AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

ACE-Merging: Data-Free Model Merging with Adaptive Covariance Estimation

Bo Xu, Haotian Wu, Hehai Lin, Weiquan Huang, Beier Zhu, Yao Shu, Chengwei Qin

CVPR 2026 · Architectures and efficiency › Pruning and efficient tuning · approach: Low-rank compression and merging · 4 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 adaptive covariance normalization (ablation)basic closed-form merge · GLUE (RoBERTa-Large)avg score (%): 80.05 -> 88.04Improved
add adaptive covariance normalization (ablation)basic closed-form merge · GLUE (GPT-2)avg score (%): 68.72 -> 71.50Improved
add collective structural prior (ablation)ACE-Merging E2 · GLUE (RoBERTa-Large)avg score (%): 88.04 -> 86.79Hurt
add collective structural prior (ablation)ACE-Merging E2 · GLUE (GPT-2)avg score (%): 71.50 -> 71.51No clear effect
add collective structural prior (ablation)ACE-Merging E2 · Vision (ViT-B/16, 8 tasks)avg score (%): 89.91 -> 90.60Improved
add spectral refinement (ablation)ACE-Merging E3 · GLUE (RoBERTa-Large)avg score (%): 86.79 -> 91.68Improved
add spectral refinement (ablation)ACE-Merging E3 · GLUE (GPT-2)avg score (%): 71.51 -> 74.09Improved
raise heterogeneity threshold from 0.3 to 0.5 (ablation)ACE-Merging · GLUE (GPT-2)avg accuracy (%): 74.09 -> 69.35Hurt
raise heterogeneity threshold from 0.3 to 0.5 (ablation)ACE-Merging · GLUE (RoBERTa-Large)avg accuracy (%): 91.68 -> 88.09Hurt
vary spectral rank fraction 0.1 to 0.5 (ablation)ACE-Merging · GLUE (GPT-2)avg accuracy (%): 74.17 / 74.09 / 73.78No clear effect
use ACE-Merging instead of WUDI-Mergingdata-free merging · GLUE normalized (RoBERTa-Base)avg score: 85.3 -> 90.4Improved
use ACE-Merging instead of Iso-CTSdata-free merging · Vision (ViT-L/14, 20 tasks)avg accuracy (%): 90.1 -> 89.5Hurt

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