
apply PaRaMS to Llama2 full-finetune merged by TAPaRaMS · Emotionaccuracy of merged model: 97.6 -> 21.4Improved
apply scaling-only PaRaMS to PEM-TA merged LoRAPaRaMS · Emotionaccuracy of merged model: 96.1 -> 52.7Improved

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