AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

Towards Performance Consistency in Multi-Level Model Collaboration

Qi Li, Runpeng Yu, Xinchao Wang

ICCV 2025 · Architectures and efficiency › Pruning and efficient tuning · approach: Parameter-efficient transfer analysis · 0 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
use NeuLig CoopVecs instead of simple averaging for mergingCLIP-ViT-B/32 5 models · 5-dataset avgmerging accuracy: 78.19 -> 95.44Improved
use NeuLig instead of Simple-AveragingCLIP-RN50 5 models · 5-dataset avgmerging accuracy: 50.39 -> 92.07Improved
use NeuLig instead of WeMoECLIP-ViT-B/32 5 models · 5-dataset avgmerge-ensemble gap: 1.32 -> 0.02Improved
use NeuLig instead of Simple-AveragingCLIP-RN50 5 models · 5-dataset avgmerge-ensemble gap: 39.34 -> 0.61Improved
use supervised instead of semi-supervised boosting term (ablation)NeuLig CLIP-ViT-B/32 · 5-dataset avgmerging accuracy: 95.44 -> 95.37No clear effect
use supervised instead of semi-supervised boosting term (ablation)NeuLig CLIP-RN50 · 5-dataset avgmerge-ensemble gap: 0.61 -> 1.22Hurt
increase collaborating CLIP-RN50 models from 2 to 5 (ablation)NeuLig · avg accuracymerge-ensemble gap: 0.17 -> 1.22Hurt
reduce unlabeled data scale from 1.0 to 0.01 (ablation)NeuLig CLIP-ViT-B/32 · semi-supervisedmerging accuracy: 95.44 -> 91.43Hurt
reduce unlabeled data scale from 1.0 to 0.01 (ablation)NeuLig CLIP-RN50 · semi-supervisedmerging accuracy: 92.07 -> 80.93Hurt
reduce unlabeled data scale from 1.0 to 0.1 (ablation)NeuLig CLIP-ViT-B/32 · semi-supervisedmerging accuracy: 95.44 -> 93.46Hurt
train models from random initialization instead of shared checkpoint (ablation)NeuLig · 2 models GTSRB-MNISTmerge vs ensemble: 82.17 vs 82.19No clear effect
compare ViT with ResNet backboneSimple-Averaging · 5-dataset avgmerging accuracy: 50.39 -> 78.19Mixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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